Hyperglycemia, nutrition and health outcomes in preterm infants

Hyperglycemia, nutrition and

health outcomes in preterm infants

Itay Zamir

Department of Clinical Sciences, Pediatrics,

Umeå 2020

Responsible publisher under Swedish law: the Dean of the Medical Faculty This work is protected by the Swedish Copyright Legislation (Act 1960:729) Dissertation for PhD ISBN: 978-91-7855-226-9 ISSN: 0346-6612 New Series No.: 2079 Cover photo: Hadas Zamir Electronic version available at: http://umu.diva-portal.org/ Printed by: UmU Print Service, Umeå University Umeå, Sweden 2020

To my grandfather, Yitzhak

“Let me be contented in everything except in the great

science of my profession. Never allow the thought to

arise in me that I have attained to sufficient

knowledge, but vouchsafe to me the strength, the

leisure and the ambition ever to extend my

knowledge. For art is great, but the mind of man is

ever expanding.”

The prayer of the physician, Maimonides

i

Table of Contents

Table of contents ................................................................................ i

Abstract ............................................................................................ iv

Enkel sammanfattning på svenska ................................................... vi

Abbreviations ................................................................................. viii

Original papers ................................................................................. x

Background ........................................................................................ 1 Extremely preterm and very low birth weight infants .................................................... 1 Why is glucose so important and how much is needed? ................................................ 2 Definition(s) of Hyperglycemia in preterm infants ........................................................ 2 Prevalence of Hyperglycemia ...........................................................................................3 Pathogenesis of hyperglycemia ........................................................................................ 5

Normal glucose metabolism ...................................................................................... 5 Insulin resistance ...................................................................................................... 6 Relative insulin deficiency – β-cell dysfunction ....................................................... 7 Relative insulin deficiency – deficient incretin response ........................................ 8 Insufficient control of glucose production ............................................................... 8 Other mechanisms ..................................................................................................... 9 The pathophysiological effects of hyperglycemia ................................................... 9

Risk factors for hyperglycemia ........................................................................................ 9 Prematurity, low birth weight and growth retardation ........................................ 9 Asphyxia .................................................................................................................... 9 Respiratory distress syndrome and ventilator support ........................................ 10 Parenteral glucose intake ........................................................................................ 10 Parenteral lipid intake ............................................................................................. 11 Steroid treatment ..................................................................................................... 11 Inotrope treatment ................................................................................................... 12 Sepsis ......................................................................................................................... 12 Stress ......................................................................................................................... 12 Hypophosphatemia .................................................................................................. 13 Neonatal diabetes mellitus ...................................................................................... 13

Outcomes of hyperglycemia ........................................................................................... 13 Mortality ................................................................................................................... 13 Retinopathy of prematurity .................................................................................... 14 Growth ...................................................................................................................... 14 Intraventricular hemorrhage .................................................................................. 14 Brain function and neurodevelopmental disability ............................................... 15 Sepsis ......................................................................................................................... 16 Necrotizing enterocolitis .......................................................................................... 16 Blood pressure .......................................................................................................... 16

Management of hyperglycemia ...................................................................................... 18 Reduction of glucose infusion rate .......................................................................... 18 Adjusting parenteral amino acid and lipid intakes ............................................... 18

ii

Insulin treatment...................................................................................................... 19 Hypoglycemia ................................................................................................................ 23

Definition ................................................................................................................. 23 Prevalence ................................................................................................................ 24 Pathophysiology ...................................................................................................... 24 Risk factors .............................................................................................................. 25 Outcomes.................................................................................................................. 25 Treatment ................................................................................................................ 25

Blood pressure ............................................................................................................... 26 Mechanisms ............................................................................................................. 26 Early programming hypothesis ............................................................................. 26

Objectives ........................................................................................ 29

Materials and Methods ................................................................... 30 Study population............................................................................................................ 30

The EXPRESS cohort (Papers I-III) ....................................................................... 30 The LIGHT study (Paper IV) ................................................................................... 31

Data collection ............................................................................................................... 32 Perinatal data .......................................................................................................... 32 Glucose data ............................................................................................................ 32 Insulin data .............................................................................................................. 32 Anthropometric data .............................................................................................. 33 Nutrition data (Papers I, II) ................................................................................... 33 Blood pressure data (Paper II) ............................................................................... 33 Neurodevelopmental outcomes data (Paper III) .................................................. 33 Continuous glucose monitoring data (Paper IV) .................................................. 34 Quality control......................................................................................................... 36

Definition of hyperglycemia .......................................................................................... 36 Statistical analysis........................................................................................................... 37 Ethical approval ............................................................................................................. 38 Power analysis ............................................................................................................... 39

Results ............................................................................................ 40 Patient characteristics ................................................................................................... 40 Prevalence and duration of glucose disturbances ........................................................ 40

Neonatal hyperglycemia in EPT infants................................................................ 40 Glucose disturbances at postmenstrual age 36 weeks in VLBW infants .............. 41

Risk factors for glucose disturbances ............................................................................ 41 Nutrition and Neonatal hyperglycemia in EPT infants ........................................ 41 Glucose disturbances at postmenstrual age 36 weeks in VLBW infants .............. 41

Outcomes of neonatal hyperglycemia in EPT infants .................................................. 44 Mortality .................................................................................................................. 44 Blood pressure ......................................................................................................... 44 Neurodevelopmental disabilities ............................................................................ 45

Insulin treatment in EPT infants ................................................................................... 47 Blood pressure at 6.5 years of age in EPT-born children .............................................. 47

Early postnatal nutrition and blood pressure ........................................................ 47 Early postnatal growth and blood pressure ......................................................... 48

iii

Discussion ....................................................................................... 49 Discussion of methodology ........................................................................................... 49

Study design ............................................................................................................ 49 Study population ..................................................................................................... 49 Drop-out and attrition bias .................................................................................... 49 Confounding ............................................................................................................ 50 Multiple testing ........................................................................................................ 50 Data quality .............................................................................................................. 51 Continuous glucose monitoring method ................................................................. 51 Point-of-care glucometers ...................................................................................... 52

Ethical considerations ....................................................................................................53 Discussion of the results .................................................................................................53

Prevalence and duration of glucose disturbances .................................................53 Risk factors for glucose disturbances .................................................................... 54 Outcomes of neonatal hyperglycemia in EPT infants........................................... 56 Definition of neonatal hyperglycemia – a continuum .......................................... 58 Insulin treatment in EPT infants ............................................................................ 58 Blood pressure at 6.5 years of age in children born EPT ..................................... 60

Main findings .................................................................................. 62

Clinical implications and future research ....................................... 63

Acknowledgements ......................................................................... 64

Funding ............................................................................................67

References ...................................................................................... 68

iv

Abstract

Background

Survival among very low birth weight (VLBW) and extremely preterm (EPT)

infants has increased markedly during the last decades. Neonatal hyperglycemia

is common in these infants and is known to be associated with adverse outcomes.

However, much about neonatal hyperglycemia is still unknown: which

mechanisms are responsible for it, its long-term risk profile, how to treat it and

even how to define it. This thesis focuses on neonatal hyperglycemia in preterm-

born infants - its prevalence, possible causes (including postnatal nutrition and

its possible programming effect of later outcomes), consequences and treatment.

Methods

Two cohorts were studied in this thesis. The EXtremely PREterm infants in

Sweden Study (EXPRESS) – a national population-based cohort - included all

infants born before completed 27 gestational weeks during the years 2004-2007

in Sweden. Nutritional data as well as glucose measurements (n=9850) and

insulin treatment data for the first 28 days of life were obtained for 580 infants.

In a sub-cohort of 171 children, blood pressure measurements were obtained at a

follow-up visit at 6.5 years of age. Neurodevelopment was assessed at 6.5 years of

age in 436 children. Intellectual ability was assessed using Wechsler Intelligence

Scale for Children IV (WISC-IV; n=355). Gross and fine motor function were

assessed using Movement Assessment Battery for Children 2 (MABC-2; n=345).

The very Low birth weight Infants - Glucose and Hormonal profiles over Time

(LIGHT) study was a prospective cohort study that included 50 VLBW infants

born in Umeå, Sweden, between 2016-2019. Infants were placed on continuous

glucose monitoring (CGM) for a 48-hour period when a postmenstrual age (PMA)

of 36 gestational weeks was reached (n=35).

Results

Daily prevalence of hyperglycemia > 10 mmol/L in EPT infants was high during

the first two weeks of life (up to 30%), followed by a slow decrease thereafter.

Protracted hyperglycemia > 8 mmol/L was detected in more than half of VLBW

infants at PMA 36 weeks.

In EPT infants, glucose concentrations during the first 28 days of life were

increased by 1.6% on the day following an increase of 1 g/kg/day in parenteral

carbohydrate intake. Male sex, amnionitis and prior hypoglycemia and

v

hyperglycemia episodes were risk factors for hyperglycemia at PMA 36 weeks in

VLBW infants.

In EPT infants, neonatal hyperglycemia > 10 mmol/L was associated with

increased 28-day mortality. Neonatal hyperglycemia, its severity and duration

were associated with increased diastolic blood pressure (DBP) and lower

MABC-2 scores at 6.5 years of age. Insulin treatment was associated with

improved 28- and 70-day survival but not with BP or neurodevelopmental

outcomes at 6.5 years of age.

Higher protein intake during the first eight weeks of life was associated with

higher DBP at 6.5 years of age. An increase of the carbohydrate intake by

1 g/kg/day during this period was associated with an increase of 0.18 SDS in

systolic (SBP) and 0.14 SDS in DBP at 6.5 years. Growth during the same period

was not associated with BP at 6.5 years.

Conclusions

Neonatal hyperglycemia during the first four weeks of life was more common in

EPT infants than has previously been described. Remaining glucose disturbances

were common at PMA 36 weeks in VLBW infants. Parenteral glucose intakes

within the range given seems to have had low contribution to glucose

concentration variability. Neonatal hyperglycemia was associated with increased

28-day mortality as well as with increased blood pressure and reduced motor

skills at 6.5 years of age. Carbohydrate intake in the immediate postnatal period

was indepentently associated with increased blood pressure at 6.5 years of age.

The results add to the knowledge regarding important risk factors and health

effects of neonatal hyperglycemia. Different treatment options for neonatal

hyperglycemia should be evaluated in animal models and in randomized

controlled clinical trials in premature infants.

vi

Enkel sammanfattning på svenska

Bakgrund

Överlevnaden hos för tidigt födda barn har förbättrats markant under de senaste

decennierna. Hos dessa barn är förhöjda blodsockernivåer under de första 28

dagarna av livet (neonatal hyperglykemi) vanligt förekommande och är

förknippat med sämre kortsiktiga hälsoutfall. Mycket kvarstår som inte är känt

kring neonatal hyperglykemi: vilka mekanismer som orsakar det, vad de

långsiktiga hälsoutfallen blir, hur hyperglykemi bör behandlas samt hur den bör

definieras. Denna avhandling fokuserar på neonatal hyperglykemi hos för tidigt

födda barn, dess förekomst, eventuella riskfaktorer, konsekvenser och

behandling. Gällande riskfaktorer undersöktes bland annat den näring som

tillförs spädbarnet i början av livet och dess möjliga påverkan på hälsoutfall

senare i livet.

Metoder

Delarbeten I-III är baserade på data från EXtremely PREterm infants in Sweden

Study (EXPRESS) – en nationell populationsbaserad studie som inkluderade alla

barn som föddes i Sverige före 27 graviditetsveckor under åren 2004–2007. Data

kring nutrition, blodsockernivåer (totalt 9850 värden) samt insulinbehandling

registrerades dagligen för de första 28 dagarna av livet hos 580 barn. Hos en

grupp som inkluderade 171 av dessa barn mättes blodtryck vid en uppföljning vid

6,5-års ålder. Neurologisk utveckling utvärderades hos 436 barn vid 6,5-års ålder.

Kognitiv utveckling utvärderades med hjälp av Wechsler Intelligence Scale for

Children IV (WISC-IV) hos 355 barn. Grov- och finmotorisk funktion

utvärderades med hjälp av Movement Assessment Battery for Children 2

(MABC-2) hos 345 barn.

Delarbete IV är baserat på the very Low birth weight Infants - Glucose and

Hormonal profiles over Time (LIGHT) studien, en prospektiv studie som

inkluderade 50 barn med mycket låg födelsevikt (< 1500 g) som föddes i Umeå

under åren 2016-2019. En kontinuerlig blodsockermätare sattes på 35 av barnen

för en 48-timmars period vid en ålder motsvarande 36 graviditetsveckor.

Resultat

Förekomsten av neonatal hyperglykemi hos extremt för tidigt födda barn ökade

under de första två veckorna av livet och sjönk därefter långsamt under

efterföljande veckor. Hyperglykemi förekom hos mer än hälften av barnen med

mycket låg födelsevikt, kontrollerat vid ålder motsvarande 36 graviditetsveckor.

vii

Blodsockernivåerna hos extremt för tidigt födda barn ökade med 1,6% dagen efter

en ökning av glukos (socker i näringslösning) med 1 g/kg/dag. Hos barn med

mycket låg födelsevikt hade pojkar större risk än flickor att få hyperglykemi vid

en ålder motsvarande 36 graviditetsveckor. Även infektion i fostervattnet och

tidigare blodsockerrubbningar under vårdperioden var riskfaktorer för

hyperglykemi.

Neonatal hyperglykemi var förknippat med ökad dödlighet före 28 dagars ålder

hos extremt för tidigt födda barn. Insulinbehandling var förknippad med bättre

överlevnad vid både 28- och 70 dagars ålder. Förekomsten, svårighetsgraden och

durationen av neonatal hyperglykemi visade samband med högre diastoliskt

blodtryck och lägre MABC-2 score vid 6,5-års ålder. Insulinbehandling hade inget

statistiskt samband med dessa utfall.

Högre proteinintag under de första åtta veckorna av livet visade samband med

högre diastoliskt blodtryck vid 6,5-års ålder. Ökning i kolhydratintaget under

denna period var förknippad med ökat systoliskt och diastoliskt blodtryck vid

6,5-års ålder. Tillväxten under samma period var inte statistiskt kopplad till

senare blodtrycksnivå.

Slutsatser

Neonatal hyperglykemi visades vara vanligare hos extremt för tidigt födda barn

än man tidigare trott. Resultaten tyder på att blodsockerrubbningar verkar vara

vanliga även vid en ålder motsvarande 36 graviditetsveckor hos barn med mycket

låg födelsevikt. Tillförsel av glukos (sockerdropp) verkar vara associerad med

enbart ringa påverkan på blodsockernivåerna. Neonatal hyperglykemi var

förknippat med ökad dödlighet före 28 dagars ålder samt med förhöjt blodtryck

och sämre motorisk utveckling vid 6,5-års ålder. Ökat kolhydratintag under de

första åtta veckorna i livet visade statistiskt samband med förhöjt blodtryck vid

6,5-års ålder. Resultaten bidrar med kunskap kring viktiga riskfaktorer för

neonatal hyperglykemi och dess påverkan på hälsoutfall upp till 6,5-års ålder.

Olika behandlingsalternativ för neonatal hyperglykemi bör utvärderas i

djurmodeller samt i randomiserade kontrollerade kliniska studier hos för tidigt

födda barn.

viii

Abbreviations

BMI – Body mass index

BP – Blood pressure

CGM – Continuous glucose monitoring

CPAP – Continuous positive air pressure

DBP – Diastolic blood pressure

ELBW – Extremely low birth weight

EPT – Extremely preterm

EXPRESS – Extremely preterm infants in Sweden study

FSIQ – Full scale intelligence quotient

GLUT – Glucose transporter

HINT – Hyperglycaemia and insulin in neonates trial

IUGR – Intrauterine growth retardation

IVH – Intraventricular hemorrhage

LIGHT – Very low birth weight infants – glucose and hormonal profiles over time

MABC-2 – Movement assessment battery for children 2

NDD – Neurodevelopmental disabilities

NEC – Necrotizing enterocolitis

NICU – Neonatal intensive care unit

NIRTURE – Neonatal insulin replacement therapy in Europe

PMA – Postmenstrual age

ix

RCT – Randomized controlled trial

ROP – Retinopathy of prematurity

SBP – Systolic blood pressure

SGA – Small for gestational age

VLBW – Very low birth weight

VPT – Very preterm

WISC-IV – Wechsler intelligence scale for children IV

x

Original papers

I Zamir I, Tornevi A, Abrahamsson T, Ahlsson F, Engström E,

Hallberg B, Hansen-Pupp I, Stoltz Sjöström E, Domellöf M.

Hyperglycemia in extremely preterm infants – Insulin treatment,

mortality and nutrient intakes. J Pediatr. 2018;200:104-110. DOI:

10.1016/j.jpeds.2018.03.049.

II Zamir I, Stoltz Sjöström E, Edstedt Bonamy AK, Mohlkert LA,

Norman M, Domellöf M. Postnatal nutritional intakes and

hyperglycemia as determinants of blood pressure at 6.5 years of

age in children born extremely preterm. Pediatr Res.

2019;86(1):115-121. DOI: 10.1038/s41390-019-0341-8.

III Zamir I, Stoltz Sjöström E, Ahlsson F, Hansen-Pupp I, Serenius

F, Domellöf M. Hyperglycemia associated with worse motor

outcomes in extremely preterm infants. (Submitted manuscript).

IV Zamir I, Stoltz Sjöström E, van der Bergh J, Naumburg E,

Berhane Y, Domellöf M. Dysglycemia at postmenstrual age 36

weeks in preterm infants born with very low birth weight.

(Manuscript)

Paper I is reprinted with permission from The Journal of Pediatrics, published by

© 2020 Elsevier. Paper II is distributed under the Creative Commons Attribution

(CC-BY) license.

1

Background

Extremely preterm and very low birth weight infants

Approximately 7000 infants are born prematurely (before completed 37

gestational weeks) in Sweden each year, which accounts for about 6% of all births

in the country. Over 100 000 children who currently live in Sweden have been

born prematurely. Very low birth weight (VLBW) infants are those who weigh

< 1500 grams at birth and extremely low birth weight (ELBW) infants are those

who weigh < 1000 grams at birth. Extremely preterm (EPT) infants are those

born < 28 completed weeks of gestation and they represent about 0.3% of all

births (1).

A dramatic improvement in the prognosis for premature infants has been brought

in recent decades by the advances in neonatal intensive care, especially with

regards to EPT infants. In Sweden, infants born before completed 27 gestational

weeks between 2004-2007 had a 1-year survival rate of 70%, while those born

between 2014-2016 had a 1-year survival rate of 77% (2). Still, half of the surviving

EPT infants suffer from neonatal morbidities and 34% of them have moderate to

severe neurodevelopmental disabilities (NDD) at 6.5 years of age (3, 4). Brain

injuries occurring early during the postnatal life (i.e. after birth) are associated

with high rates of chronic conditions and special healthcare needs later in life,

and those children have been shown to have higher risk for motor skills deficits

and poor visual-motor integration at 6.5 years of age (5-7).

Already from birth, these infants have increased energy requirements due to

growth and low content of energy storage tissues. Up to 120 kcal/kg/day might

be needed by these infants (8). This high energy requirement, as well as the

infant’s basal need for maintenance of body tissues, need to be provided by the

nutritional intake of these infants. Early nutritional intake is an independent

predictor of the infant’s growth (9). Since preterm infants do not tolerate

receiving large amounts of nutrition enterally (i.e. via the gastrointestinal tract),

it is therefore recommended to begin parenteral nutrition (i.e. via infusion) for

these infants as soon as possible after birth.

Key points

The survival of preterm infants has increased but these infants have

higher risk for morbidity throughout their childhood.

Preterm infants have low energy stores while their energy needs are

extremely high and should be supplied by parenteral nutrition.

2

Why is glucose so important and how much is needed?

Glucose is the primary energy source in fetal life and in the early neonatal period

(10). It is a crucial energy source for the brain, the renal medulla and the

erythrocytes, and is utilized by the muscles, liver, heart, kidneys and gut. The

ELBW infant utilizes three to four times more glucose per kg body weight than

adults due to their higher brain to body ratio (11). The endogenous production of

glucose (i.e. what the infant can produce from other nutrients and energy stores)

provides only 1/3 of the glucose needed. The stable ELBW infant has been

estimated to require a glucose supply rate of 6 mg/kg/min (8.6 g/kg/d) with an

additional 2-3 mg/kg/min (2.9-4.3 g/kg/d) to support protein anabolism (i.e.

protein building, thereby growth) (10). On the other hand, glucose intakes beyond

12 mg/kg/min (17.3 g/kg/d) would be converted to fat, a process which is not

energy efficient.

EPT infants were shown to experience low glucose concentrations earlier after

birth than term-born infants and thus exogenous glucose (i.e. from an outside

source) should be provided as soon as possible after birth (12). Carbohydrates are

usually supplied in parenteral nutrition in the form of dextrose (D-Glucose). The

latest European Society for Pediatric Gastroenterology, Hepatology and Nutrition

(ESPGHAN) guidelines for parenteral nutrition in children published 2018

recommend a carbohydrate intake of 4-8 mg/kg/min (5.8-11.5 g/kg/d) on day 1

of life and a target intake of 8-10 mg/kg/min (11.5-14.4 g/kg/d) thereafter (13). It

is also recommended to avoid hyperglycemia > 8 mmol/L due to increased

morbidity and mortality, and repetitive blood glucose concentrations > 10

mmol/L should be treated with insulin infusion after first adjusting the glucose

infusion rate.

Definition(s) of Hyperglycemia in preterm infants

Healthy term-born neonates maintain glucose concentrations between

1.3-8.3 mmol/L even when unfed up to six hours (14). However, there is no

definitive definition of neonatal hyperglycemia, especially in preterm infants, and

a wide range of definitions is being used by different neonatal units (15). Some of

the different definitions used in research publications are presented in Figure 1.

Key points

Glucose is an important energy source in preterm infants, especially

for the brain.

Preterm infants cannot produce enough glucose and thus need to

receive it via parenteral nutrition.

3

Figure 1. Different definitions for neonatal hyperglycemia used in the literature. Glucose

concentrations expressed in mmol/L. 1 Pertierra-Cortada et al. (16); 2 Scheurer et al. (17); 3 Sabzehei et al. (18); 4 Alsweiler et al. (19); 5 Meetze et al. (20); 6 Yoon et al. (21); 7 Fernández-Martínez et al. (22); 8 Beardsall et al. (23); 9 Yoo et al. (24); 10 Bermick et al.

(25).

The American Society for Parenteral and Enteral Nutrition has recommended in

its 2012 guidelines to maintain serum glucose concentrations < 8.3 mmol/L in

neonates receiving parenteral nutrition (26). In clinical practice though, plasma

glucose concentrations < 10 mmol/L are many times overlooked, and even higher

glucose concentrations are sometimes not treated as long as the hyperglycemia

episode is not prolonged and can be controlled to some extent by decreasing the

glucose infusion rate.

Prevalence of Hyperglycemia

Different patient populations and definitions used for hyperglycemia have led to

various prevalence rates being reported in the literature. Studies have thus far

focused mainly on hyperglycemia occurring during the first two weeks of life

(Table 1).

≥ 7.81

> 8.32,3

Twice > 8.5

4 hours apart4

> 8.9 lasting 4 hours5

> 106,7

Twice > 108

> 11.19

Persistently high

(usually >

11.1) requiring

insulin infusion10

> 13.35

Key point

There is no clear definition for hyperglycemia in preterm infants.

4

Table 1. Prevalence of hyperglycemia in preterm infants.

Study Type of study N Definition Postnatal period

Prevalence

VLBW infants Scheurer et al. (17)

Prospective 53 > 8.3 Weeks 1-2 45%

Yoon et al. (21)

Prospective 141* ≥ 10 Week 1 30%

Sabzehei et al. (18)

Retrospective 564 > 8.3 Admission period

32%

ELBW infants Hays

et al. (27) Retrospective 82 > 8.3 Week 1** 57%

> 13.9 32% Blanco

et al. (28) Retrospective 169 ≥ 8.3; twice Weeks 1-2 88%

Bermick

et al. (25)

Retrospective

216

> 8.3 Days 0-10

98% > 11.1;

insulin-treated

35%

Yoo

et al. (24)

Retrospective

260

> 11.1 Weeks 1-2

85% ≥ 16.7; insulin-treated

46%

* Preterm infants born < 34 gestational weeks; ** excluding 1st day of life; Glucose

concentrations expressed in mmol/L.

Continuous glucose monitoring (CGM; see Methods) allows monitoring glucose

concentrations continuously during a certain time period, as opposed to relying

on blood samples taken at certain time points without knowing the trend of the

glucose concentrations before or after these specific time points. Results from

studies using CGM are shown in Table 2.

Table 2. Prevalence of hyperglycemia in VLBW infants using prospective CGM-derived

data.

Study N Definition Postnatal period Prevalence

Beardsall et al. (23)

188

> 8

Week 1

80% > 10 57% > 15 23% > 20 9% > 10

10% of the time* 32%

Szymonska et al. (29)

63 > 8.3 Week 1 43%

Fernández-Martínez et al. (22)

60 > 10 Week 1 37%

Iglesias Platas et al. (30)

38 > 10 Weeks 1-2 58%

* Equivalent of 15 hours; Glucose concentrations expressed in mmol/L.

5

In VLBW infants who were clinically stable at postmenstrual age (PMA) 32-33

weeks, CGM-derived data showed that 37% had hyperglycemia > 8.3 mmol/L and

further 29% had combined dysglycemia (both hyperglycemia and hypoglycemia)

(31). Hyperglycemia > 11.1 mmol/L was detected in 32%. Infants who were ELBW

experienced longer and more severe dysglycemia episodes. Hyperglycemia lasted

for more than 30 minutes in 39% of the episodes.

A cohort of 60 very preterm infants (VPT; < 32 gestational weeks) were

monitored for glucose concentrations using CGM during 48 hours around PMA

38 weeks (16). Hyperglycemia > 10 mmol/L was found in 23% and further 13%

had combined dysglycemia. Hyperglycemia lasted for a mean duration of four

hours.

Pathogenesis of hyperglycemia

Normal glucose metabolism

During pregnancy, the mother is the primary supplier of glucose to the fetus via

the placenta (32). The fetal pancreas starts to secrete insulin in response to

glucose after 20 weeks of gestation and insulin secretion increases thereafter

during the third trimester (33). Insulin has an important role in glucose

homeostasis, and is a major anabolic hormone active in growth. It enhances fatty

acid synthesis in the liver and glucose uptake in adipose tissue promoting fat

deposition (34).

After birth, glucose supply from the mother suddenly ceases and blood glucose

concentrations drop. The insulin secretion by the neonate must then adapt to the

new glucose intakes that the neonate is exposed to. This insulin-glucose

stabilization process is crucial in order to be able to supply the brain and other

vital organs with sufficient glucose as their energy source. The stabilization

process occurs through activation of hepatic glycogenolysis (i.e. breaking down

glycogen to glucose) and gluconeogenesis (i.e. producing new glucose) in

response to a reduction in insulin secretion and an increase in counterregulatory

hormones (cortisol, glucagon, growth hormone and catecholamines) (35-37). The

last step in both glycogenolysis and gluconeogenesis is catalyzed by the same

Key points

Hyperglycemia is common in preterm infants during the first two

weeks of life.

Two studies have suggested that hyperglycemia might also be

prevalent later during the admission period.

6

enzyme, glucose-6-phosphatase. The activity of this enzyme is upregulated by the

decrease in insulin levels. In healthy term newborns, this metabolic maturation

process happens over the first few days of life: the activity level of the enzyme is

low before birth, it increases to approximately 10% of the activity levels measured

in adults by term age and rapidly increases up to adult activity levels by three days

of age (38-40). During this time enteral intakes increase but until effective enteral

feeds are established, glycogenolysis and gluconeogenesis are the only sources of

glucose in the neonate.

To be able to utilize glucose, it needs to be transported into cells. This is facilitated

by glucose transporters (GLUTs), “channels” transporting glucose into cells.

GLUT-1 is expressed especially in the blood-brain barrier and on erythrocytes

(41). GLUT-2 is found primarily on pancreatic β-cells, liver and kidney. GLUT-3

is found in the brain. GLUT-4 is expressed in heart, skeletal muscle and adipose

tissues, and responds to insulin. This means that while insulin facilitates the

influx of glucose into muscle and fat cells via GLUT-4, it has no role in promoting

the uptake of glucose in the brain. During fetal and early neonatal life, GLUT-1 is

the predominant type of glucose transporter, enabling available glucose to be

taken up by the brain and erythrocytes, and is replaced gradually by the other

tranporter types as the infant matures (42).

Insulin resistance

Pollak et al. postulated that hyperglycemia might be caused due to insufficient

effect of insulin on the peripheral tissues (muscle, fat) (43). In a preterm baboon

model, Blanco et al. have shown that reduced peripheral insulin sensitivity was

present in the first weeks of life, with reduced response to insulin along its

signaling pathway and lower content of GLUT-4, which led to impaired glucose

uptake (44). Insulin resistance might also be induced by increased levels of

proinflammatory cytokines (Tumor necrosis factor-α, Interleukin-1,

Interleukin-6) due to, for example, sepsis and necrotizing enterocolitis (NEC)

(45). Some concerns were raised that high amino acid intake, by enhancing

glucose production, might exacerbate insulin resistance (46). Furthermore,

premature infants were shown to have lower insulin sensitivity even later in life,

at 4 to 10 years of age (47). Table 3 summarizes the evidence supporting the

theory that insulin resistance is common in preterm infants.

On the other hand, Meetze et al. found that baseline serum insulin levels before

onset of hyperglycemia did not correlate with later development of hyperglycemia

(20). Furthermore, insulin levels increased in most infants in response to

hyperglycemia which argues for an adequate pancreatic response to changes in

glucose concentrations. VPT neonates born small for gestational age (SGA) were

found to have lower insulin and C-peptide levels than appropriate for gestational

7

Table 3. Evidence for insulin resistance in preterm infants.

Study N Population Finding Hawdon

et al. (48) 16 Preterm (< 34 weeks) ↑ insulin:glucose ratio

Salis et al. (49)

n.a. VPT (≤ 32 weeks) ↑ insulin, C-peptide

Ahmad et al. (50)

19 VPT (< 32 weeks) ↑ insulin, HOMA-IR

Bagnoli et al. (51)

16 VPT (≤ 30 weeks) ↑ glucagon, insulin, HOMA-IR

Goldman et al. (52)

4 VLBW ↓ response to insulin

Rao et al. (53)

16 VLBW ↓ glucose tolerance during weeks 1-2 of life

Beardsall et al. (54)

283 VLBW - insulin treatment and dose did not affect IGF-1 levels - ↓ IGF-1 in infants with hyperglycemia

Cekmez et al. (55)

30 ELBW ↑ insulin, HOMA-IR, visfatin

HOMA-IR – Homeostatis Model Assessment for Insulin Resistance; IGF-1 – Insulin-like

Growth Factor 1

age preterm neonates, and their quantitative insulin check index (a measure of

insulin sensitivity) was higher, indicating higher insulin sensitivity (56).

Relative insulin deficiency – β-cell dysfunction

Bansal et al. have shown in a model of preterm lambs that preterm birth reduced

β-cell mass both at 4 weeks after term age and at 12 months (equivalent to

adulthood) (57). This was associated with a decrease in insulin secretion at four

months (equivalent to juvenile age) and reduced insulin mRNA expression in

adulthood. On top of that, hyperglycemia further downregulated pancreatic gene

expression in adulthood, suggesting that – if transferred to human

pathophysiology - preterm infants might have a reduced β-cell mass which can

lead to diabetes at a later age, and hyperglycemia might exacerbate this process.

Low endogenous C-peptide production (a marker for insulin production) was

found to be prevalent in critically ill children with both respiratory and

cardiovascular failure who developed hyperglycemia (58). In VPT neonates, boys

were shown to have lower insulin secretion than girls, and insulin levels did not

differ between hyperglycemic and normoglycemic neonates (59, 60). Proinsulin

(a less active protein that cannot effectively regulate glucose concentrations and

needs to be processed to insulin first in order to exert its regulatory effect) levels

in preterm neonates were higher than in term-born neonates, with proinsulin

levels falling in hyperglycemic neonates after insulin infusion (48, 60). These

8

findings led to the conclusion that processing of proinsulin in β-cells is partially

defective in hyperglycemic VPT infants.

Relative insulin deficiency – deficient incretin response

Incretins are hormones secreted by cells in the gastrointestinal system and

include glucagon-like peptide 1 and glucose-dependent insulinotropic

polypeptide. In adults, incretins augment insulin secretion in response to meals.

Normal establishment of enteral feeds is often delayed in preterm infants and

thus normal stimulation of incretins by nutrition does not occur. This might

contribute to the relative insulin deficiency in these infants as Salis et al. found

that enteral feeds in premature neonates result in an increase in glucagon-like

peptide 1 levels (49). Yet even when fed enterally, Lucas et al. found that preterm

neonates do not present an adequate incretin response compared to term-born

neonates (39). The delayed introduction of enteral feeds in these infants might

hinder the maturation of insulin secretion in response to blood glucose, as it has

been shown in neonatal rat models (61). Thus, when enteral feeds are finally

started, the response of increasing glucagon-like peptide 1 levels might not be

enough to regulate the blood glucose since insulin is not secreted in sufficient

amounts.

Insufficient control of glucose production

It is postulated that hyperglycemia might be caused by persisting hepatic glucose

production, supported by evidence in infants with severe respiratory disease

and/or who were exposed to severe bacterial infection (43, 62). Farrag et al. found

that neonates have persistent glucose production (63). It was shown that VPT

neonates had incompletely suppressed glucose production following glucose

infusion, and insulin levels affected the regulation of plasma glucose more than

plasma glucose level itself (64, 65). Chacko et al. investigated VPT neonates who

received total parenteral nutrition providing glucose at rates exceeding normal

infant glucose turnover rates, i.e. infants who were supposed to react to the high

intake of glucose by lowering the rate of gluconeogenesis. It was shown that

gluconeogenesis continued to produce glucose in a rate of 1.3 mg/kg/min

(1.9 g/kg/d) and this rate was not affected by the total glucose infusion rate and

glucose concentration (66). A persistent endogenous glucose production in spite

of glucose infusion was found in low birth weight infants at age 2-5 weeks as well

(67). It was also found that gluconeogenesis in ELBW neonates remained

unchanged despite a large reduction in the glucose infusion rate (instead of

increasing, as expected) (68). This might support the findings by Hume et al.,

where preterm infants had lower activity of glucose-6-phosphatase, thus

implying lower activity of gluconeogenesis (69).

9

Other mechanisms

Cortisol levels are increased during the first days of life in preterm infants, and

the lower the gestational age is at birth - the higher the cortisol levels (70, 71).

High cortisol levels might lead to hyperglycemia by stimulating gluconeogenesis

and by decreasing glucose uptake by muscle and fat tissue cells.

The pathophysiological effects of hyperglycemia

Hyperglycemia leads to increased uptake of glucose by cells. Glucose overload in

the cell increases the generation of oxygen free radicals (“glucotoxicity”), which

in turn might cause mitochondrial dysfunction and increased apoptosis (i.e. cell

death) (33). Other effects of hyperglycemia include reduced neutrophil activity,

decreased complement function and a change in the balance of proinflammatory

cytokines, which might lead to an increased susceptibility for infections (72).

Risk factors for hyperglycemia

Prematurity, low birth weight and growth retardation

It has been shown in animal models that intrauterine growth retardation (IUGR)

leads to impairment of fetal pancreatic function and altered glucose-stimulated

insulin secretion (73). Conflicting results have been published regarding

gestational age (as a measure for prematurity) and relative size at birth

(represented by birth weight z-scores) as risk factors for neonatal hyperglycemia

(Table 4).

Asphyxia

Asphyxia is the term used for a prolonged oxygen deprivation in the newborn. A

marker for asphyxia is a persistently low Apgar score after birth. In VLBW infants,

Apgar score < 6 at 5 minutes after birth was associated with more than four times

higher risk for neonatal hyperglycemia (in unadjusted models) (18). In ELBW

infants, low Apgar scores at 1 minute were associated with hyperglycemia (20,

24).

Key points

Conflicting evidence exists regarding the mechanisms causing

neonatal hyperglycemia.

Scientific evidence points to a combination of insulin resistance,

relative insulin deficiency (β-cell dysfunction, deficient incretin

response) and incomplete control of glucose production.

10

Table 4. Prematurity, low birth weight and growth retardation as risk factors for neonatal

hyperglycemia in preterm infants.

Study N Population Risk factor for hyperglycemia

Pertierra-Cortada et al. (16)

60 VPT IUGR

Yoon et al. (21) 141 Preterm (< 34 weeks) SGA

Iglesias Platas et al. (30) 38 VLBW ELBW, EPT

Beardsall et al. (23) 188 VLBW ↓ GA, ↓ BW z-score

Scheurer et al. (17) 53 VLBW ↓ GA, ↓ BW, ↓ BL, ↓ BHC, ↔ z-scores

Szymonska et al. (29) 63 VLBW ↓ GA

Sabzehei et al. (18) 564 VLBW ELBW, ↔ EPT

Fernández-Martínez et al. (22)

60 VLBW ↓ GA, ↓ BW, ↔ IUGR

Blanco et al. (28) 169 ELBW ↓ GA, ↔ BW

Yoo et al. (24) 260 ELBW ↓ GA, ↓ BW, ↔ SGA

Stensvold et al. (74) 195 ELBW ↓ GA, ↔ BW z-score

Meetze et al. (20) 56 ELBW ↓ BW

GA – gestational age; BW – birth weight; BL – birth length; BHC – birth head

circumference.

Respiratory distress syndrome and ventilator support

VLBW infants with hyperglycemia during the first week of life were more likely

to have received ventilator support (22). Similarly, respiratory distress syndrome

and ventilator support in ELBW infants during the first two weeks of life were

associated with 11.2 and 3.6 times higher risk for hyperglycemia > 16.7 mmol/L,

respectively (24).

Parenteral glucose intake

Hawdon et al. showed that glucose production rate in neonates was not affected

by insulin or glucagon levels and thus was a relatively constant factor determining

blood glucose concentration, but the change in blood glucose concentration was

rather associated with the rate of glucose infusion (75). The conclusion was that

blood glucose concentrations are affected mostly by the exogenous glucose given

to the infant.

A retrospective study by Tottman et al. compared 190 VLBW infants born before

and 267 infants born after a new parenteral nutrition protocol (more protein, less

fat and carbohydrates, but also less energy) was implemented (76). Infants in the

“after” group had lower glucose concentrations and experienced less

hyperglycemia > 8.5 mmol/L both during the first week of life and during the first

11

four weeks of life (First week: 27% versus 42%; first four weeks: 39% versus 51%).

Insulin usage rates did not differ though. Increased carbohydrate intake during

the first week of life, but not during the entire period of the first four weeks of life,

was associated with increased risk for hyperglycemia (25% increase in risk for

each 1 g/kg/d increase in the intake).

It was shown that VLBW neonates receiving parenteral glucose infusion (either

as a supplement to oral feedings or as the only source of nutrition) had more

hyperglycemia than exclusively oral-fed neonates, and hyperglycemia was related

to glucose infusion rates > 6.7 mg/kg/min (9.6 g/kg/d) (77, 78). Another study in

VLBW neonates showed that infants who received a glucose infusion rate of

8 mg/kg/min (11.5 g/kg/d) did not have hyperglycemia, while infants who

received glucose infusion rates of 11 and 14 mg/kg/min (15.8 and 20.2 g/kg/d)

did develop hyperglycemia (79). In ELBW infants, a maximum daily glucose

infusion rate of > 5.1 mg/kg/min (7.3 g/kg/d) (but not the carbohydrate intake)

was an independent predictor of hyperglycemia > 12 mmol/L during the first

week of life (74).

On the other hand, other studies have not been able to detect a significant

association between glucose infusion rate and neonatal hyperglycemia. In the

Neonatal Insulin Replacement Therapy in Europe (NIRTURE) study, a

randomized controlled trial (RCT) in VLBW infants that used CGM registration

during the first week of life, no association was found between mean glucose

infusion rate during the first week of life and the risk for hyperglycemia (23). A

large retrospective study as well as a recently published prospective study using

CGM in VLBW infants have shown similar results (18, 22).

Parenteral lipid intake

It was shown that lipid infusion contributes to hyperglycemia, possibly by

stimulating gluconeogenesis due to provision of more available substrate for the

process (80, 81). In the NIRTURE study, parenteral lipid infusion was shown to

be an independent risk factor for hyperglycemia, perhaps due to the lipid infusion

impairing insulin sensitivity (23). Table 5 presents conflicting results from

observational studies regarding change in parenteral lipid intake and its effect on

the risk of neonatal hyperglycemia.

Steroid treatment

Antenatal (i.e. pre-birth) synthetic corticosteroid treatment given to mothers was

not associated with hyperglycemia during the first three days of life, but long

exposure (> one week) for antenatal steroids was associated with a four times

higher risk for hyperglycemia (82). The use of systemic steroids (hydrocortisone

12

Table 5. Parenteral lipid intake and the risk of neonatal hyperglycemia in preterm infants.

Study Population (N)

Change in parenteral lipid intake

Hyperglycemia definition

Change in odds for

hyperglycemia Tottman et al. (76)

VLBW (457)

↑ 1 g/kg/d

> 8.5 mmol/L ↓ 44%

Stensvold et al. (74)

ELBW (195)

> 12 mmol/L ↑ 41%

and dexamethasone) in VLBW infants was shown to be associated with an

increased risk for hyperglycemia > 10 mmol/L (83). Cochrane meta-analyses

have shown that postnatal systemic steroid treatment given to preterm infants at

any time was associated with hyperglycemia (84, 85). On the other hand, a

retrospective study in ELBW neonates did not detect an association between

postnatal steroid use and hyperglycemia (28). Furthermore, Cochrane

meta-analyses concluded that ventilated VLBW infants treated with inhaled

corticosteroids during the first week of life for preventing bronchopulmonary

dysplasia had a 48% lower risk of hyperglycemia, but no difference was noted

when the treatment was given later than the first week of life (86, 87).

Inotrope treatment

The use of inotropes in the NIRTURE study was associated with a 3.4 times

higher risk for hyperglycemia (23). A recent prospective study in VLBW infants

has confirmed this association, reporting that the need for inotropic drugs

increased the risk of hyperglycemia by more than 4 times (22). This association

might be caused by inotropes increasing insulin resistance as well as reducing

insulin secretion.

Sepsis

Sepsis was found to be an independent risk factor for hyperglycemia in the

NIRTURE study (23). Another study showed that early-onset sepsis was

associated with hyperglycemia in VLBW neonates (29). Antibiotics use during the

first two weeks of life was found to be associated with ten times higher odds for

hyperglycemia in ELBW infants (24). However, a prospective study in VLBW

neonates and a retrospective study in ELBW neonates did not observe similar

results (22, 28).

Stress

Both disease processes and the medical treatment provided (blood sampling,

intubation, infusion etc.) increase physiologic stress in the sick neonate (88).

Stress leads to a counterregulatory hormone reaction, which in turn leads to

13

decreased insulin secretion from the pancreas and increased glycogenolysis and

gluconeogenesis, thereby to hyperglycemia.

Hypophosphatemia

An observational study including 148 ELBW infants found that the hazard risk of

hyperglycemia > 8.3 mmol/L at any given time was 3 times higher for each 0.41

mmol/L decrease of phosphate level at that time and 3.85 times higher for each

0.41 mmol/L decrease of phosphate level on the previous day (89). The authors

suggested that hypophosphatemia might thus be associated with hyperglycemia.

Neonatal diabetes mellitus

Type 1 (insulin-dependent) diabetes mellitus presenting in neonates is rare and

autoimmune antibodies are not frequently detected in infants diagnosed with

diabetes before 6 months of age (90). Thus, most of the patients with neonatal

diabetes mellitus will have a monogenic form of diabetes. Early insulin

administration is crucial in treating this group of patients. Approximately 50% of

these patients will have permanent diabetes requiring life-long treatment with

insulin, while in the remaining patients, the diabetes will be transient and remit

within a few weeks or months.

Outcomes of hyperglycemia

Mortality

In VLBW infants, neonatal hyperglycemia and its duration were found to be

associated with a higher mortality risk (18, 22, 29, 30). Hays et al. showed that

increased blood glucose concentration in ELBW infants was associated with an

increased mortality risk (27). Large observational studies in ELBW infants

demonstrated that hyperglycemia > 10 mmol/L during the first week of life and

> 12 mmol/L lasting for at least three hours were risk factors for death (91, 92).

On the other hand, some retrospective studies in ELBW infants did not observe

such an association between hyperglycemia and mortality, after adjusting for

gestational age and major morbidities (24, 28).

Key points

Many risk factors have been proposed for neonatal hyperglycemia,

including prematurity, growth retardation, parenteral nutrition,

perinatal morbidities and medical treatment.

Conflicting results have been shown for most of these risk factors.

14

A small study including 33 VPT infants found that mortality during admission

was higher among the subgroup of hyperglycemic EPT infants (93). EPT infants

who had blood glucose concentrations > 8.3 mmol/L on the first day of life and

those who had blood glucose concentrations ≥ 8.3 mmol/L four or more times

during the first week of life were shown to have had significantly increased

mortality rates (94, 95).

Retinopathy of prematurity

Hyperglycemia in VLBW infants, its episode frequency, mean blood glucose

concentration in the first week of life, and increased insulin use were all shown to

be independent risk factors for the development of retinopathy of prematurity

(ROP) (18, 30, 96-98). Similarly, ROP patients were found to have previously had

higher average glucose concentrations and higher incidence of hyperglycemia

> 8.3 mmol/L than non-ROP patients (99).

Retrospective studies in ELBW infants have shown similar results, with

hyperglycemia being an independent risk factor for ROP (28, 100). A

retrospective cohort study including 114 ELBW infants found that ROP patients

had higher time-weighted glucose concentrations during the first 10 and 30 days

of life (101). Longer periods of time having blood glucose concentrations > 6.6

mmol/L were associated with severe ROP, rather than isolated exposures to

glucose concentrations > 8.3 mmol/L.

A possible explanation for the mechanism responsible for these findings might be

that hyperglycemia induces inflammation and apoptosis in the retina which leads

to neuronal degeneration (102).

Growth

In VLBW infants, hyperglycemia > 8.3 mmol/L during the first two weeks of life

was associated with lower weight, shorter length and head circumference, and

lower percent body fat at four months corrected age, as well as with poor growth

up to two years of age (17, 103). On the other hand, van der Lugt et al. did not find

an association between hyperglycemia and growth at two years of age (93).

Intraventricular hemorrhage

It was previously shown that preterm infants who have hyperglycemia have an

increased risk for intraventricular hemorrhage (IVH) (18, 27). Auerbach et al.

reported that longer duration of hyperglycemia > 6.9 mmol/L during the first four

days of life was associated with higher incidence of severe IVH (104).

15

The presence of insulin-treated hyperglycemia during the first 10 days of life in

ELBW infants was associated with a 2.6 times higher risk for IVH (25). It is worth

noting though that hyperglycemia in itself was not an independent predictor of

IVH, rather only in the simultaneous presence of hypernatremia. Also worth

mentioning is that it was not established whether IVH preceeded or followed the

development of hyperglycemia. Other studies did not find an association between

IVH and hyperglycemia (22, 28, 95).

Brain function and neurodevelopmental disability

Hyperglycemic rats were found to have lower brain weight and more severe brain

damage, especially in the hippocampus (105). In these animal models, neonatal

hyperglycemia was found to induce oxidative stress and damage, apoptosis, and

increase in inflammatory cytokines, resulting in microglial activation and

astrocytosis, which led to long-term changes in synaptogenesis and behaviour

(106). Higher glucose concentrations were associated with higher methylation of

leucine-rich alpha-2-glycoprotein 1, a protein known to be associated with

neurodevelopment (107). In neonatal rats, hyperglycemia was also associated

with upregulation of Poly (ADP-ribose) polymerase-1 (a nuclear enzyme involved

in DNA repair and cell survival) and NF-kB expression (which regulates

microglial activation), which led to microgliosis in the brain (108).

A Swedish study including 94 EPT infants who underwent brain magnetic

resonance imaging scan at term age found that hyperglycemia > 8.3 mmol/L on

the first day of life was an independent risk factor for white matter reduction (95).

A prospective study including 41 infants born < 31 gestational weeks found that

increased blood glucose concentrations were associated with decreased total

absolute band power, a measure expressing background brain activity (109).

van der Lugt et al. found more abnormal neurological and behavioural

development in VLBW infants with insulin-treated hyperglycemia (93).

Hyperglycemia > 8.3 mmol/L was shown to be associated with lower survival

without NDD at 18 months as well as with poorer gross motor outcomes at 4 years

of age in term-born children who had moderate-severe hypoxic ischemic

encephalopathy or asphyxia (110, 111). A retrospective study including 443 VLBW

infants showed that hyperglycemia during the first week of life was associated

with lower survival without NDD at two years of age, but this association did not

remain significant after adjusting for gestational age, birth weight z-score and

socioeconomic status (112).

16

Sepsis

Kao et al. showed that ELBW infants who had hyperglycemia > 10 mmol/L during

the first week of life had higher risk for dying or developing a culture-proven

infection during the same period (91). On the other hand, van der Lugt et al. found

that sepsis was more prevalent in VLBW infants with birth weight > 1000 grams

who had hyperglycemia, but not in ELBW infants (93). In line with these findings,

Alexandrou et al. did not find an association between hyperglycemia and sepsis

in EPT infants (95).

Necrotizing enterocolitis

One study by Kao et al. found that hyperglycemia > 10 mmol/L during the first

week of life in ELBW infants increases the risk for developing stage II/III NEC,

but this finding was not repeated by others (91).

Blood pressure

In adults with hypertension, impaired glucose tolerance was associated with

increased arterial stiffness (113). Sustained hyperglycemia was shown to activate

the renin-angiotensin system in adults with type 1 (insulin-dependent) diabetes

mellitus, thereby increasing systemic and renal vasomotor tonus, increasing the

blood pressure (BP) (114, 115). Rat models showed that hyperglycemia exerts its

blood pressure-elevating effect in two phases: an early phase due to an efferent

sympathetic discharge, and a delayed phase due to activation of the

renin-angiontensin system (116). Hyperglycemia in preterm baboons was shown

to accelerate kindey maturation and result increased oxidative stress, which later

might lead to increased BP (117). No study has investigated the association

between hyperglycemia and later BP in preterm infants.

Key points

Neonatal hyperglycemia is associated with many adverse outcomes,

including increased mortality and higher risk of ROP, diminished

growth, IVH, white matter reduction, lower brain activity, sepsis and

NEC, though conflicting results have been shown.

Very little or no evidence exists regarding the possible associations

between neonatal hyperglycemia and NDD and BP later in life.

17

Insu

lin r

esis

tance

Rel

ativ

e

insu

lin d

efic

iency

Insu

ffic

ien

t

contr

ol

of

glu

cose

pro

duct

ion

Hy

per

gly

cem

ia

Mo

rtal

ity

Ret

ino

pat

hy o

f pre

mat

uri

ty

Dim

inis

hed

gro

wth

Intr

aven

tric

ula

r h

emo

rrhag

e

Whit

e m

atte

r re

du

ctio

n

Low

er b

rain

act

ivit

y

Sep

sis

Nec

roti

zin

g e

nte

roco

liti

s

Neu

rodev

elo

pm

enta

l dis

abil

ity

Pre

mat

uri

ty

Lo

w b

irth

wei

gh

t

Gro

wth

ret

ard

atio

n

Asp

hyxia

Sep

sis

Res

pir

ato

ry d

istr

ess

syn

dro

me

Mec

han

ical

ven

tila

tion

Par

ente

ral

glu

cose

an

d l

ipid

in

takes

Ste

roid

tre

atm

ent

Ino

tro

pe

trea

tmen

t

Str

ess

Fig

ur

e 2

. Neo

na

tal h

yp

erg

lyce

mia

– m

ech

an

ism

s, r

isk

fa

cto

rs a

nd

co

mp

lica

tio

ns.

Ris

k f

act

ors

:

18

Management of hyperglycemia

Different neonatal units and different clinicians have different approaches for

treatment of neonatal hyperglycemia in preterm infants. The majority would

decrease parenteral glucose intake as the initial measure, and only if this would

prove to be insufficient, start insulin treatment (15). Insulin treatment instead of

glucose intake reduction, on the other hand, could help optimize early energy

intakes, which have been shown to be important for postnatal growth (9, 118).

The starting dose of insulin differs greatly between different units, with the most

common starting dose being 0.05 IU/kg/h. Current evidence for the benefits of

this treatment is sparse and contradicting (see below), and a cochrane review has

concluded that there is insufficient evidence published to determine whether

treatments for neonatal hyperglycemia have any effect (119).

Reduction of glucose infusion rate

The American Society for Parenteral and Enteral Nutrition recommended to

avoid excess energy and glucose delivery in hyperglycemic infants receiving

parenteral nutrition (26). However, no trials have been done comparing

reduction of parenteral glucose infusion with no reduction as treatment for

hyperglycemia. A cochrane review has thus concluded that there is insufficient

evidence from trials in VLBW infants to determine whether lower or higher

glucose infusion rates affect mortality and morbidity (120).

In a study including 195 ELBW infants, Stensvold et al. compared infants born

before and after a strict strategy aiming at avoiding fluctuations in glucose

infusion rates was implemented at the University hospital in Oslo, Norway (74).

In the “after” group, glucose infusion rate was reduced to a minimum of

4 mg/kg/min (5.8 g/kg/d) if blood glucose concentrations approached

10-12 mmol/L. The maximum glucose infusion rate decreased from

6.3 mg/kg/min (9.1 g/kg/d) to 5.8 mg/kg/min (8.4 g/kg/d), severe

hyperglycemia (two consecutive glucose values > 12 mmol/L at least three hours

apart) during the first week of life decreased from 48% to 23%, insulin use

decreased from 39% to 16% and mortality decreased from 26% to 10%.

Hyperglycemia > 10 mmol/L decreased as well, from 68% to 35%. Total protein

and energy intakes were higher in the “after” group.

Adjusting parenteral amino acid and lipid intakes

Some amino acids (e.g. lysine, threonine and branched chain amino acids)

stimulate insulin secretion. This was shown for arginine as well, whereas low

plasma arginine levels were associated with poorer blood glucose control (121). A

small study comparing preterm neonates receiving only glucose infusion with

neonates receiving both glucose and amino acid infusion showed that neonates

19

receiving amino acid infusion had higher insulin levels, thereby suggesting that

amino acid infusion has an insulinogenic glucose-decreasing effect (122). Studies

investigating the association between a parenteral amino acid intake intervention

and hyperglycemia are summarized in Table 6. A cochrane meta-analysis

concluded that higher amino acid intake in parenteral nutrition was associated

with a reduction in hyperglycemia > 8.3 mmol/L, but without reduction in the

incidence of hyperglycemia treated with insulin (123).

Table 6. Associations between parenteral amino acid intake interventions and

hyperglycemia in preterm infants.

Study Type of study

Population (N)

Intervention Finding in intervention group

Thureen et al. (124)

RCT ≤ 1300 g (28)

3 g/kg/d (versus 1 g/kg/d)

↑ insulin levels

Burattini et al. (125)

RCT < 1250 g (131)

4 g/kg/d (versus 2.5 g/kg/d)

↓ hyperglycemia > 9.7 mmol/L

Tottman et al. (76)

Before-after

VLBW (457)

↑ 1 g/kg/d on weeks 1-4

↓ hyperglycemia > 8.5 mmol/L

Mahaveer et al. (126)

Before-after

< 29 weeks (76)

↑ 3.5-4 g/kg on week 1

↓ insulin treatment

A study in 29 VLBW infants found that introducing parenteral lipid infusion at

day one versus day eight of life was associated with decreased blood glucose

concentrations (127).

Insulin treatment

In VLBW infants

A number of small case series studies investigated continuous insulin infusion as

treatment for neonatal hyperglycemia in VLBW infants and showed positive

results (Table 7). A retrospective study in hyperglycemic VLBW infants supported

these results (128).

Two RCTs were carried out in VLBW infant populations. The first included

42 infants and compared a more aggressive nutrition regimen (higher protein and

non-protein intakes and insulin treatment for hyperglycemia) with a control

group (lower intakes and glucose infusion rate reduction as treatment for

hyperglycemia). Only a temporary benefit of improved weight gain in the first

week of life was found. Furthermore, higher incidence of hyperglycemia was

found in the intervention group (129). The other RCT, the Hyperglycaemia and

Insulin in Neonates Trial (HINT), included 88 infants with hyperglycemia (> 8.5

mmol/L twice at least four hours apart). A tight glycemic control group

20

Table 7. Continuous insulin infusion as treatment for neonatal hyperglycemia in VLBW

infants.

Study Indication Findings Vaucher

et al. (130)

(N = 10)

Hyperglycemia > 7 mmol/L +

GIR < 4.2 mg/kg/min (age 0-3d)

or < 7.5 mg/kg/min (age ≥ 4d)

- GIR: 5.8 mg/kg/min → 11.2 mg/kg/min - Energy intake: 29 kcal/kg/d → 56 kcal/kg/d - Net weight gain in 80% of infants - Hypoglycemia < 1.4 mmol/L: < 1% of samples

Ostertag et al. (131)

(N = 10)

Hyperglycemia or

GIR < 6 mg/kg/min

- GIR: up to 11.2 mg/kg/min - Energy intake: 50 kcal/kg/d → 70 kcal/kg/d - Weight change: -23 g/d → +13 g/d - Hypoglycemia ≤ 1.4 mmol/L: none

Heron et al.

(132)*

(N = 15)

Hyperglycemia - GIR: 7 mg/kg/min → 9.2 mg/kg/min - Energy intake: 61 kcal/kg/d → 80 kcal/kg/d - Hypoglycemia < 2 mmol/L: 2.8% of samples

* Birth weight < 1250 g; GIR – glucose infusion rate.

immediately receieved insulin infusion at 0.05 IU/kg/h with a target blood

glucose of 4-6 mmol/L. A standard practice group was treated with insulin at the

same infusion rate but only if blood glucose was > 10 mmol/L or if the infant had

persistent glycosuria > 2+, energy intake < 100 kcal/kg/d, was older than

72 hours of age and was not acutely stressed; target blood glucose was

8-10 mmol/L. The study showed that tight glycemic control resulted in greater

weight gain and head circumference growth up to PMA 36 weeks, but linear

growth (leg growth rate) during the same period was compromised and more

hypoglycemia events were noted (19). Furthermore, no effect was found on

survival without NDD, intelligence scores or motor skills at seven years of age

(133). However, treated infants who had their blood glucose concentrations in the

range 4-6 mmol/L had a higher survival rate and were more likely to survive

without NDD and low full scale intelligence quotient (FSIQ).

In ELBW infants

A prospective cohort study found that ELBW infants with hyperglycemia (> 12

mmol/L twice at least four hours apart; N = 88) required longer duration and

higher infusion rates of insulin to maintain normoglycemia, compared with

hyperglycemic infants with birth weight > 1000 g (N = 27) (134).

An RCT in 24 ELBW infants with hyperglycemia (> 9.9 mmol/L with glycosuria)

assigned the infants to receive either total parenteral nutrition and insulin

infusion or a reduction of glucose infusion rate (135). Insulin-treated infants

21

tolerated higher glucose infusion rates (20.1 mg/kg/min versus 13.2 mg/kg/min),

had greater non-protein energy intake (125 kcal/kg/d versus 86 kcal/kg/d) and

greater weight gain (20 g/kg/d versus 8 g/kg/d), without increasing the rate of

hypoglycemia (0.2%).

An RCT by Meetze et al. enrolled 56 ELBW infants (20). Hyperglycemic infants

(> 13.3 mmol/L or > 8.9 mmol/L lasting at least four hours) received either

insulin infusion or a reduction of glucose intake while normoglycemic infants

served as a control group. Insulin infusion was started at a rate of 0.1 IU/kg/h.

Glucose infusion rate was reduced until blood glucose decreased or until a

minimum of 5 mg/kg/min (7.2 g/kg/d) was reached (in which case insulin

infusion was started). Glucose target range was 5-8.9 mmol/L. None of the

infants receiving insulin had hypoglycemia < 3.3 mmol/L. Lower energy intake

was found in the glucose reduction group and these infants remained in a

catabolic state (energy intake < 60 kcal/kg/d) longer than the control group

(8.6 days versus 4.1 days). On the other hand, insulin-treated infants received low

intakes as well (mostly due to delay in enteral feeds), but they were in a catabolic

state for a shorter period than infants in the glucose reduction group (5.5 days).

The authors concluded that while glucose reduction is an efficient treatment for

hyperglycemia, it affects the nutrition intakes markedly, while insulin treatment

is also an efficient treatment without an increased risk for hypoglycemia.

On the other hand, in a retrospective study by Binder et al., 34 ELBW infants with

hyperglycemia who were treated with insulin infusion were compared to

42 hyperglycemic infants who did not receive insulin treatment (136). No

difference was found between the groups regarding non-protein energy intake,

time to energy intake of 100 kcal/kg/d, time to regain birth weight and change in

weight during the first two months of life. Hypoglycemia < 2.2 mmol/L was found

in < 0.5% of the blood samples in insulin-treated infants.

Preventive insulin therapy

Only one RCT to date, the NIRTURE study, has looked into insulin infusion as a

preventive therapy for hyperglycemia. This study was preceded by a pilot study

that included 17 VLBW infants within 24 hours from birth and randomized them

to receive either early elective fixed-dose insulin during the first week of life or

standard care (137). All infants were monitored by CGM that did not affect the

clinical decision making. In the standard care group, hyperglycemia (two

measurements > 10 mmol/L) was managed by switching to a less concentrated

glucose infusion (5%) if glucose infusion rate was > 5 mg/kg/min (7.2 g/kg/d).

Otherwise, an insulin infusion was started at 0.05 IU/kg/h and titrated according

to a sliding scale. Infants in the intervention group received a fixed-dose

(0.025 IU/kg/h) insulin infusion within 24 hours of birth (and if blood glucose

was > 3.5 mmol/L). A 20% glucose infusion was given as long as glucose

22

concentrations were < 4 mmol/L. Insulin infusion was stopped if glucose

concentrations were < 2.6 mmol/L, followed by increasing the glucose infusion.

Hyperglycemia was treated as in the standard care group. Target glucose range

was 2.6-10 mmol/L. Both groups received the same energy intake. The

intervention group had better glucose control and less time was spent in

hyperglycemia (7.6% versus 35.9%). No difference was found in hypoglycemia

rates (0.2% versus 0.4%).

The NIRTURE study itself used the same design as described above, apart from a

higher insulin infusion rate of 0.05 IU/kg/h in the intervention group (138).

Infants in the intervention group had less hyperglycemia, received a higher

carbohydrate intake and lost less weight during the first week of life. However,

hypoglycemia was more common in the intervention arm (29% versus 17%). It

was more prevalent among infants who had birth weight > 1000 g in the

intervention arm, but not among ELBW infants (34% versus 12% in non-ELBW

infants, 26% versus 23% in ELBW infants). No differences were found in

morbidity and mortality at term age or in growth at 28 days of life, but the protein

intake in the study was substantially low (1.23 g/kg/d). The recruiting of patients

was suspended due to concerns for excess periventricular leukomalacia. No

significant difference was found regarding this condition but it was found that

28-day mortality was higher in the intervention group (12% versus 6%). Data

regarding neurodevelopmental outcomes has not been published yet.

The American Society for Parenteral and Enteral Nutrition recommended

therefore against the use of early insulin therapy to prevent hyperglycemia and

recommended the use of insulin infusion only for those patients with persistent

hyperglycemia where reduction of glucose infusion rates, elimination of

medications predisposing to hyperglycemia and correction of underlying causes

of hyperglycemia have failed to correct the hyperglycemia (26). Ogilvy-Stuart et

al. suggested that neonatal hyperglycemia should be treated by maintaining

glucose infusion rates between 6-12 mg/kg/min and glucose concentrations < 10

mmol/L, if necessary, with the use of insulin (45).

Problems with insulin treatment

As mentioned above, rates of hypoglycemia associated with insulin treatment

were quite low in smaller studies (20, 130-132, 135, 136). In a prospective study

using CGM during the first weeks of life in 38 VLBW infants, no relationship

between insulin treatment and hypoglycemia could be found (30). A retrospective

study in ELBW infants found a 4% prevalence of hypoglycemia < 2.2 mmol/L in

hyperglycemic infants treated with insulin (24). Another retrospective study in

infants born < 29 gestational weeks found hypoglycemia ≤ 2 mmol/L in 1.3% of

the samples (12% of the infants) but no association to survival or

23

neurodevelopmental outcomes at one year of age was observed (139). However,

large RCTs in VLBW infants have shown high rates of hypoglycemia (19, 138).

Information regarding the bioavailability and half-life time of insulin is derived

from studies in adults and older children but no information is available for

insulin treatment kinetics in preterm infants. Furthermore, it was shown that

insulin infusion in ELBW infants decreases proteolysis, but also protein

synthesis, resulting no net reduction in protein catabolism, and concerns were

also raised for accompanying lactic acidosis (140).

It was previously described that more than 75% of the insulin in an insulin

infusion is adsorbed by glass and polyvinylchloride containers and by the infusion

set (141). This delays the time to reach steady state of insulin delivery, i.e. longer

time is needed for the treatment to be effective. Therefore, it was suggested to

administrate insulin in combination with human serum albumin. Priming the

tubing used with an insulin flush before initiation of insulin infusion therapy

blocks nonspecific binding sites, enhances the delivery of insulin to the patient

and might accelerate the correction of hyperglyclemia (142).

Hypoglycemia

Definition

As with hyperglycemia, hypoglycemia also lacks a clear definition, especially in

the preterm infant population, and the range of different definitions used is wide

(143). A meta-analysis suggested the following definitions for hypoglycemia in

full-term newborns (144): < 1.6 mmol/L (age 1-2 hours), < 2.2 mmol/L (age

3-47 hours), < 2.6 mmol/L (age 48-72 hours). The American Academy of

Pediatrics defined the operational threshold for hypoglycemia as 2.2 mmol/L in

the first four hours of life and 2.6 mmol/L therafter (145). In a recently published

meeting abstract, CGM data from the first five days of life in 67 term healthy

Key points

Both adjusting parenteral nutrition and insulin infusion as treatment

for neonatal hyperglycemia have shown promising results in smaller

studies.

However, larger studies have raised concerns for adverse outcomes

associated with both treatments.

Thus, no clear recommendation for treatment of neonatal

hyperglycemia could be made.

24

infants included in the Glucose in Well Babies (GLOW) study showed that glucose

concentrations are stable around 3.3 mmol/L during the first 48 hours of life and

increase to a new plateau at 4.6 mmol/L after 72 hours (146).

Prevalence

Preliminary results from the GLOW study show that hypoglycemia < 2.6 mmol/L

in healthy newborns is common mainly during the first 72 hours of life and only

a few patients have later episodes of hypoglycemia (146). CGM registration

detected hypoglycemia in approximately 40% of VLBW infants during the first

two weeks of life, and mostly during the first two days of life (30, 147). A large

study by Lucas et al. demonstrated that 67% of infants born

< 1850 g experienced hypoglycemia < 2.6 mmol/L during their admission period,

with only 10% experiencing hypoglycemia after the first month of life (148).

Table 8 presents prevalence rates for hypoglycemia later during the admission

period in different populations of preterm infants.

Table 8. Prevalence of hypoglycemia during later periods of admission in preterm infants.

Study Population (N)

Definition Postnatal period

Prevalence

Mola-Schenzle et al. (31)

VLBW (41)

< 2.5 PMA 32-33 weeks

41% - 37% of episodes lasted 30 min

Pertierra-Cortada

et al. (16)

VPT (60)

< 2.6 lasting 30 min

PMA 36-39 weeks

23%

Staffler et al. (149)

VLBW (> 1000 g)

(44)

< 2.5 Exclusively enterally

fed

23%

ELBW (54)

44%

Glucose concentrations expressed in mmol/L. Mola-Schenzle et al. and Pertierra-Cortada

et al. used CGM-derived data while Staffler et al measured blood glucose concentrations.

Pathophysiology

Preterm infants are vulnerable to hypoglycemia due to lack of energy stores and

a lower capacity for production of alternative energy sources (150, 151). Glycogen

storage in the liver does not occur until the third trimester, and therefore the

ELBW infant has a limited glycogen reserve (10). It was postulated that poorly

coordinated counterregulatory hormone response, as well as peripheral

insensitivity to these hormones, might also contribute to hypoglycemia (152).

25

Risk factors

SGA status, Apgar score < 5 at 5 minutes after birth and lower body mass index

(BMI) at birth were found to be risk factors for hypoglycemia (148).

Outcomes

Much controversy exists regarding potential brain damage caused by

hypoglycemia. On the one hand, hypoglycemia has been associated with adverse

neurological outcomes and radiological findings in multiple studies (148, 153-

157). Frequent, moderate hypoglycemia was associated with greater

developmental deficits than severe but less frequent hypoglycemia, but the

specific threshold in which hypoglycemia exerts its effect on the brain is

unknown. It was recently shown that children born moderate to late preterm or

term who were exposed to severe (< 2 mmol/L), recurrent (≥ 3 episodes) or

clinically undetected (< 2 mmol/L for at least 10 minutes detected only by CGM)

neonatal hypoglycemia episodes had reduced executive and visual motor function

at 4.5 years of age (158).

On the other hand, large studies in children born preterm did not find differences

in neurodevelopmental outcomes at an array of ages from 2 to 18 years between

children who were and were not exposed to hypoglycemia (158-160).

Furthermore, a recently published large RCT in otherwise healthy newborns with

asymptomatic moderate hypoglycemia (< 2.6 mmol/L) showed that using a lower

glucose treatment threshold of 2 mmol/L was not inferior to a treatment

threshold of 2.6 mmol/L with regard to neurodevelopmental outcomes at

18 months of age (161). A recent systematic review and meta-analysis concluded

that neonatal hypoglycemia was associated with visual-motor impairment but not

with neurodevelopmental impairment in early childhood, while it was associated

with neurodevelopmental impairment and low literacy and numeracy in

mid-childhood (162).

Treatment

In 2017, Swedish national guidelines for the prevention and treatment of

hypoglycemia in neonates born > 35 gestational weeks were published (163).

Hypoglycemia should be prevented by promoting breastfeeding and skin-to-skin

contact with the parent(s) as soon as possible after birth (within 1 hour).

Hypoglycemia < 2.6 mmol/L should be treated with dextrose gel and feeding. If

the hypoglycemia persists, or if blood glucose is < 1.5 mmol/L, glucose infusion

should be started.

26

Blood pressure

Increased BP is a known risk factor for stroke and death from cardiovascular

disease. It has been previously shown that children (at all age ranges) and adults

born preterm have increased BP as well as increased risk for cerebrovascular

disease compared with term-born peers (164-170). A small reduction of 2 mm Hg

in the BP at the individual level leads to a reduction of 17% in the prevalence of

hypertension, 6% in the risk for congestive heart disease and 15% in the risk for

stroke and transient ischemic attacks in the population level (171).

Mechanisms

The mechanisms by which BP levels are affected by preterm birth are not fully

understood. Different theories include underdevelopment of the kidneys with

lower number of nephrons, disturbed growth of the vascular tree leading to

increased peripheral vascular resistance, and increased activity along the

sympathetic-adrenal axis (172-174). A recently published study suggested that

decreased growth in the postnatal period induces higher cortisol and

dehydroepiandrosterone levels at six years of age, which in turn were associated

with higher systolic blood pressure (SBP) at this age (175).

Early programming hypothesis

Barker et al. formulated the early programming hypothesis: events or insults

occurring during sensitive periods of development early in life, so called critical

periods, can affect long-term metabolism (176). Thus, IUGR, low birth weight and

prematurity might cause hypertension, coronary heart disease and diabetes. In

the same way, postnatal nutrition and early growth might affect long-term

outcomes. The causality implied in the hypothesis has thus far been difficult to

prove.

Key points

Hypoglycemia is common in preterm infants, especially in the first

days of life.

It might affect the neurodevelopment of the infants, but this effect is

debated.

27

Growth and blood pressure

Birth weight and birth weight z-score

In the general population, it was found that lower birth weight z-scores were

associated with higher SBP and diastolic blood pressure (DBP) at adult age (177).

At 9-11 years of age, SBP and DBP were lower by 2.8 and 1.4 mm Hg for each 1 kg

increase in birth weight, respectively (178). Other studies, including a large

meta-analysis, have shown similar results (179-187). Among adolescents and

adults born preterm, however, multiple studies have failed to show a significant

association between birth weight and birth weight z-scores and cardiovascular

measures, especially BP, later in life (167, 188-190).

Growth

Weight gain and weight gain velocity during the first years of life in preterm

infants were shown to be associated with higher SBP and DBP at 6.5 and 16 years

of age (191, 192). A subanalysis of 38 VLBW-born children included in the

NIRTURE study found that SBP at 2 years of age was significantly positively

correlated with increase in length and length z-scores between 6-12 months of age

and with increase in weight, length, and length z-scores between term age and

2 years of age (but not between term age and 6 months of age) (193). DBP was

not associated with growth. Contrarily, earlier results from a regional subset of

the EXtremely PREterm infants in Sweden Study (EXPRESS) including 68

children born EPT showed that increased weight z-score change between PMA 36

weeks and 2.5 years of age was associated with lower SBP and DBP z-scores at 6.5

years of age (164). A longitudinal cohort study in adolescents born preterm did

not find any association at all between weight gain during the immediate

postdischarge period and later BP (194). Finally, a Finnish longitudinal cohort

study including 125 adults born VLBW did not find an association between weight

gain and linear growth during weeks 1-3 and 4-6 of life and BP at the ages of 22.5

and 25.1 years (195).

Nutrition and blood pressure

In mice models, underfeeding during the early postnatal period led to defective

insulin secretion, while overfeeding led to insulin resistance. Both underfed and

overfed mice had increased BP (196). Multiple studies have provided evidence for

the association between nutrition and BP in term infants, especially with regards

to breastfeeding which was repeatedly shown to be associated with decreased BP

at later age (197-201). Increased maternal carbohydrate intake during pregnancy

was associated with increased SBP in the child at four years of age (202).

In preterm infants, however, more conflicting results were found. The Helsinki

study of VLBW-born adults (N = 125) found that neither higher protein, fat nor

carbohydrate intakes during postnatal weeks 1-3, 4-6 or 7-9 were associated with

28

BP in adulthood (195). An RCT including 758 preterm infants who received early

diets with different nutrient contents (preterm formula versus standard formula

versus donor breast milk) did not find a programming effect of the nutritional

intake and weight gain on BP at 7.5-8 years of age (203). Surprisingly, in the very

same cohort it was later found that breast milk consumption was associated with

lower mean BP by 4.2 mm Hg at age 13-16 years, thereby supporting the theory

of early programming of later cardiovascular outcomes by early-life nutrition

(204).

Nutrition, growth and blood pressure

Results from the EXPRESS cohort have previously shown that energy and protein

intakes are independent predictors of neonatal growth (9). An RCT that used

protein-energy supplementation versus low energy/no protein supplementation

in 450 infants born full term in the 1970s showed no effect of the nutrition given

or of length gain z-scores during the first two years of life on BP in adulthood

(205). On the other hand, an RCT in 6-8 years old children born SGA who as

neonates received either standard or a nutrient-enriched formula found that

children who received the higher protein intake had higher DBP by 3.5 mm Hg

than the control group, and faster weight gain in infancy was associated with

higher BP later in life (206). This raised a concern for possible “overnutrition”,

where overzealous nutrition in order to promote growth might affect long-term

cardiovascular disease programming.

Key points

Increased blood pressure is common in children and adults born

preterm.

Programming of blood pressure by early-life nutrition and growth

has been suggested.

29

Objectives

The primary aim of this doctoral thesis was to investigate neonatal hyperglycemia

in VLBW and EPT infants, including its epidemiology, risk factors (including the

possible associations with nutritional intakes), consequences and treatment. A

secondary aim was to investigate the programming effect of early-life nutrition

and growth on cardiovascular outcomes in children born EPT.

The specific aims were:

- To describe the trends in blood glucose concentrations during the first

four weeks of life in EPT infants, to determine its relation to the

nutritional intake, and to investigate its association with neonatal

mortality, as well as any differences associated with insulin treatment

(Paper I).

- To investigate the associations between early-life nutrition, postnatal

growth and neonatal hyperglycemia, and BP at 6.5 years of age in

EPT-born children (Paper II).

- To explore the associations between neonatal hyperglycemia in EPT

infants and neurodevelopmental outcomes at 6.5 years of age (Paper III).

- To determine the prevalence and duration of and risk factors for

hyperglycemia episodes at PMA 36 weeks in VLBW infants (Paper IV).

30

Materials and Methods

Study population

The EXPRESS cohort (Papers I-III)

The EXPRESS cohort is a population-based study including all infants born

before 27 completed gestational weeks in Sweden between April 1, 2004 and

March 31, 2007. All neonatal units in Sweden, including all seven University

hospitals with their respective uptake areas, participated in the study. The study

included 707 live-born infants, of them 602 survived the first 24 hours of life.

Included and excluded infants in papers I and II are described in Figures 3 and

4. For Paper III this is described in Figure 1 in the respective paper.

Figure 3. Included and excluded infants in Paper I.

Live-born EPT infants

N = 707

Died before 24 hours (N = 105)

Missing perinatal data (N = 14)

Chromosomal or congenital

anomalies (N = 8)

Included in analyses

N = 580

Died before 48 hours (N = 11)

Included in analyses involving

hyperglycemia on 2 consecutive days

N = 569

Died before 72 hours (N = 12)

Included in analyses involving

hyperglycemia on 3 consecutive days

N = 557

31

Figure 4. Included and excluded infants in Paper II.

The LIGHT study (Paper IV)

The Very Low Birth Weight Infants - Glucose and Hormonal Profiles over Time

(LIGHT) study is a prospective cohort study including 50 VLBW infants born

between October 1, 2016 and November 30, 2019 and admitted to the neonatal

intensive care unit (NICU) at Umeå University hospital, Sweden. Included and

excluded infants in paper IV are described in Figure 1 in the respective paper.

Live-born EPT infants

N = 707

Died before 6.5 years (N = 213)

Live in non-participating

regions (N = 232)

Congenital heart/lung

malformations (N = 12)

Invited to cardiovascular

follow-up at 6.5 years

N = 250

Performed follow-up

N = 205

Declined participation (N = 38) Lost to follow-up (N = 7)

Included in analyses

N = 171

No measurements obtained (N = 5)

Ongoing cardiovascular disease

(N = 2)

Follow-up visit not

within time period (N = 27)

32

Data collection

A summary of the data collected in the different papers is presented in Table 9.

Table 9. Data collection in Papers I-IV.

Paper I Paper II Paper III Paper IV

Perinatal data X X X X

Daily glucose data X X X X *

Daily insulin data X X X X *

Anthropometric measurements at birth

X X ** X X

Nutrition data X X

Blood pressure data X

Neurodevelopmental outcomes data

X

CGM data X

* Only data regarding hyperglycemia, hypoglycemia and insulin treatment during the

entire admission period (yes/no). ** Also on days 28 and 56 of life.

Perinatal data

For all papers, perinatal data was prospectively collected from the medical

journals and from the Swedish Neonatal Quality register (SNQ).

Glucose data

For Papers I-III, all available glucose measurements were retrospectively

retrieved from the medical records of the patients, and data for the first four

weeks of life was used. For each calender day, first, highest and lowest glucose

concentrations were registered, regardless if the blood sample was arterial,

venous or capillary. Most commonly, glucose concentrations were analyzed by

blood gas analyzer at the neonatal unit. Samples taken from a venous line where

glucose was simultaneously infused were excluded. For Paper IV, prospectively

collected dichotomous data regarding the occurrence of hyperglycemia and

hypoglycemia during admission period (yes/no) was registered.

Insulin data

For Papers I-III, data regarding insulin treatment were registered retrospectively

from the medical records of the patients for each treatment day. As no guidelines

were in place for hyperglycemia treatment, insulin was given according to the

clinical judgement of the attending neonatologist. For Paper IV, dichotomous

33

data regarding the administration of insulin during the admission period

(yes/no) were extracted.

Anthropometric data

For Papers I-III, prospectively registered weight and length measurements were

used, including measurements registered at birth and on days 28 and 56 of life.

Missing data were interpolated linearly based on existing data. Z-score

calculations were based on a Swedish growth reference for infants born > 24

gestational weeks (207). A Canadian reference was used for infants born < 24

gestational weeks (208). SGA was defined as birth weight < -2 SD. For paper II,

we used prospectively registered height and weight measurements taken on a

cardiovascular follow-up visit at 6.5 years of age. For Paper IV, only data

regarding measurements at birth were extracted.

Nutrition data (Papers I, II)

Data regarding nutrition were retrospectively collected from hospital records

daily for the first four weeks of life and weekly thereafter (on days of life 35, 42,

etc.) (9). The data included macronutrient intakes (carbohydrates, protein and

fat) - enteral, parenteral and total. Intakes included flush solutions and saline

infusions. Transfused blood products were included only in paper I. Nutrient

intakes were calculated using manufacturer data and breast milk analyses (using

midinfrared spectrophotometry analyses). When breast milk analyses were not

available, an average content was assumed for the sample. Daily nutrition intakes

were calculated between 6am and 6am the following day.

Blood pressure data (Paper II)

We used prospectively collected data from a cardiovascular follow-up visit at

6.5 years of age. The follow-up visit was scheduled during morning time. After

15 minutes of rest in the examination room while sitting, a validated oscillometric

device (Omron HEM 907; Omron Healthcare, Kyoto, Japan) was used to measure

SBP and DBP in the right arm which was placed at the level of the heart. Three

measurements, at least two minutes apart, were performed. Mean and z-scores

were calculated using age-, sex-, and height-adjusted BP coefficients for children

(209).

Neurodevelopmental outcomes data (Paper III)

Prospectively collected data were used from a neurodevelopmental assessment

performed at a follow-up visit at 6.5 years of age. Intellectual ability was assessed

using Wechsler Intelligence Scale for Children IV (WISC-IV) (210). This is a

four-domain test administered by a psychologist. The domains tested include

verbal comprehension, perceptual reasoning, working memory and processing

34

speed. Summarizing scores from all four domains, a full scale index was

calculated and was converted to a gender- and age-specific scale, the IQ score

(FSIQ).

Motor skills were assessed by using the Movement Assessment Battery for

Children 2 (MABC-2) test (211). This is a standardized three-domain test,

administered by a physiotherapist, which aims to identify children who have

motor function impairment by assessing ball skills, static and dynamic balance

and manual dexterity through eight tasks. It is validated for use in children

3-16 years old. Higher scores indicate better motor function.

Functional level in children with cerebral palsy was assessed by using the Gross

Motor Function Classification System (GMFCS) (212). The need of hearing aids

was registered. An ophthalmologist performed a visual acuity test. Data from all

of the aforementioned evaluations (except MABC-2), as well as a clinical

examination and a medical record review, were used to assign the children an

NDD category (Table 10).

Table 10. Definitions of neurodevelopmental disability categories.

Criteria No or mild NDD Moderate to severe NDD FSIQ score ≥ -2 SD < -2 SD Cognitive disability on clinical examination or medical record review

No/mild Moderate/severe

GMFCS score 1 ≥ 2 Visual acuity ≥ 20/63 < 20/63 Need of hearing aid No Yes

Continuous glucose monitoring data (Paper IV)

Rationale

Monitoring glucose concentrations in premature infants is done at a higher

frequency on the first days and weeks of life, but when the clinical condition of

the infants stabilize and they do not seem to be at risk of glucose disturbances any

longer, this monitoring is done in increasing intervals until it eventually stops.

This means that there are prolonged periods in which we do not know if, how

much and for how long these infants are actually exposed to glucose disturbances.

CGM is a method used primarily in type 1 (insulin-dependent) diabetes mellitus

patients which allows for continuous registration of glucose concentrations over

time. Currently commercially available CGM devices are not recommended by the

manufacturer for use under two years of age (213).

35

The method

The device uses a disposable oxidase-based platinum electrode, the sensor, which

is inserted under the skin. Interstitial glucose catalyzed generates an electrical

signal sent to the transmitter unit that is placed on the skin. From the transmitter,

data are being transmitted further to a monitor device. The “first generation”

models collect real time data every 10 seconds and a mean value calculated for

the previous 5 minutes is the final output. “Newer generation” algorithms are able

to provide real-time data.

Safety

CGM is able to detect more hyperglycemia and hypoglycemia episodes than blood

sampling, and it has been reported to be a safe and clinically adequate method to

estimate glucose concentrations in preterm infants (22, 30, 138, 147, 214-218).

The procedural pain associated with insertion of the CGM device has previously

been shown to be milder (lower pain scores) than the pain associated with heel

stick (219). Disadvantages of the method are discussed later, see Discussion.

Study procedure

All infants included in the LIGHT study who remained in the study at PMA

36 weeks ± 2 days (n=35) were continuously monitored for glucose by using a

CGM system. This time point was chosen because: a) this is a period where

glucose concentrations are rarely routinely monitored for; b) the infants are

nearing term age; and c) the infants are not usually discharged from the hospital

prior to this age. A Dexcom G4 PLATINUM sensor (Dexcom Inc., San Diego,

California), a “first generation” device, was inserted in the frontolateral aspect of

the thigh of the infant. Registration time was 48 hours and clinical decisions were

not made based on CGM data. The procedure was performed at the University

hospital in Umeå as well as at four county hospitals. Calibration of the CGM

device was done every 12 hours by pediatric nurses using point-of-care

glucometers (Accucheck Inform II, Roche Diagnostics, Basel, Switzerland or

HemoCue Glucose 201+, HemoCue, Ängelholm, Sweden, depending on the

hospital).

Post hoc correction

CGM values calibrated by Accucheck point-of-care glucometers were significantly

lower by 1.0 mmol/L (95% CI of difference 0.98 – 1.09) compared to CGM values

calibrated by HemoCue devices. We therefore corrected CGM values calibrated

by Accucheck devices with a factor of +0.5 mmol/L while CGM values calibrated

by HemoCue devices were corrected with a factor of -0.5 mmol/L.

36

Figure 5. A Dexcom G4 PLATINUM CGM transmitter. Dimensions (including sensor

pod): 3.8 x 2.3 x 1.0 cm. Weight (including sensor pod): 8.5 g. © Picture by Itay Zamir.

Quality control

Trained staff at the neonatal units collected the data in the EXPRESS study and

double controls were performed continuously with the original records. At each

of the seven health care regions in the EXPRESS study there were two study

coordinators (one obstetric and one pediatric) responsible for quality control of

the data collection. Furthermore, internal and external controls were performed

on a random sample of infants. Unreasonable values identified in descriptive

statistics were cross-checked against the original records. All glucose records in

the LIGHT study were collected by trained staff at the neonatal unit as well. All

data registration was supervised by the author of this thesis.

Definition of hyperglycemia

Since no clear definition for neonatal hyperglycemia exists, we tested different

combinations of glucose concentration thresholds commonly used in the

literature (see Background) and frequencies.

37

Papers I-III

In these papers, four thresholds were used to define neonatal hyperglycemia:

> 8, 10, 12 and 14 mmol/L. The threshold of 14 mmol/L was not used in Paper I,

while results for hyperglycemia > 8 mmol/L were reported only in Paper III (see

Results). Three frequency definitions for neonatal hyperglycemia were used:

occurring at least once, on 2 and on 3 consecutive days during the first 28 days of

life. Results for the first of these definitions were not reported in Paper II (see

Results). Thus, 12 definitions in total were used (Table 11).

Table 11. Various definitions of neonatal hyperglycemia used in Papers I-III.

Frequency Glucose threshold (mmol/L)

> 8 > 10 > 12 > 14

At least once Paper III Papers I, III Papers I, III Paper III

On 2 consecutive days Paper III Papers I-III Papers I-III Papers II-III

On 3 consecutive days Paper III Papers I-III Papers I-III Papers II-III

Paper IV

Protracted hyperglycemia was defined as interstitial glucose concentrations > 8

mmol/L lasting for at least 30 minutes. Protracted hypoglycemia was defined as

interstitial glucose concentrations < 2.6 mmol/L (< 2.8 mmol/L in

Accucheck-calibrated CGM data) lasting for at least 30 minutes.

Statistical analysis

In all papers, data were analyzed by using SPSS Statistical software (IBM corp.

Released 2016. IBM SPSS Statistics for Windows, version 24.0. Armonk, NY: IBM

Corp.). In paper I, R (version 3.3.2, R Foundation for Statistical Computing,

Vienna, Austria) was used as well. The significance level in all papers was set to

p < 0.05.

Paper I

Generalized additive models including a random effect for each patient and a

smooth spline adjusting for an average time trend in observed glucose

concentrations were used to investigate the associations between nutritional

intakes and plasma glucose concentrations. In practice, we investigated how the

daily nutritional intakes were associated with glucose concentrations on the

following day. Days on which the infants received insulin treatment were

excluded from these analyses in order to avoid the effect of insulin on glucose

concentrations. Logistic regression models were used to explore the associations

between hyperglycemia and insulin treatment, and mortality up to 28 and 70 days

38

of age. All models were adjusted for gestational age at birth and birth weight, with

some of the models adjusted for further morbidity factors.

Paper II

Multiple linear regression models were used to investigate the associations

between nutritional intakes, growth parameters and duration of neonatal

hyperglycemia, and BP z-scores at 6.5 years of age. Total intakes were calculated

for weeks 1-4, 5-8 and 1-8 of life. Analyses of covariance were used to analyze

differences in BP z-scores between children who were and were not exposed to

neonatal hyperglycemia. All models were adjusted for gestational age at birth and

heart rate at follow-up.

Paper III

Generalized linear mixed models including a random effect for twins/triplets

(and in some models also for treating hospital) were used to evaluate the

associations between neonatal hyperglycemia, its duration and insulin treatment,

and neurodevelopmental outcomes at 6.5 years of age. Models were adjusted for

variables that were shown to be associated with both exposure and outcome

variables and that significantly affected the association of interest.

Paper IV

Pearson correlation coefficient was calculated in order to evaluate the agreement

between CGM and capillary glucose values and a Bland-Altman plot was

constructed. Mann-Whitney and chi-square tests were used to compare

continuous and binary variables between groups, respectively. Linear regression

models were used to assess possible risk factors for time spent in hyperglycemia

and hypoglycemia.

Ethical approval

All studies in this thesis were carried out according to the Declaration of Helsinki

– Ethical principles for medical research involving human subjects, developed by

the World medical association. The EXPRESS study (including the follow-up

described in Paper III) was approved by the regional ethics committee in Lund,

Sweden (no. 42/2004, 138/2008, 524/2009, 488/2015, 18/2016, 970/2016).

The follow-up described in Paper II was approved by the regional ethics

committee in Stockholm, Sweden (no. 520-31/2/2010, amendment

no. 376-32/2011). Parents provided informed consent for data acquisition. The

ethical committee granted waiver of consent for additional data acquisition

regarding nutrition data. The LIGHT study was approved by the regional ethics

committee in Umeå, Sweden (no. 226-31M/2016). Parents provided informed

consent for all study procedures.

39

Power analysis

The LIGHT study was planned to include 50 VLBW infants. A large drop-out

margin of 40%, leaving 30 infants in the study, was taken into consideration. It

was calculated that including 30 infants would provide the study a power of 80%

at a significance level of 5% to detect a significant association between glucose

concentrations and total carbohydrate intake (based on results presented in

Paper I – an association between a 3.0% increase in glucose concentration

following a 1 g/kg/d increase in total carbohydrate intake; see Results).

40

Results

Patient characteristics

Table 12 describes baseline charecteristics of the infants included in the different

studies. In paper IV, out of the 35 VLBW infants included in the CGM registration

at PMA 36 weeks, 22 were ELBW.

Table 12. Baseline characteristics of infants included in papers I-IV.

Paper I Paper II Paper III Paper IV

Infants included, N 580 171 533 35

Gestational age (weeks), mean (SD)

25.3 (1.1) 25.4 (1.0) 25.3 (1.1) 27.3 (2.6)

Birth weight (g), mean (SD) 763 (168) 782 (165) 761 (170) 929 (276)

Birth weight z-score, mean (SD)

-0.79 (1.2) -0.71 (1.2) -0.80 (1.2) -1.41 (1.5)

Males, N (%) 317 (54.7) 95 (56.0) 288 (54.0) 12 (34.3)

Multiple gestation, N (%) 121 (20.9) 35 (20.5) 116 (21.8) 3 (8.6)

Apgar score at 5 minutes after birth, mean (SD)

7 (2) 7 (2) 7 (2) 7 (2)

IVH grade 3-4, N (%) 78 (13.4) 19 (11.1) 76 (14.3) 3 (8.6)

NEC, N (%) 33 (5.7) 10 (5.8) 29 (5.4) 0 (0.0)

Duration of mechanical ventilation treatment (days), mean (SD)

14.2 (17.1) 14.5 (15.6) 14.5 (17.5) 7.7 (12.9)

Culture-verified sepsis, N (%)

263 (45.4) 92 (53.8) 245 (46.0) 3 (8.6)

Patent ductus arteriosus, N (%)

339 (58.4) 103 (60.2) 310 (58.2) 21 (60.0)

ROP stage 3-5, N (%) 173 (34.5) (N = 501)

60 (35.7) (N = 168)

158 (35.2) (N = 449)

6 (17.1)

Bronchopulmonary dysplasia, N (%)

390 (67.2) 132 (77.2) 357 (67.0) 16 (45.7)

Insulin treatment, N (%) 86 (14.8) 15 (8.8) 77 (14.4) 7 (20.0)

Prevalence and duration of glucose disturbances

Neonatal hyperglycemia in EPT infants

We observed that 51% of EPT infants had hyperglycemia > 10 mmol/L at least

once during the first week of life. The prevalence of neonatal hyperglycemia (i.e.

during the first 28 days of life) in EPT infants according to different threshold

and frequency definitions is presented in Table 13.

41

Table 13. Prevalence of neonatal hyperglycemia (according to various definitions) in EPT

infants.

At least once On 2 consecutive days On 3 consecutive days

> 8 86.7% 65.4% 46.1%

> 10 69.8% 44.3% 29.4%

> 12 52.6% 29.4% 16.2%

> 14 41.6% 18.5% 8.3%

Glucose concentrations expressed in mmol/L.

Mean glucose concentrations, as well as the daily prevalence of hyperglycemia,

increased steadily during the first and second weeks of life, reaching a peak of

hyperglycemia in 30% of tested infants on day 12 of life (Paper I). Thereafter, a

slow decrease was noted during weeks three and four of life, but nevertheless,

> 20% of tested infants had hyperglycemia at week four of life.

Glucose disturbances at postmenstrual age 36 weeks in VLBW

infants

Most (63%) of the VLBW infants who were monitored with CGM at PMA

36 weeks had protracted dysglycemia (hyperglycemia, hypoglycemia or both;

Paper IV). Protracted hyperglycemia was registered in 54% of the infants, with

11% having had protracted hyperglycemia > 10 mmol/L. Protracted hypoglycemia

was registered in 29% of the infants. Hyperglycemia and hypoglycemia episodes

lasted for a mean of 4.3 and 2.0 hours, respectively. Almost half of the patients

who experienced protracted dysglycemia had only one or two episodes, while over

40% experienced five episodes or more.

Risk factors for glucose disturbances

Nutrition and Neonatal hyperglycemia in EPT infants

Table 14 summarizes the associations observed between macronutrient

parenteral and total intakes during the neonatal period and glucose concentration

in EPT infants.

Glucose disturbances at postmenstrual age 36 weeks in VLBW

infants

Hyperglycemia

Amnionitis and prior hypoglycemia episodes during the admission period were

risk factors for protracted hyperglycemia at PMA 36 weeks in VLBW infants

(Paper IV). Other possible risk factors, such as sepsis, were not significantly

42

Table 14. Associations between change in mean daily macronutrient intakes and glucose

concentration in EPT infants during the first 28 days of life.

Intake Change in mean daily intake

Change in mean daily highest glucose concentration

Parenteral Carbohydrate

↑ 1 g/kg/d ↑ 1.6%*

Protein ↓ 1.4%**

Lipid ↓ 3.0%*

Total Carbohydrate

↑ 1 g/kg/d ↑ 3.0%*

Protein ↑ 0.8%

Lipid ↓ 0.3%

* p < 0.001; ** p < 0.05. Generalized additive models including all macronutrient intakes

simultaneosly, a random effect for each patient and a smooth spline adjusting for an

average time trend in observed glucose concentrations. Adjusted for gestational age at

birth and birth weight.

associated with hyperglycemia. Infants with protracted hyperglycemia at PMA

36 weeks were exposed to prior hyperglycemia (> 8 mmol/L) episodes during the

admission period more often than infants who were normoglycemic (or who were

non-hyperglycemic). Factors observed to be associated with time spent in

hyperglycemia at PMA 36 weeks in univariable linear regression models are

presented in Table 15. In multivariable models, male sex remained significantly

associated with longer time spent in hyperglycemia at PMA 36 weeks (Paper IV).

Table 15. Risk factors associated with time spent in hyperglycemia > 8 mmol/L at

postmenstrual age 36 weeks in VLBW infants.

Median time (IQR) spent in hyperglycemia, min

p value

Sex

Males 250 (96-459)

Females 50 (0-110)

0.002

SGA

Yes 15 (0-69)

No 100 (40-230)

0.046

Prior hyperglycemia > 8 during admission

Yes 100 (46-283)

No 5 (0-25)

0.004

IQR – interquartile range (25th-75th percentiles). Glucose concentrations expressed in

mmol/L. Univariable linear regression analyses.

43

Hypoglycemia

No risk factors differentiating infants with protracted hypoglycemia at PMA

36 weeks from those without hypoglycemia could be identified (Paper IV).

Factors observed to be associated with time spent in hypoglycemia at PMA 36

weeks in univariable linear regression models are presented in Tables 16 and 17.

Table 16. Risk factors associated with time spent in hypoglycemia at postmenstrual age

36 weeks in VLBW infants.

Median time (IQR) spent in hypoglycemia, min

p value

Prior hyperglycemia > 12 during admission

Yes 28 (8-185)

No 0 (0-25)

0.008

Prior insulin treatment during admission

Yes 165 (20-190)

No 0 (0-32.5)

0.016

Prior hypoglycemia < 2.6 during admission

Yes 0 (0-18.8)

No 20 (0-170)

0.039

IQR – interquartile range (25th-75th percentiles). Glucose concentrations expressed in

mmol/L. Univariable linear regression analyses.

Table 17. Risk factors associated with time spent in hypoglycemia at postmenstrual age

36 weeks in VLBW infants.

Unit of change Change in time spent in hypoglycemia, min

p value

Gestational age + 1 week - 11.6 0.035

Birth weight + 100 g - 14.9 0.003

Birth length + 1 cm - 10.7 0.009

Birth head circumference

+ 1 cm - 17.7 0.001

Apgar at 5 min + 1 point - 14.7 0.037

Apgar at 10 min + 1 point - 22.5 0.013

CPAP treatment duration

+ 1 day + 1.4 0.017

Univariable linear regression analyses.

In multivariable models, lower Apgar scores at 10 min after birth and prior

hyperglycemia > 12 mmol/L during the admission period remained significantly

associated with longer time spent in hypoglycemia at PMA 36 weeks (Paper IV).

Combined dysglycemia

Lower Apgar scores at 10 minutes after birth and longer treatment (in days) with

continuous positive airway pressure (CPAP) were risk factors associated with

protracted combined dysglycemia (having both protracted hyperglycemia and

hypoglycemia) at PMA 36 weeks (Paper IV).

44

Outcomes of neonatal hyperglycemia in EPT infants

Mortality

Hyperglycemia > 10 mmol/L occurring at least once during the first 28 days of

life in EPT infants was associated with a 2.45 times higher 28-day mortality risk

(p = 0.006; Paper I). Similarly, neonatal hyperglycemia > 10 mmol/L on

2 consecutive days was associated with a 2.55 times higher 28-day mortality risk

(p = 0.005).

Blood pressure

SBP z-scores were higher by 0.31 (equivalent to 2.4 mm Hg) in EPT-born children

who were exposed as infants to neonatal hyperglycemia > 12 mmol/L on

2 consecutive days (Paper II). Higher DBP z-scores were found in children who

were exposed as infants to neonatal hyperglycemia of various definitions,

independently of carbohydrate intake (Table 18 and Paper II). Longer duration

of hyperglycemia > 12 and 14 mmol/L was associated with higher SBP and DBP

z-scores (Table 19 and Paper II). However, hyperglycemia > 8 mmol/L

(regardless of its duration) was not associated with SBP or DBP z-scores. SBP z-

scores were not associated with neonatal hyperglycemia occurring at least once

during the first 28 days of life (regardless of threshold).

Table 18. Associations between neonatal hyperglycemia of various definitions and

diastolic blood pressure at 6.5 years of age in children born EPT.

Definition of neonatal hyperglycemia

Change in DBP z-score

Equivalent change in DBP, mm Hg

> 10

at least once - 0.12 - 0.9

on 2 consecutive days - 0.10 - 0.7

on 3 consecutive days + 0.24* + 1.7

> 12

at least once + 0.06 + 0.4

on 2 consecutive days + 0.23* + 1.6

on 3 consecutive days + 0.34* + 2.4

> 14

at least once + 0.22* + 1.6

on 2 consecutive days + 0.33* + 2.3

on 3 consecutive days + 0.27 + 1.9

* p < 0.05; Glucose concentrations expressed in mmol/L. Analyses of covariance adjusted

for gestational age at birth and heart rate at follow-up.

45

Table 19. Associations between duration of neonatal hyperglycemia of various definitions

and blood pressure at 6.5 years of age in children born EPT.

Definition of neonatal

hyperglycemia

Change in duration,

day

Change in SBP z-score

Equivalent change in

SBP, mm Hg

Change in DBP z-score

Equivalent change in

DBP, mm Hg

> 10 + 1

+ 0.02 + 0.16 + 0.02 + 0.14

> 12 + 0.03 + 0.24 + 0.04* + 0.28

> 14 + 0.05** + 0.4 + 0.05** + 0.35

* p < 0.01; ** p < 0.05; Glucose concentrations expressed in mmol/L. Multivariable linear

regression models adjusted for gestational age at birth and heart rate at follow-up.

Neurodevelopmental disabilities

Neonatal hyperglycemia was not associated with higher odds for moderate to

severe NDD or with lower odds for survival without moderate to severe NDD at

6.5 years of age in children born EPT (Paper III). Neonatal hyperglycemia was

not associated with WISC-IV scores at 6.5 years of age (Paper III).

Lower MABC-2 scores were found in children exposed to neonatal hyperglycemia

> 8, 10 and 14 mmol/L on 3 consecutive days, > 10 and 12 mmol/L on

2 consecutive days and > 14 mmol/L occurring at least once (Figure 6 and

Paper III). Longer exposure to neonatal hyperglycemia, regardless of threshold,

was associated with lower MABC-2 scores (Table 20 and Paper III).

Table 20. Associations between duration of neonatal hyperglycemia of various definitions

and MABC-2 scores at 6.5 years of age in children born EPT.

Definition of neonatal hyperglycemia

Change in duration

Change in MABC-2 score

> 8 + 1 day

- 0.8* > 10 - 1.1* a > 12 - 1.0** b > 14 - 1.5* c

* p < 0.001; ** p < 0.005; Glucose concentrations expressed in mmol/L. For reference,

MABC-2 total scores ≤ 56 and ≤ 67 are equivalent to the 5th and 15th centiles (i.e. high-

risk for, and at risk of motor impairment), respectively. Generalized linear mixed models

including a random effect for twins/triplets. All analyses were adjusted for clinical risk

index for babies (CRIB), no. of days without available glucose measurements during weeks

1-4 of life, sex, patent ductus arteriosus operated during weeks 1-2 of life, duration of

mechanical ventilation treatment during weeks 1-2 of life, educational status of mother

and age at follow-up, unless otherwise specified. a Not adjusted for duration of mechanical

ventilation treatment during weeks 1-2 of life; b Not adjusted for sex; c Not adjusted for no.

of days without available glucose measurements, sex and educational status of mother.

46

Figure 6. Exposure to neonatal hyperglycemia of various definitions and MABC-2 scores

at 6.5 years of age in children born EPT.

30

35

40

45

50

55

60

65 p = 0.039

MA

BC

-2 s

core

Hyperglycemia > 8 mmol/L

Hyperglycemia No hyperglycemia

30

35

40

45

50

55

60

65 p = 0.01 p = 0.032

MA

BC

-2 s

core

Hyperglycemia > 10 mmol/L

Hyperglycemia No hyperglycemia

30

35

40

45

50

55

60

65 p = 0.031

MA

BC

-2 s

core

Hyperglycemia > 12 mmol/L

Hyperglycemia No hyperglycemia

30

35

40

45

50

55

60

65 p = 0.005 p = 0.005

MA

BC

-2 s

core

Hyperglycemia > 14 mmol/L

Hyperglycemia No hyperglycemia

47

Insulin treatment in EPT infants

EPT infants who were exposed to hyperglycemia > 10 and 12 mmol/L (regardless

of its frequency) during the first 28 days of life and who received insulin

treatment had higher 28- and 70-day survival rates, even when adjusting for

multiple perinatal morbidities (Paper I).

Insulin-treated hyperglycemic infants had lower gestational age at birth and birth

weight, higher clinical risk index for babies (CRIB) score and received mechanical

ventilation for a longer period, compared to non-insulin-treated hyperglycemic

infants (Paper I). We did not observe differences in NEC, IVH or sepsis outcomes

between treated and non-treated hyperglycemic infants (Paper I).

We did not observe differences in the odds for moderate to severe NDD or for

survival without moderate to severe NDD at 6.5 years between insulin-treated

and non-insulin-treated hyperglycemic infants (Paper III). Neither did we

observe differences in WISC-IV nor in MABC-2 scores. Insulin treatment was not

associated with BP at 6.5 years of age.

More than a third (36.7%) of the infants who received insulin treatment

experienced hypoglycemia < 2.6 mmol/L on the same day they received insulin

treatment. The daily prevalence of hypoglycemia in insulin-treated infants was

higher than in non-insulin-treated hyperglycemic infants during most of the first

28 days of life, but it stabilized at 5% (0-10%) from day 4 of life and onwards.

Blood pressure at 6.5 years of age in EPT-born children

Early postnatal nutrition and blood pressure

Carbohydrate intake

Total carbohydrate intakes during weeks 1-4, 5-8 and 1-8 of life were significantly

associated with SBP and DBP z-scores at 6.5 years of age in children born EPT

(Paper II). Increasing the carbohydrate intake by 1 g/kg/d during the first eight

weeks of life was associated with an increase of 0.18 SD in SBP and 0.14 SD in

DBP at 6.5 years of age (Table 21).

Protein intake

Total protein intake during the first eight weeks of life was associated with DBP

z-scores at 6.5 years of age (Paper II). Increasing the protein intake by 1 g/kg/d

during the first eight weeks of life was associated with an increase of 0.4 SD in

DBP at 6.5 years of age (Table 21). No associations with SBP were observed.

48

Lipid intake

Total lipid intakes during either weeks 1-4, 5-8 or 1-8 of life were not associated

with SBP or DBP z-scores at 6.5 years of age (Table 21 and Paper II).

Table 21. Associations between change in mean daily macronutrient intakes and blood

pressure at 6.5 years of age in children born EPT.

Intake / Postnatal

period

Change in mean daily

intake

Equivalent change in SBP, mm Hg

Equivalent change in DBP, mm Hg

Carbohydrate Weeks 1-4

↑ 1 g/kg/d ↑ 1.1* ↑ 0.9*

Weeks 5-8 ↑ 0.9* ↑ 0.6*

Weeks 1-8 ↑ 1.4* ↑ 1.1*

Protein Weeks 1-4

↑ 1 g/kg/d ↑ 1.1 ↑ 1.6

Weeks 5-8 ↑ 1.7 ↑ 1.7

Weeks 1-8 ↑ 2.9 ↑ 2.8**

Lipid Weeks 1-4

↑ 1 g/kg/d 0 ↓ 0.3

Weeks 5-8 ↓ 0.1 ↓ 0.2

Weeks 1-8 ↓ 0.2 ↓ 0.4

* p < 0.01; ** p < 0.05. Multivariable linear regression models adjusted for gestational age

at birth and heart rate at follow-up.

Early postnatal growth and blood pressure

No significant associations were observed between weight and weight change

during the first eight weeks of life and BP at 6.5 years of age (Paper II). However,

increase in weight z-score between week 5 of life and PMA 36 weeks, as well as

greater BMI and ponderal index at PMA 36 weeks, were associated with increased

SBP z-scores (Table 22). Shorter length measurements at four and eight weeks of

life (but not length gain) were associated with incrased DBP z-scores (Paper II).

Table 22. Associations between growth parameters and systolic blood pressure at 6.5

years of age in children born EPT.

Unit of change Equivalent change in SBP, mm Hg

Weight z-score change, week 5 of life to PMA 36 weeks

+ 1 ↑ 1.1*

BMI at PMA 36 weeks + 1 kg/m2 ↑ 0.8*

Ponderal index at PMA 36 weeks + 1 kg/m3 ↑ 0.3*

* p < 0.05. Multivariable linear regression models adjusted for gestational age at birth and

heart rate at follow-up.

49

Discussion

Discussion of methodology

Study design

The studies in this project all used an observational approach with, to varying

extent, retrospective and prospective data collection. Major strengths of this

project include a population-based observational design, large population size

and high participation rate in follow-up. The use of well validated psychological

and motor skill tests, as well as a systematic procedure for measuring blood

pressure, add to the validity of the results. This supports the internal validity of

the studies, i.e. the outcomes of the studies, with good approximation, could

describe the true situation in the study population. In paper IV, the study was

powered to detect an association between carbohydrate intake and glucose

concentrations, but not to detect differences in secondary outcomes, such as risk

factors for hyperglycemia.

Study population

Since these studies included almost all children born EPT in Sweden between

2004 and 2007 even at follow-up, there are good grounds to believe that the

findings are representative for this patient population in a western world NICU

setting. Different NICUs in Sweden differ from one another by their treatment

traditions while clinical outcomes are relatively comparable. For example, while

the NICU at Umeå University hospital implements CPAP and insulin treatments

more often than other NICUs in the country, the survival rate for EPT infants is

comparable at all NICUs (1).

Some changes in clinical practice in the NICUs in Sweden have occurred since the

EXPRESS study was first started. Nutrition is now routinely calculated in most of

the Swedish NICUs using the same computer program, which was not the

standard during the EXPRESS study period. This leads to nutrition intakes closer

to those recommended by ESPGHAN (220). No specific change regarding

neonatal hyperglycemia and its treatment was introduced during or after the

study period.

Drop-out and attrition bias

Longitudinal studies (studies following a group of patients over a period of time)

are bound to be affected to some extent by drop-out of participants due to

different reasons (death, loss to follow-up, decline to participate in a follow-up).

This might lead to attrition bias, i.e. only a specific group of patients, that may

50

differ in its baseline characteristics or outcome variables from the rest of the

cohort, is followed-up and thus might not represent the entire population studied.

In paper I, this bias might have been caused by the fact that a decreasing number

of neonates was tested for glucose towards the end of the first month of life. It is

thus possible that glucose concentrations were monitored mainly in those infants

who were suspected to have glucose concentration disturbances in the first place.

To address this possible bias, we have compared hyperglycemia rates between

different NICUs based on their frequency of glucose testing and did not observe

any significant differences. Therefore, it is unlikely that hyperglycemia rates were

overestimated due to attrition bias. In papers II-IV, a bias might have been

introduced due to drop-out of patients before the follow-up was performed.

However, drop-out analyses have confirmed that the infants remaining in the

different studies did not differ significantly in their baseline characteristics from

the infants that did not remain in the studies. Thus, there are good grounds to

believe that the results are generalizeable for the entire original study population.

Confounding

Observational studies are prone to confounder bias, limiting the ability to

investigate causality. A confounder is a variable external to the association

studied, which is associated with both the exposure and the outcome, as well as

had occurred before the exposure could have affected the outcome. Confounders

might affect associations studied, and although we have carefully considered,

included and accounted for different confounders in our analyses, residual

confounding effect might still remain. Similarly, “overadjusting” might also be a

problem, leading to unnecessary diminished associations. However, we have

carefully chosen the confounding variables accounted for and they were not

inserted to models based on results in univariable models only.

Multiple testing

When many associations are being tested for, some might appear to be significant

by chance. Some suggest that in these cases the p-value should be diminished

(e.g. p-value should be < 0.01 to be regarded as significant) to reduce the risk for

this problem. However, different post hoc tests for multiple testing problem

(Benjamini-Hochberg, Bonferroni) showed that most of the results in our studies

remained significant. A problem associated with applying these post hoc tests is

that the power of the study is then being reduced markedly, which is of

importance especially in this patient group.

51

Data quality

Glucose concentrations, nutritional intakes, anthropometric measurements as

well as morbidity data are meticulously documented at NICUs in Sweden.

However, some misclassification of morbidities and erroneous documentation of

data in clinical charts might always be present to some extent. Adding or

revalidating data is a difficult task in such a large cohort as in Papers I-III,

especially since the medical journals are paper-written and physically located in

almost 40 different hospitals throughout the country. Data quality was controlled

as previously described in the Methods section.

Continuous glucose monitoring method

Advantages

CGM might help to delineate and define what should be regarded as

“normal” regarding glucose metabolism, as was shown in the Glucose in Well

Babies (GLOW) study (146).

Studies have shown that more dysglycemia episodes can be detected using

CGM than with blood sampling (22, 30, 221).

Using CGM reduces the number of blood samples needed to monitor glucose

concentrations as well as involves minimal pain (219, 221).

Using CGM in the NICU might lead to better management of glucose

concentrations by guiding glucose and insulin infusion practices (216, 221-

224).

The CGM method allows for many more glucose concentration

measurements, which in turn increases the number of data points available

and thus the power of studies.

Safety

CGM systems have been studied in the population of preterm infants and shown

to be a safe and clinically adequate method to estimate glucose concentrations

(22, 147, 214-218). The NIRTURE study used CGM during the first week of life in

infants born as early as 23 gestational weeks and the method was well tolerated

(138).

Disadvantages

Limited point accuracy: This due to delayed diffusion of glucose to the

interstitial space, especially at the lower ranges of glucose concentrations

(30). This might lead to a time delay in the detection of hypoglycemia.

Furthermore, calibrations done during dysglycemia periods can be

inaccurate due to this lag between blood and interstitial glucose

concentrations.

52

Drift phenomenon: There is a drift in the readings between calibrations due

to changes in the probe surface (biofilm build-up) (225). This causes sensor

inaccuracy over time, and if the calibration algorithm relies on previous

calibration points, then new calibration points will still be “tainted” by the

drift that affected the previous ones.

Lowest detection threshold: CGM devices have lowest detection threshold of

2.2 mmol/L.

Margin of error: A typical CGM device has a mean absolute relative

difference of 7-12% (226). Furthermore, CGM error might be greater due to

error in the calibration, using point-of-care glucometers (227).

Not recommended for use: CGM systems are not yet recommended by the

manufacturers for use in preterm infants.

Not real-time data: “Older” generation devices, such as the one used in Paper

IV, present data for the past five minutes and not real-time data.

The results of our study thus need to be interpreted with caution. CGM is a good

method for research of the physiology of glucose disturbances in preterm infants.

However, further technical advancements and greater knowledge as to how to

interpret the large database generated by this method are needed to be achieved

before any recommendation regarding the clinical management of glucose

disturbances based on CGM data can be made.

Point-of-care glucometers

As was shown in Paper IV, capillary glucose concentrations registered in hospitals

that used Accucheck Inform II as their point-of-care glucometer differed

significantly from those registered in hospitals that used HemoCue Glucose 201+.

A higher correlation with the CGM-derived measurements was shown with the

HemoCue Glucose 201+ glucometer (r = 0.77) than with AccuCheck Inform II (r

= 0.54). That is despite that both devices have been shown to be precise and

reliable (228, 229). However, it was previously shown that Accucheck Inform II

tends to overreport hypoglycemia in newborn infants (227). HemoCue Glucose

201+ used in a NICU setting, on the other hand, was shown to report slightly

(though not statistically significant) higher glucose concentrations (by 0.33

mmol/L) than laboratory values (230). We have not yet compared the reported

glucose concentrations with laboratory values due to ongoing data extraction.

CGM-generated values determined by the calibration values obtained using these

glucometers have undoubtedly been affected by the difference between the two

glucometers used, and the introduction of a correction factor (see Methods)

might have ameliorated the problem but have probably not completely eliminated

it.

53

Ethical considerations

Insertion of the CGM sensor used in Paper IV was done by a medical team

member (physician or nurse) after being trained in performing the procedure by

a nurse specially trained in the care of diabetes mellitus in children. Inserting the

CGM sensor, as well as taking capillary glucose concentrations every 12 hours for

calibration of the sensor, involved some pain and discomfort for the infants. On

the other hand, adding to the growing body of knowledge regarding CGM use in

preterm infants might lead to more common use of the method in younger infants

in the future, thereby markedly reducing the number of blood sampling

performed, and thereby reducing pain and discomfort in these infants.

The children (and their parents) had no personal benefit of participating in the

studies, apart from the right to receive any information acquired in connection

with the studies. These families have contributed to research that will hopefully

improve future diagnostics and treatment of dysglycemia in preterm infants.

Important to note is that consent was given by the parents/guardians to the

children and not by the children themselves which might raise an ethical

discussion regarding research studies involving participants who were not able to

give consent to their own participation. However, the need to improve the

medical treatment of these infants using evidence from human-based medical

research necessitates performing studies without the consent of the actual

participant who is not able to consent rather with the consent of a

parent/guardian.

Discussion of the results

Prevalence and duration of glucose disturbances

Neonatal hyperglycemia in EPT infants

We observed that 87% of EPT infants experienced hyperglycemia > 8 mmol/L

during the first 28 days of life. A retrospective chart review of 169 ELBW infants

found that 88% had hyperglycemia > 8.3 mmol/L during the first two weeks of

life (28). Another retrospective study including 93 ELBW infants found that more

than 50% had hyperglycemia > 8.3 mmol/L during the first week of life (27). Yet

another retrospective study including 114 ELBW infants observed that

approximately 80% of the infants had hyperglycemia > 8.3 mmol/L during the

first 30 days of life (101). Our study, being the largest cohort of the

aforementioned studies (and the only one focusing on EPT and not ELBW

infants), supports these earlier results.

54

We observed that 51% of EPT infants experienced hyperglycemia > 10 mmol/L

during the first week of life. Furthermore, 53% and 42% of the infants

experienced hyperglycemia > 12 and > 14 mmol/L during the first 28 days of life,

respectively. A steady increase in the hyperglycemia rate up to day 12 of life was

observed, with a slow rate reduction thereafter. These results are in line with

previously published results from a retrospective study that observed persistent

blood glucose concentrations > 11.1 mmol/L that required insulin treatment in

35% of 216 ELBW infants during the first 10 days of life (25). Another

retrospective study including 260 ELBW infants found that 39% had

hyperglycemia > 11.1 mmol/L during the first two weeks of life, and further 46%

had hyperglycemia > 16.7 mmol/L and were treated with insulin (24). Daily

prevalence of hyperglycemia peaked on days 3-4 of life and then peaked again at

days 7-11 of life. These results too are congruent with our findings, apart from our

finding of one peak (and not two peaks) of daily hyperglycemia prevalence.

Glucose disturbances at postmenstrual age 36 weeks in VLBW infants

We have shown that more than half of VLBW infants have protracted

(≥ 30 minutes) episodes of hyperglycemia > 8 mmol/L at PMA 36 weeks, and

many have protracted hypoglycemia as well. These results resemble previously

published rates of glucose disturbances in VLBW infants monitored by CGM at

PMA 32-33 and 36-39 weeks (16, 31). These observed glucose disturbances occur

during a period where glucose concentrations are not usually monitored due to

the perceieved clinical stability of the patients and might imply the need for a

closer monitoring of glucose concentrations in these infants even later during the

admission period.

Risk factors for glucose disturbances

Neonatal hyperglycemia in EPT infants

Parenteral glucose infusion

It was previously shown in ELBW infants that higher glucose infusion rate and

early enhanced parenteral nutrition are risk factors for hyperglycemia (74, 92). A

“before-after” study in VLBW infants found that increasing the carbohydrate

intake during the first week of life (but not later) by 1 g/kg/d was associated with

a 25% increased risk for hyperglycemia (76).

Contrary to our hypothesis and the common clinical belief, our results showed

that parenteral and total carbohydrate intakes were associated with a very small

effect on glucose concentrations - only a 1.6% increase in glucose concentration

for every 1 g/kg/d increase in the parenteral carbohydrate intake. Other studies

have found similar results. Blanco et al. have found that only 21% of

hyperglycemia episodes could be explained by iatrogenic (treatment-induced)

55

causes, while other studies did not find any association between the rate of

glucose infusion and carbohydrate intake and hyperglycemia (16, 18, 22, 23, 28).

Furthermore, we showed that hyperglycemia was common during weeks three

and four of life, a period in which parenteral nutrition intake is usually low,

suggesting that etiologies other than parenteral nutrition might contribute to the

occurrence of hyperglycemia.

The magnitude of the association between the intake and hyperglycemia might

have been attenuated by analyzing the summation of intake during an entire day

versus the highest glucose concentration value on the following day. Another

possible explanation might have been that glucose infusion rates were adjusted

according to previous glucose concentrations, but this potential bias has been

minimized by the pairing described above. Nutritional intakes that were adjusted

according to a certain glucose concentration value could not have affected the

analysis involving the intake on the previous day. Furthermore, we have excluded

days on which the infants received insulin treatment in order to avoid the effect

of insulin on glucose concentrations and to isolate as much as possible the

association between nutritional intakes and glucose concentrations.

Our conclusion is that glucose infusion affects glucose concentrations, but this

effect is smaller than previously thought, at least within the range of glucose

infusion rates given in our cohort, and hyperglycemia in EPT infants does not

seem to be solely the result of the nutrition given to the infant.

Parenteral lipid infusion

We observed that an increase in the parenteral lipid intake was associated with

lower glucose concentrations on the following day. This is in contrast to results

from previous studies that observed that parenteral lipid infusion increased the

risk for hyperglycemia (23, 74, 80, 81). Our findings are supported by the

previously mentioned “before-after” study in VLBW infants by Tottman et al.,

where it was found that increasing the lipid intake by 1 g/kg/d during the first

four weeks of life was associated with a 44% decrease in the odds for

hyperglycemia > 8.5 mmol/L (76). The effect of lipid infusion might depend on

the overall nutritional status of the infant. These findings call for further studies

in order to delineate the effects of parenteral lipid intakes on glucose metabolism.

Parenteral amino acid infusion

In our study, higher parenteral amino acid intake was associated with lower

glucose concentrations. These findings support multiple previous studies,

including a meta-analysis, that reported that parenteral amino acid infusion in

preterm infants have an insulinogenic glucose-decreasing effect, thereby

decreasing the risk for hyperglycemia (76, 122-124).

56

Hyperglycemia at postmenstrual age 36 weeks in VLBW infants

Exposure to amnionitis and glucose disturbances earlier during the admission

period was associated with increased risk for hyperglycemia at PMA 36 weeks.

Male sex and prior hyperglycemia were risk factors for spending longer time in

hyperglycemia at PMA 36 weeks. We observed that infants born SGA spent less

time in hyperglycemia at PMA 36 weeks than non-SGA infants, while

Pertierra-Cortada et al. found that IUGR/SGA was a risk factor for hyperglycemia

(16). This difference in the results might be due to slightly different definitions

for hyperglycemia, as well as different outcome variables (occurrence of

hyperglycemia versus time spent in hyperglycemia).

Hypoglycemia at postmenstrual age 36 weeks in VLBW infants

We found that prior exposure to hypoglycemia was a protective factor, while prior

exposure to hyperglycemia as well as prior insulin treatment were risk factors for

spending longer time in hypoglycemia at PMA 36 weeks. Lower gestational age at

birth, smaller size at birth, as well as being a sicker infant (lower Apgar scores,

longer CPAP treatment), put the infant at risk for spending longer time in

hypoglycemia at PMA 36 weeks. However, Pertierra-Cortada et al. did not find

anthropometric measurements at birth to be associated with hypoglycemia at

term age in VLBW infants (16). This difference might be due to different

definitions and outcome variables, as mentioned previously for hyperglycemia,

but our results might also reflect the well-known fact that these infants have lower

energy reserves and are thus prone for hypoglycemia. Allthough in our study

smaller size at birth (in absolute measurements) was shown to be associated with

spending longer time in hypoglycemia, gestational age-adjusted z-scores of the

respective anthropometric measurements were not associated with such an

outcome, implying that it is prematurity (i.e. gestational age) that explains the

association between smaller size at birth and time spent in hypoglycemia at PMA

36 weeks rather than the actual size of the infant. In multivariable models, we

have observed that lower Apgar scores and prior exposure to hyperglycemia

rather than the size of the infant remained significant risk factors for longer

periods of hypoglycemia at PMA 36 weeks.

Outcomes of neonatal hyperglycemia in EPT infants

Mortality

In our study, a 2.45 times higher 28-day mortality risk was observed in infants

exposed to hyperglycemia > 10 mmol/L during the first 28 days of life. Our results

support previous findings, including hyperglycemia > 8.3 mmol/L during the first

days of life being a risk factor for increased mortality in EPT infants

(93-95). This evidence is strengthened by a series of retrospective studies that

57

showed that increased mortality risk was associated with hyperglycemia of

different definitions in ELBW infants (27, 91, 92).

Blood pressure and a possible early metabolic programming effect of glucose

disturbances

The association between neonatal hyperglycemia and higher BP z-scores at

6.5 years of age described in Paper II is a novel finding. A possible mechanism

could be that hyperglycemia activates the renin-angiotensin system and increases

the setpoint of the vasomotor tone, resulting in increased BP. Increased

vasomotor tone due to transient hyperglycemia was previously shown in young

adults with type 1 (insulin-dependent) diabetes mellitus (114, 115). Furthermore,

in Paper IV we have observed that prior exposure to glucose disturbances was

associated with longer periods with glucose disturbances at PMA 36 weeks. This

leads the author of this thesis to hypothesize that early-life glucose disturbances

and/or their treatment might induce a compensatory programming effect on

metabolic regulation mechanisms later in life. Further studies are needed to

investigate this hypothesis.

Neurodevelopmental outcomes

We did not observe an association between neonatal hyperglycemia and survival

without NDD at 6.5 years of age. These results support previous results published

by Tottman et al., who did not find an association between neonatal

hyperglycemia and survival without NDD at two years of age (112).

We have reported a novel finding – a significant association between neonatal

hyperglycemia and its duration and lower MABC-2 scores at 6.5 years of age. This

association was observed already when a threshold of 8 mmol/L was used for

defining hyperglycemia. This is in contrast to results from the HINT trial, where

no association was found between glucose concentrations and MABC-2 scores

(133). However, all infants in the HINT trial were hyperglycemic to begin with in

order to be included in the study, while our study compared hyperglycemic

infants with non-hyperglycemic infants. Our results are supported by a recently

published meeting abstract by Goldner Pérez et al., where it was reported that

hyperglycemia (8.3-10 mmol/L) during the first week and month of life was

associated with worse motor skills at 18-24 months of age in 68 children born

VLBW (231). Such hyperglycemia during the first week of life was also associated

with language deficits.

As described previously (see Background), hyperglycemia might affect the brain

by inducing oxidative damage, apoptosis and microglial activation, while also

causing increased methylation associated with neurodevelopmental problems in

animal models (106-108). In EPT infants, hyperglycemia was shown to be

associated with white matter reduction (95).

58

Definition of neonatal hyperglycemia – a continuum

While we observed that reduced motor skills at 6.5 years of age were associated

with glucose concentrations > 8 mmol/L, increased BP was only associated with

glucose concentrations > 10 and 12 mmol/L and longer exposure to

hyperglycemia (Table 23). Thus, it seems that different tissues have different

sensitivity to glucose toxicity and are affected to a different extent by glucose

concentrations. It might be so that the brain is affected already at lower glucose

concentrations and after a shorter duration of hyperglycemia, while the blood

vessels and the renin-angiotensin axis (if these actually are the mechanisms

causing the increased BP in these children) are more resilient to the effects of

glucose, and higher glucose concentrations and longer periods of hyperglycemia

are needed to affect them. Thus, in the opinion of the author of this thesis, the

definition of neonatal hyperglycemia should rather be seen as a continuum,

combining both glucose concentration threshold and duration, rather than a

single threshold at a certain timepoint.

Table 23. Significant associations between neonatal hyperglycemia of various definitions

and adverse outcomes at 6.5 years of age in children born EPT.

At least once 2 consecutive days 3 consecutive days

> 8 ↓ Motor skills

> 10 ↓ Motor skills ↓ Motor skills

↑ Blood pressure

> 12 ↓ Motor skills ↑ Blood pressure

↑ Blood pressure

> 14 ↓ Motor skills ↑ Blood pressure

↑ Blood pressure ↓ Motor skills

Glucose concentrations expressed in mmol/L.

Insulin treatment in EPT infants

We described a novel finding of an association between insulin treatment in EPT

infants exposed to hyperglycemia > 10 mmol/L and increased 28- and 70-day

survival. The association remained significant after adjusting for multiple

perinatal comorbidities, and was present despite the fact that the infants treated

with insulin were generally sicker than the non-treated group.

A number of small studies in VLBW and ELBW infants have shown that

continuous intravenous insulin infusion is an effective treatment for

hyperglycemia, allowing an increase in the so-much-needed energy intake, as well

as an increase in weight gain, with hypoglycemia being detected in only as much

as 3% of the blood samples (128, 130-132, 135, 232). An RCT in hyperglycemic

VLBW infants found that a tighter glucose control with insulin infusion enabled

a greater weight gain, and while no difference in mortality was noted, more

59

hypoglycemia was seen (58% versus 27%) (19). No data was published for the

subgroup of ELBW infants in that study. An RCT by Meetze et al. compared

glucose infusion reduction with insulin infusion as treatment for hyperglycemia

in ELBW infants and observed higher energy intakes and no hypoglycemia

episodes in the insulin-treated group (20).

The NIRTURE study, investigating a preventive insulin treatment for

hyperglycemia during the first week of life, did show an increased carbohydrate

intake and decreased weight loss during the first week of life in treated infants

(138). However, a higher 28-day mortality rate was found in the intervention

group, while no differences were found in morbidity and mortality at term age.

Furthermore, treated infants had more hypoglycemia. Interestingly, the

increased hypoglycemia rate was due to increased risk for hypoglycemia among

infants who had a birth weight > 1000 g, while no difference in hypoglycemia

rates was found in ELBW infants.

It is important to have in mind that the NIRTURE study treated all infants with

insulin as long as the infants were not hypoglycemic. Thereby, it is possible that

this might explain the higher hypoglycemia rate among treated infants compared

to other studies. On the other hand, of all the studies mentioned above, the

NIRTURE study was the only one using CGM which might have led to unveiling

of hypoglycemia episodes that otherwise would have gone by undetected.

We did not observe any differences in comorbidities that have previously been

associated with hyperglycemia (NEC, IVH, sepsis) between insulin-treated and

non-insulin-treated infants, but these diagnoses were quite rare in our cohort,

which might have affected the power of our study to detect such differences. Many

of the infants were hypoglycemic on the same day they received insulin infusion,

a somewhat higher rate than the hypoglycemia rate seen among ELBW infants in

the NIRTURE study. However, hypoglycemia occurred mostly during the first

days of life, whereas less than 10% of the insulin-treated infants were

hypoglycemic from day four of life and onwards.

It might be so that ELBW infants are in greater need of insulin treatment than

VLBW infants are. This might explain why treated ELBW infants did not have

more hypoglycemia than control infants in the NIRTURE study, and why our

findings, being the largest study to date in this patient population, show a higher

survival rate in treated infants. However, it is important to remember that insulin

infusion might easily lead to hypoglycemia, and thus glucose concentrations must

be carefully monitored while insulin infusion is ongoing.

60

We did not observe any advantage for insulin treatment regarding NDD, survival

without NDD, FSIQ or MABC-2 scores at 6.5 years of age. This is in contrast with

results from the HINT study, where hyperglycemic infants treated with insulin

and having their blood glucose concentrations in the range 4-6 mmol/L were

more likely to survive without NDD and low FSIQ at seven years of age (133).

However, due to the retrospective design of our studies, no comment can be made

regarding the effectiveness of insulin treatment and if glucose concentrations

were actually effectively controlled by it. Our definition for NDD was slightly

different than in the HINT study (lower FSIQ score, higher visual acuity

threshold, no MABC-2 criterion). Furthermore, we adjusted for factors that were

not regarded as confounders in the HINT study. Also important to have in mind

is that tight glycemic control was not shown to be advantageous in the HINT

study either (no difference in NDD, survival without NDD, FSIQ or

MABC-2 scores). Our results are also supported by previous results by Heald et

al., where insulin treatment was not associated with neurodevelopmental

outcomes at one year of age (139).

Blood pressure at 6.5 years of age in children born EPT

Nutrition and blood pressure

Our results show an association between increased protein intake during the first

eight weeks of life and increased DBP at 6.5 years of age. Our finding that

increased carbohydrate intake during the first eight weeks of life is associated

with increased BP at 6.5 years of age is a novel finding. Studies in animals have

shown that overfeeding is associated with increased BP (196). In preterm infants

though, studies have shown conflicting results regarding protein intake and BP

(195, 203, 204). It might be so that EPT infants are more vulnerable to the

programming effects of early nutrition than more mature infants. A possible

mechanism for that would be that increased nutritional intakes lead to

hyperfiltration in the kidneys with resulting glomerulosclerosis. This might

activate the renin-angiotensin system and result increased BP.

Growth and blood pressure

We did not observe an association between birth weight and BP. Multiple large

studies have previously found that higher birth weight and birth weight z-scores

are associated with increased BP at later age (177-187). Yet these studies did not

focus on preterm infants. Our findings are in line with other studies focusing on

patients born preterm, including a meta-analysis including 1571 adults born

VLBW (167, 188-190, 233).

61

We did not observe an association between weight and length gain during the first

eight weeks of life and BP at 6.5 years of age. Yet we did observe an association

between increased weight gain from week 5 of life to PMA 36 weeks and increased

SBP. A multicenter longitudinal study including 911 preterm infants showed that

for each additional z-score of weight gain between term age and one year of age,

SBP at 6.5 years of age was higher by 0.7 mm Hg (191). An observational cohort

study including 296 children born VLBW found that weight gain velocity between

birth and 3 years of age, but not between birth and 18 months of age, was a

predictor of increased SBP and DBP at 16 years of age (192). Another study did

not find any association between infant weight gain and later BP in preterm borns

(194). The Helsinki cohort study of VLBW-born adults found no relation between

weight gain and linear growth during weeks 1-3 and 4-6 of life and BP in

adulthood (195). It might be so that growth during later months of life has a

programming effect on BP but early postnatal growth does not seem to have such

an effect in preterm infants.

Significance

Our results support the theory that early-life “overnutrition” might have a

programming effect on outcomes later in life, in this case BP. This effect is not

necessarily exerted via growth. As mentioned previously, even a small reduction

of 2 mm Hg in the BP in the individual level might lead to substantial reductions

in the prevalence of hypertension and its complications in the population level

(171). Thus, our results might be of significant clinical importance in the

population level.

62

Main findings

Neonatal hyperglycemia was very common in EPT infants throughout the

first 28 days of life

Protracted (≥ 30 minutes) hyperglycemia and hypoglycemia episodes

were common in VLBW infants at postmenstrual age 36 weeks

Parenteral carbohydrate intake was associated with glucose

concentrations but the effect size was small

Male sex, amnionitis and prior hyperglycemia and hypoglycemia were

risk factors for hyperglycemia at postmenstrual age 36 weeks

Infants who had lower Apgar scores and who were exposed to

hyperglycemia during the admission period spent more time in

hypoglycemia at postmenstrual age 36 weeks

Neonatal hyperglycemia was associated with increased mortality at

28 days of life

Neonatal hyperglycemia was associated with increased blood pressure

and reduced motor skills at 6.5 years of age

Insulin treatment was associated with increased survival at 28 and

70 days of life, but not with neurodevelopmental or cardiovascular

outcomes at 6.5 years of age

Increased protein and carbohydrate intakes during the first eight weeks

of life were associated with higher blood pressure at 6.5 years of age

Growth during the first eight weeks of life was not associated with blood

pressure at 6.5 years of age

63

Clinical implications and future research

The present study shows that neonatal hyperglycemia is a common condition in

preterm infants and not only during the first few weeks of life. Our results stress

the need to continue and monitor preterm infants, especially those with specific

risk factors, for glucose disturbances during later weeks and months of the

admission period. Parenteral nutrition is a powerful tool in the “toolbox” of the

neonatologist but it must be used carefully, in consideration with possible

long-term health outcomes. Further studies are needed to delineate the effect of

parenteral nutrition on glucose concentrations and to clarify its role in the

treatment of hyperglycemia. Our results also suggest that it might be prudent to

closely monitor the blood pressure and motor skills development of these infants

as they grow up.

The LIGHT study will look into the associations between nutrition and glucose

concentrations. Furthermore, blood samples were collected at seven days of age

and at PMA 36 weeks. These will be analyzed for different hormone levels,

exploring the mechanisms contributing to neonatal hyperglycemia in preterm

infants. In addition, the effects of hyperglycemia on the brain will be investigated

using brain magnetic resonance imaging scans that were performed at term age.

A broader use of CGM in the NICU may introduce artificial intelligence-guided

therapy for neonatal hyperglycemia in conjunction with parenteral nutrition.

Studies such as HINT2 and REACT (REAl time ConTinuous glucose monitoring

in the newborn) are already studying the possibility of using CGM systems to

support computer-guided insulin treatment (216, 223, 234). Novel treatments for

neonatal hyperglycemia might also be a promising research direction.

Questions still remain regarding the definition of neonatal hyperglycemia, both

in blood glucose threshold and in duration. The mechanisms causing

hyperglycemia are still poorly understood and the benefit of different treatment

modalities is debated. Further research is needed, both in animal models and in

humans, to clarify the pathophysiology of neonatal hyperglycemia. Large

randomized controlled trials are needed in order to assess different treatment

modalities, all while harnessing the technology of continuous glucose monitoring.

64

Acknowledgements

As most of you readers know, this is usually the first (and somtimes also the last)

part of the thesis that many will ever have had the patience to read through, so I

thought I should thank some important people here.

My main supervisor, Magnus Domellöf – thank you for “taking a chance on

me”, giving me the possibility to rediscover research and fall in love with it. Thank

you for having your door always open for me to come in and ask questions, no

matter how stupid it felt they were; for your constant encouragement and positive

feedback; for supporting me far and beyond what you obliged to do as a

supervisor; for your patience and wise help; and for many conversations about

spex! You gave me the opportunity to come to Sweden to do research, but we both

soon realized that I am “stuck” here for good. I will never forget it.

Elisabeth Stoltz Sjöström, my co-supervisor – thank you for helping with the

data collection, for all those “statistical moments” we’ve shared (see Elisabeth’s

facebook), for always having a practical way of doing things. Thank you also (or

maybe you should thank me..?) for those shopping tours in Venice and Paris.

Never have I ever thought that I would spend over an hour walking around

Sephora on Champs Elisé. Thank you for the yearly dinners at your utekök – I am

not exagerating, that is our culinary highlight of the year.

Thank you both for being the perfect duo as supervisors, a completion to one

another - you know always how to fill in the gaps the other has. Thank you for

always knowing exactly the right thing to say when I needed it the most.

Cornelia Späth, my “older sister” here at the University – thank you for endless

discussions about confounding, for always being there if I needed your help, for

your constant positivity and for being able to talk to you when I needed it.

Stina Alm, my “younger sister” here at the University – Thank you for great

conversations, often not directly work-related. It is so exciting to watch your

journey from a medical student intreseted in some research to a physician and

doctoral student who is already half-way her project.

Ania Chmielewska – my colleague, we have so much in common, from country

of origin (well, my grandparents, but it counts!) to why we came to Sweden, and

now also colleagues at the clinic as well. Always great to have a friend who is going

through kind of the same as you, I am glad to have had you by my side.

65

Andreas, Fredrik A, Ingrid, Thomas, Fredrik S, Johannes and my other

co-authors – thank you for input, wise comments and motivating encouragement.

It was a pleasure cooperating with you, and I am sure that we will continue doing

so in the future.

To all of the colleagues who helped recruiting infants to the LIGHT study as well

as collecting and double-checking data in the EXPRESS study – Elisabeth

Stoltz Sjöström, Johannes van der Bergh, Marie Lindvall, Gunilla

Strinnholm, Yonas Berhane, Estelle Naumburg, Ulrika Ekman,

Kerstin Andersson, Ann-Cathrine Berg, Cecilia Ewald, Sanna

Klevebro, Joesfine Lennhed, Katarina Patriksson, Johan Robinsson,

Christina Fuxin, Birger Malmström, Katarina Strand-Brodd, Magnus

Fredriksson, Jiri Kofron, Per Häggström, Anna Kasemo, Magnus

Ljungcrantz, Lars Alberg, Anna Hedlund, Eva Albinsson as well as my

co-authors - thank you for your time and valuable help. A speciell thanks to the

NICU (barn 4) nursing staff at Umeå for your help with the LIGHT study!

To everyone who has proofread this thesis – Magnus Domellöf, Elisabeth

Stoltz Sjöström, Stina Alm, Maja Nilsson, Torbjörn Lind, Staffan

Berglund, Elena Lundberg, Berit Kriström – thank you for a multitude of

great ideas and a lot of help. You made this thesis better!

My fellow doctoral students in Pediatrics through the years – Frida Karlsson

Videhult, Anna Winberg, Åsa Strinnholm, Josefine Starnberg, Jenny

Alenius Dahlqvist, Cornelia Späth, Lena Hansson, Urban Johansson

Kostenniemi, Maria Björmsjö, Ulrica Johansson, Tove Grip, Stina

Alm, Marie Adamsson, Jonas Österlund, Pontus Challis and Marie

Fredriksson – thank you for great discussions, both on journal clubs and on

fikas, work-related and not work-related. It was a pleasure to share this road with

you.

My roommate – Åsa Strinnholm – after five years of being your other sambo

they have finally succeeded in separating us. I loved every moment of it, my wise,

outspoken, ever-so-positive, trying-to-teach-me-burträsk-dialect friend.

Carina Forslund – thank you for many hours of Melodifestivalen reviews and

for always letting me be myself when talking to you, you know what I mean!

Our administrative staff - Tove Mårs, Anna Englund, Angelica Jonsson,

Mia Olsson, Marléne Engström, Karin Moström and the all-mighty Ulla

Norman – thank you for all your help and support through the years, you are

my superheroes!

66

The team at the research lab and clinic, dietitians, psychologists as well as fellow

researchers – no one named, no one forgotten – thank you for your help and for

many hours of fika and lunch times that distracted me from work just for a little

while.

My colleagues at the pediatric clinic in Umeå – thank you for your understanding

during the last year leading to the dissertation, for supporting me and always

giving me that extra time I needed in order to be ready.

My friends at Medicinarspexet, Umespexarna, Nationskören, but most of all,

Snösvänget – I cannot picture my life here in Umeå without you guys. It was

priceless to know that no matter what there is always that safe environment, same

weekday, same time. Thanks for being an open-minded and accepting group, and

for the confidence and trust you put in me through the years. I have made friends

for life!

The Nilsson, Ågren and Vermeer families – thanks for numerous interesting

questions about my project and for many happy times!

My dear parents, Ofira and Benny, and siblings, Ohad, Hila and Hadas –

where would I have been without you? Certainly not here. Words like support,

encouragement and unconditional love feel almost empty to use when it comes to

describing what you have given me through the years. You were and always are

there for me. This thesis is not only my own accomplishment. It is yours too.

My partner, my other half, Maja – you joined me on this journey after it had

already started, but I cannot imagine how I would have made it to the finish line

without you by my side. You are my source of strength when I need it, you support

me always unconditionally and constantly remind me of what that is important

in life. Thank you for helping me become a better me.

67

Funding

This project was funded by:

The Swedish Research Council

Västerbotten County Council

Stockholm County Council

Skåne County Council

J C Kempe Memorial Stipend Foundation

HRH Crown Princess Lovisa’s Association for Child Medical Care

The Swedish Heart-Lung Foundation

The Childhood Foundation of the Swedish Order of Freemasons

Stockholm Odd Fellow Foundation

Lilla Barnets fond

Stiftelsen Samariten

Oskarfonden

Arnerska Research Fund

68

References

1. Swedish Neonatal Quality Register. Yearly report 2017 [Internet]. Stockholm: Swedish Neonatal Quality Register; 2018. [cited 2020 Apr 14]. Available from: https://www.medscinet.com/pnq/Uploads/SNQ%20%C3%85rsrapport%202017_180909.pdf.

2. Norman M, Hallberg B, Abrahamsson T, Björklund LJ, Domellöf M, Farooqi A, et al. Association Between Year of Birth and 1-Year Survival Among Extremely Preterm Infants in Sweden During 2004-2007 and 2014-2016. JAMA. 2019;321(12):1188-99.

3. Serenius F, Ewald U, Farooqi A, Fellman V, Hafström M, Hellgren K, et al. Neurodevelopmental Outcomes Among Extremely Preterm Infants 6.5 Years After Active Perinatal Care in Sweden. JAMA Pediatr. 2016;170(10):954-63.

4. EXPRESS Group. Incidence of and risk factors for neonatal morbidity after active perinatal care: extremely preterm infants study in Sweden (EXPRESS). Acta Paediatr. 2010;99(7):978-92.

5. Bolk J, Farooqi A, Hafström M, Ådén U, Serenius F. Developmental Coordination Disorder and Its Association With Developmental Comorbidities at 6.5 Years in Apparently Healthy Children Born Extremely Preterm. JAMA Pediatr. 2018;172(8):765-74.

6. Bolk J, Fredriksson Kaul Y, Hellström-Westas L, Stjernqvist K, Padilla N, Serenius F, et al. National population-based cohort study found that visual-motor integration was commonly affected in extremely preterm born children at six-and-a-half years. Acta Paediatr. 2018;107(5):831-7.

7. Holsti A, Serenius F, Farooqi A. Impact of major neonatal morbidities on adolescents born at 23-25 weeks of gestation. Acta Paediatr. 2018;107(11):1893-901.

8. Joosten K, Embleton N, Yan W, Senterre T, ESPGHAN/ESPEN/ESPR/CSPEN working group on pediatric parenteral nutrition. ESPGHAN/ESPEN/ESPR/CSPEN guidelines on pediatric parenteral nutrition: Energy. Clin Nutr. 2018;37(6 Pt B):2309-14.

9. Stoltz Sjöström E, Öhlund I, Ahlsson F, Engström E, Fellman V, Hellström A, et al. Nutrient intakes independently affect growth in extremely preterm infants: results from a population-based study. Acta Paediatr. 2013;102(11):1067-74.

10. Thureen PJ. Early aggressive nutrition in the neonate. Pediatr Rev. 1999;20(9):e45-55.

11. Tyrala EE, Chen X, Boden G. Glucose metabolism in the infant weighing less than 1100 grams. J Pediatr. 1994;125(2):283-7.

12. Kaiser JR, Bai S, Rozance PJ. Newborn Plasma Glucose Concentration Nadirs by Gestational-Age Group. Neonatology. 2018;113(4):353-9.

13. Mesotten D, Joosten K, van Kempen A, Verbruggen S, ESPGHAN/ESPEN/ESPR/CSPEN working group on pediatric parenteral nutrition. ESPGHAN/ESPEN/ESPR/CSPEN guidelines on pediatric parenteral nutrition: Carbohydrates. Clin Nutr. 2018;37(6 Pt B):2337-43.

69

14. Diwakar KK, Sasidhar MV. Plasma glucose levels in term infants who are appropriate size for gestation and exclusively breast fed. Arch Dis Child Fetal Neonatal Ed. 2002;87(1):F46-8.

15. Alsweiler JM, Kuschel CA, Bloomfield FH. Survey of the management of neonatal hyperglycaemia in Australasia. J Paediatr Child Health. 2007;43(9):632-5.

16. Pertierra-Cortada A, Ramon-Krauel M, Iriondo-Sanz M, Iglesias-Platas I. Instability of glucose values in very preterm babies at term postmenstrual age. J Pediatr. 2014;165(6):1146-53 e2.

17. Scheurer JM, Gray HL, Demerath EW, Rao R, Ramel SE. Diminished growth and lower adiposity in hyperglycemic very low birth weight neonates at 4 months corrected age. J Perinatol. 2016;36(2):145-50.

18. Sabzehei MK, Afjeh SA, Shakiba M, Alizadeh P, Shamshiri AR, Esmaili F. Hyperglycemia in VLBW infants; incidence, risk factors and outcome. Arch Iran Med. 2014;17(6):429-34.

19. Alsweiler JM, Harding JE, Bloomfield FH. Tight glycemic control with insulin in hyperglycemic preterm babies: a randomized controlled trial. Pediatrics. 2012;129(4):639-47.

20. Meetze W, Bowsher R, Compton J, Moorehead H. Hyperglycemia in extremely- low-birth-weight infants. Biol Neonate. 1998;74(3):214-21.

21. Yoon JY, Chung HR, Choi CW, Yang SW, Kim BI, Shin CH. Blood glucose levels within 7 days after birth in preterm infants according to gestational age. Ann Pediatr Endocrinol Metab. 2015;20(4):213-9.

22. Fernández-Martínez MDM, Gómez-Llorente JL, Momblán-Cabo J, Martin-González M, Calvo-Bonachera M, Olvera-Porcel M, et al. Monitoring the incidence, duration and distribution of hyperglycaemia in very-low-birth-weight newborns and identifying associated factors. J Perinat Med. 2020.

23. Beardsall K, Vanhaesebrouck S, Ogilvy-Stuart AL, Vanhole C, Palmer CR, Ong K, et al. Prevalence and determinants of hyperglycemia in very low birth weight infants: cohort analyses of the NIRTURE study. J Pediatr. 2010;157(5):715-9 e1-3.

24. Yoo HS, Ahn SY, Lee MS, Han YM, Sung SI, Chang YS, et al. Permissive hyperglycemia in extremely low birth weight infants. J Korean Med Sci. 2013;28(3):450-60.

25. Bermick J, Dechert RE, Sarkar S. Does hyperglycemia in hypernatremic preterm infants increase the risk of intraventricular hemorrhage? J Perinatol. 2016;36(9):729-32.

26. Arsenault D, Brenn M, Kim S, Gura K, Compher C, Simpser E, et al. A.S.P.E.N. Clinical Guidelines: hyperglycemia and hypoglycemia in the neonate receiving parenteral nutrition. JPEN J Parenter Enteral Nutr. 2012;36(1):81-95.

27. Hays SP, Smith EO, Sunehag AL. Hyperglycemia is a risk factor for early death and morbidity in extremely low birth-weight infants. Pediatrics. 2006;118(5):1811-8.

28. Blanco CL, Baillargeon JG, Morrison RL, Gong AK. Hyperglycemia in extremely low birth weight infants in a predominantly Hispanic population and related morbidities. J Perinatol. 2006;26(12):737-41.

70

29. Szymonska I, Jagla M, Starzec K, Hrnciar K, Kwinta P. The Incidence of Hyperglycaemia in Very Low Birth Weight Preterm Newborns. Results of a Continuous Glucose Monitoring Study--Preliminary Report. Dev Period Med. 2015;19(3 Pt 1):305-12.

30. Iglesias Platas I, Thio Lluch M, Pociello Alminana N, Morillo Palomo A, Iriondo Sanz M, Krauel Vidal X. Continuous glucose monitoring in infants of very low birth weight. Neonatology. 2009;95(3):217-23.

31. Mola-Schenzle E, Staffler A, Klemme M, Pellegrini F, Molinaro G, Parhofer KG, et al. Clinically stable very low birthweight infants are at risk for recurrent tissue glucose fluctuations even after fully established enteral nutrition. Arch Dis Child Fetal Neonatal Ed. 2015;100(2):F126-31.

32. James EJ, Raye JR, Gresham EL, Makowski EL, Meschia G, Battaglia FC. Fetal oxygen consumption, carbon dioxide production, and glucose uptake in a chronic sheep preparation. Pediatrics. 1972;50(3):361-71.

33. Beardsall K, Dunger D. Insulin therapy in preterm newborns. Early Hum Dev. 2008;84(12):839-42.

34. Bereket A, Lang CH, Blethen SL, Gelato MC, Fan J, Frost RA, et al. Effect of insulin on the insulin-like growth factor system in children with new-onset insulin-dependent diabetes mellitus. J Clin Endocrinol Metab. 1995;80(4):1312-7.

35. Cornblath M, Reisner SH. Blood glucose in the neonate and its clinical significance. N Engl J Med. 1965;273(7):378-81.

36. Sperling MA, DeLamater PV, Phelps D, Fiser RH, Oh W, Fisher DA. Spontaneous and amino acid-stimulated glucagon secretion in the immediate postnatal period. Relation to glucose and insulin. J Clin Invest. 1974;53(4):1159-66.

37. Kraiem Z, Sack J, Brish M. Serum cortisol levels: the first 10 days in full-term and preterm infants. Isr J Med Sci. 1985;21(2):170-2.

38. Burchell A, Gibb L, Waddell ID, Giles M, Hume R. The ontogeny of human hepatic microsomal glucose-6-phosphatase proteins. Clin Chem. 1990;36(9):1633-7.

39. Lucas A, Bloom SR, Aynsley-Green A. Metabolic and endocrine events at the time of the first feed of human milk in preterm and term infants. Arch Dis Child. 1978;53(9):731-6.

40. Hoseth E, Joergensen A, Ebbesen F, Moeller M. Blood glucose levels in a population of healthy, breast fed, term infants of appropriate size for gestational age. Arch Dis Child Fetal Neonatal Ed. 2000;83(2):F117-9.

41. Mitanchez D. Glucose regulation in preterm newborn infants. Horm Res. 2007;68(6):265-71.

42. Postic C, Leturque A, Printz RL, Maulard P, Loizeau M, Granner DK, et al. Development and regulation of glucose transporter and hexokinase expression in rat. Am J Physiol. 1994;266(4 Pt 1):E548-59.

43. Pollak A, Cowett RM, Schwartz R, Oh W. Glucose disposal in low-birth-weight infants during steady-state hyperglycemia: effects of exogenous insulin administration. Pediatrics. 1978;61(4):546-9.

44. Blanco CL, McGill-Vargas LL, Gastaldelli A, Seidner SR, McCurnin DC, Leland MM, et al. Peripheral insulin resistance and impaired insulin signaling contribute to abnormal glucose metabolism in preterm baboons. Endocrinology. 2015;156(3):813-23.

71

45. Ogilvy-Stuart AL, Beardsall K. Management of hyperglycaemia in the preterm infant. Arch Dis Child Fetal Neonatal Ed. 2010;95(2):F126-31.

46. Verbruggen SC, Coss-Bu J, Wu M, Schierbeek H, Joosten KF, Dhar A, et al. Current recommended parenteral protein intakes do not support protein synthesis in critically ill septic, insulin-resistant adolescents with tight glucose control. Crit Care Med. 2011;39(11):2518-25.

47. Hofman PL, Regan F, Jackson WE, Jefferies C, Knight DB, Robinson EM, et al. Premature birth and later insulin resistance. N Engl J Med. 2004;351(21):2179-86.

48. Hawdon JM, Hubbard M, Hales CN, Clark PM. Use of specific immunoradiometric assay to determine preterm neonatal insulin-glucose relations. Arch Dis Child Fetal Neonatal Ed. 1995;73(3):F166-9.

49. Salis ER, Reith DM, Wheeler BJ, Broadbent RS, Medlicott NJ. Insulin resistance, glucagon-like peptide-1 and factors influencing glucose homeostasis in neonates. Arch Dis Child Fetal Neonatal Ed. 2017;102(2):F162-F6.

50. Ahmad A, Srikantiah RM, Yadav C, Agarwal A, Manjrekar PA, Hegde A. Cord Blood Levels of Insulin, Cortisol and HOMA2-IR in Very Preterm, Late Preterm and Term Newborns. J Clin Diagn Res. 2016;10(5):BC05-8.

51. Bagnoli F, Vodo F, Vodo S, Conte ML, Tomasini B, Vodo Z, et al. Glucagon and insulin cord blood levels in very preterm, late preterm and full-term infants. J Pediatr Endocrinol Metab. 2014;27(5-6):419-23.

52. Goldman SL, Hirata T. Attenuated response to insulin in very low birthweight infants. Pediatr Res. 1980;14(1):50-3.

53. Rao S, Tan KL. Oral glucose tolerance tests in newborn infants. Ann Acad Med Singapore. 1983;12(3):457-61.

54. Beardsall K, Vanhaesebrouck S, Frystyk J, Ogilvy-Stuart AL, Vanhole C, van Weissenbruch M, et al. Relationship between insulin-like growth factor I levels, early insulin treatment, and clinical outcomes of very low birth weight infants. J Pediatr. 2014;164(5):1038-44 e1.

55. Cekmez F, Canpolat FE, Pirgon O, Aydemir G, Tanju IA, Genc FA, et al. Adiponectin and visfatin levels in extremely low birth weight infants; they are also at risk for insulin resistance. Eur Rev Med Pharmacol Sci. 2013;17(4):501-6.

56. Tsubahara M, Shoji H, Mori M, Matsunaga N, Ikeno M, Hisata K, et al. Glucose metabolism soon after birth in very premature infants with small- and appropriate-for-gestational-age birth weights. Early Hum Dev. 2012;88(9):735-8.

57. Bansal A, Bloomfield FH, Connor KL, Dragunow M, Thorstensen EB, Oliver MH, et al. Glucocorticoid-Induced Preterm Birth and Neonatal Hyperglycemia Alter Ovine beta-Cell Development. Endocrinology. 2015;156(10):3763-76.

58. Preissig CM, Rigby MR. Hyperglycaemia results from beta-cell dysfunction in critically ill children with respiratory and cardiovascular failure: a prospective observational study. Crit Care. 2009;13(1):R27.

59. Dickson JL, Chase JG, Pretty CG, Gunn CA, Alsweiler JM. Hyperglycaemic Preterm Babies Have Sex Differences in Insulin Secretion. Neonatology. 2015;108(2):93-8.

72

60. Mitanchez-Mokhtari D, Lahlou N, Kieffer F, Magny JF, Roger M, Voyer M. Both relative insulin resistance and defective islet beta-cell processing of proinsulin are responsible for transient hyperglycemia in extremely preterm infants. Pediatrics. 2004;113(3 Pt 1):537-41.

61. Asplund K. Effects of postnatal feeding on the functional maturation of pancreatic islet B-cells of neonatal rats. Diabetologia. 1972;8(3):153-9.

62. Burchell A, McGeechan A, Hume R. Therapeutic insulin and hepatic glucose-6-phosphatase activity in preterm infants. Arch Dis Child Fetal Neonatal Ed. 2000;82(3):F228-32.

63. Farrag HM, Nawrath LM, Healey JE, Dorcus EJ, Rapoza RE, Oh W, et al. Persistent glucose production and greater peripheral sensitivity to insulin in the neonate vs. the adult. Am J Physiol. 1997;272(1 Pt 1):E86-93.

64. Cowett RM, Oh W, Schwartz R. Persistent glucose production during glucose infusion in the neonate. J Clin Invest. 1983;71(3):467-75.

65. Sunehag A, Gustafsson J, Ewald U. Very immature infants (< or = 30 Wk) respond to glucose infusion with incomplete suppression of glucose production. Pediatr Res. 1994;36(4):550-5.

66. Chacko SK, Sunehag AL. Gluconeogenesis continues in premature infants receiving total parenteral nutrition. Arch Dis Child Fetal Neonatal Ed. 2010;95(6):F413-8.

67. Cowett RM, Andersen GE, Maguire CA, Oh W. Ontogeny of glucose homeostasis in low birth weight infants. J Pediatr. 1988;112(3):462-5.

68. Chacko SK, Ordonez J, Sauer PJ, Sunehag AL. Gluconeogenesis is not regulated by either glucose or insulin in extremely low birth weight infants receiving total parenteral nutrition. J Pediatr. 2011;158(6):891-6.

69. Hume R, Burchell A. Abnormal expression of glucose-6-phosphatase in preterm infants. Arch Dis Child. 1993;68(2):202-4.

70. Doerr HG, Sippell WG, Versmold HT, Bidlingmaier F, Knorr D. Plasma mineralocorticoids, glucocorticoids, and progestins in premature infants: longitudinal study during the first week of life. Pediatr Res. 1988;23(5):525-9.

71. Scott SM, Watterberg KL. Effect of gestational age, postnatal age, and illness on plasma cortisol concentrations in premature infants. Pediatr Res. 1995;37(1):112-6.

72. Turina M, Fry DE, Polk HC, Jr. Acute hyperglycemia and the innate immune system: clinical, cellular, and molecular aspects. Crit Care Med. 2005;33(7):1624-33.

73. Limesand SW, Rozance PJ, Zerbe GO, Hutton JC, Hay WW, Jr. Attenuated insulin release and storage in fetal sheep pancreatic islets with intrauterine growth restriction. Endocrinology. 2006;147(3):1488-97.

74. Stensvold HJ, Lang AM, Strommen K, Abrahamsen TG, Ogland B, Pripp AH, et al. Strictly controlled glucose infusion rates are associated with a reduced risk of hyperglycaemia in extremely low birth weight preterm infants. Acta Paediatr. 2018;107(3):442-9.

75. Hawdon JM, Aynsley-Green A, Bartlett K, Ward Platt MP. The role of pancreatic insulin secretion in neonatal glucoregulation. II. Infants with disordered blood glucose homoeostasis. Arch Dis Child. 1993;68(3 Spec No):280-5.

73

76. Tottman AC, Bloomfield FH, Cormack BE, Harding JE, Mohd Slim MA, Weston AF, et al. Relationships Between Early Nutrition and Blood Glucose Concentrations in Very Preterm Infants. J Pediatr Gastroenterol Nutr. 2018;66(6):960-6.

77. Anagnostakis D, Matsaniotis N. Hyperglycemia in very low birth weight infants: a common iatrogenic problem. Padiatr Padol. 1982;17(3):585-90.

78. Dweck HS, Cassady G. Glucose intolerance in infants of very low birth weight. I. Incidence of hyperglycemia in infants of birth weights 1,100 grams or less. Pediatrics. 1974;53(2):189-95.

79. Cowett RM, Oh W, Pollak A, Schwartz R, Stonestreet BS. Glucose disposal of low birth weight infants: steady state hyperglycemia produced by constant intravenous glucose infusion. Pediatrics. 1979;63(3):389-96.

80. Sann L, Divry P, Lasne Y, Ruitton A. Effect of oral lipid administration on glucose homeostasis in small-for-gestational-age infants. Acta Paediatr Scand. 1982;71(6):923-7.

81. Savich RD, Finley SL, Ogata ES. Intravenous lipid and amino acids briskly increase plasma glucose concentrations in small premature infants. Am J Perinatol. 1988;5(3):201-5.

82. Koivisto M, Peltoniemi OM, Saarela T, Tammela O, Jouppila P, Hallman M. Blood glucose level in preterm infants after antenatal exposure to glucocorticoid. Acta Paediatr. 2007;96(5):664-8.

83. Rohr SB, Sauer H, Gottschling S, Abdul-Khaliq H, Gortner L, Nunold H, et al. Non-neurological, steroid-related adverse events in very low birth weight infants: a prospective audit. Swiss Med Wkly. 2014;144:w13954.

84. Doyle LW, Cheong JL, Ehrenkranz RA, Halliday HL. Early (< 8 days) systemic postnatal corticosteroids for prevention of bronchopulmonary dysplasia in preterm infants. Cochrane Database Syst Rev. 2017;10:CD001146.

85. Doyle LW, Cheong JL, Ehrenkranz RA, Halliday HL. Late (> 7 days) systemic postnatal corticosteroids for prevention of bronchopulmonary dysplasia in preterm infants. Cochrane Database Syst Rev. 2017;10:CD001145.

86. Shah SS, Ohlsson A, Halliday HL, Shah VS. Inhaled versus systemic corticosteroids for the treatment of bronchopulmonary dysplasia in ventilated very low birth weight preterm infants. Cochrane Database Syst Rev. 2017;10:CD002057.

87. Shah SS, Ohlsson A, Halliday HL, Shah VS. Inhaled versus systemic corticosteroids for preventing bronchopulmonary dysplasia in ventilated very low birth weight preterm neonates. Cochrane Database Syst Rev. 2017;10:CD002058.

88. Hemachandra AH, Cowett RM. Neonatal Hyperglycemia. Pediatrics in Review. 1999;20(7):e 16-e 24.

89. Dreyfus L, Fischer Fumeaux CJ, Remontet L, Essomo Megnier Mbo Owono MC, Laborie S, Maucort-Boulch D, et al. Low phosphatemia in extremely low birth weight neonates: A risk factor for hyperglycemia? Clin Nutr. 2016;35(5):1059-65.

90. Rubio-Cabezas O, Ellard S. Diabetes mellitus in neonates and infants: genetic heterogeneity, clinical approach to diagnosis, and therapeutic options. Horm Res Paediatr. 2013;80(3):137-46.

74

91. Kao LS, Morris BH, Lally KP, Stewart CD, Huseby V, Kennedy KA. Hyperglycemia and morbidity and mortality in extremely low birth weight infants. J Perinatol. 2006;26(12):730-6.

92. Stensvold HJ, Strommen K, Lang AM, Abrahamsen TG, Steen EK, Pripp AH, et al. Early Enhanced Parenteral Nutrition, Hyperglycemia, and Death Among Extremely Low-Birth-Weight Infants. JAMA Pediatr. 2015;169(11):1003-10.

93. van der Lugt NM, Smits-Wintjens VE, van Zwieten PH, Walther FJ. Short and long term outcome of neonatal hyperglycemia in very preterm infants: a retrospective follow-up study. BMC Pediatr. 2010;10:52.

94. Heimann K, Peschgens T, Kwiecien R, Stanzel S, Hoernchen H, Merz U. Are recurrent hyperglycemic episodes and median blood glucose level a prognostic factor for increased morbidity and mortality in premature infants </=1500 g? J Perinat Med. 2007;35(3):245-8.

95. Alexandrou G, Skiöld B, Karlén J, Tessma MK, Norman M, Ådén U, et al. Early hyperglycemia is a risk factor for death and white matter reduction in preterm infants. Pediatrics. 2010;125(3):e584-91.

96. Ertl T, Gyarmati J, Gaal V, Szabo I. Relationship between hyperglycemia and retinopathy of prematurity in very low birth weight infants. Biol Neonate. 2006;89(1):56-9.

97. Mohsen L, Abou-Alam M, El-Dib M, Labib M, Elsada M, Aly H. A prospective study on hyperglycemia and retinopathy of prematurity. J Perinatol. 2014;34(6):453-7.

98. Kaempf JW, Kaempf AJ, Wu Y, Stawarz M, Niemeyer J, Grunkemeier G. Hyperglycemia, insulin and slower growth velocity may increase the risk of retinopathy of prematurity. J Perinatol. 2011;31(4):251-7.

99. Ahmadpour-Kacho M, Motlagh AJ, Rasoulinejad SA, Jahangir T, Bijani A, Pasha YZ. Correlation between hyperglycemia and retinopathy of prematurity. Pediatr Int. 2014;56(5):726-30.

100. Garg R, Agthe AG, Donohue PK, Lehmann CU. Hyperglycemia and retinopathy of prematurity in very low birth weight infants. J Perinatol. 2003;23(3):186-94.

101. Chavez-Valdez R, McGowan J, Cannon E, Lehmann CU. Contribution of early glycemic status in the development of severe retinopathy of prematurity in a cohort of ELBW infants. J Perinatol. 2011;31(12):749-56.

102. Kermorvant-Duchemin E, Pinel AC, Lavalette S, Lenne D, Raoul W, Calippe B, et al. Neonatal hyperglycemia inhibits angiogenesis and induces inflammation and neuronal degeneration in the retina. PLoS One. 2013;8(11):e79545.

103. Ramel SE, Long JD, Gray H, Durrwachter-Erno K, Demerath EW, Rao R. Neonatal hyperglycemia and diminished long-term growth in very low birth weight preterm infants. J Perinatol. 2013;33(11):882-6.

104. Auerbach A, Eventov-Friedman S, Arad I, Peleg O, Bdolah-Abram T, Bar-Oz B, et al. Long duration of hyperglycemia in the first 96 hours of life is associated with severe intraventricular hemorrhage in preterm infants. J Pediatr. 2013;163(2):388-93.

105. Tayman C, Yis U, Hirfanoglu I, Oztekin O, Goktas G, Bilgin BC. Effects of hyperglycemia on the developing brain in newborns. Pediatr Neurol. 2014;51(2):239-45.

75

106. Satrom KM, Ennis K, Sweis BM, Matveeva TM, Chen J, Hanson L, et al. Neonatal hyperglycemia induces CXCL10/CXCR3 signaling and microglial activation and impairs long-term synaptogenesis in the hippocampus and alters behavior in rats. J Neuroinflammation. 2018;15(1):82.

107. Cartier J, Piyasena C, Sparrow SA, Boardman JP, Drake AJ. Alterations in glucose concentrations affect DNA methylation at Lrg1 in an ex vivo rat cortical slice model of preterm brain injury. Eur J Neurosci. 2018;47(5):380-7.

108. Gisslen T, Ennis K, Bhandari V, Rao R. Recurrent hypoinsulinemic hyperglycemia in neonatal rats increases PARP-1 and NF-kappaB expression and leads to microglial activation in the cerebral cortex. Pediatr Res. 2015;78(5):513-9.

109. Schumacher EM, Larsson PG, Pripp AH, Stiris TA. The effect of blood glucose and pCO2 on spectral EEG of premature infants during the first three days of life. Neonatology. 2014;105(4):297-305.

110. Basu SK, Kaiser JR, Guffey D, Minard CG, Guillet R, Gunn AJ, et al. Hypoglycaemia and hyperglycaemia are associated with unfavourable outcome in infants with hypoxic ischaemic encephalopathy: a post hoc analysis of the CoolCap Study. Arch Dis Child Fetal Neonatal Ed. 2016;101(2):F149-55.

111. Spies EE, Lababidi SL, McBride MC. Early hyperglycemia is associated with poor gross motor outcome in asphyxiated term newborns. Pediatr Neurol. 2014;50(6):586-90.

112. Tottman AC, Alsweiler JM, Bloomfield FH, Pan M, Harding JE. Relationship between Measures of Neonatal Glycemia, Neonatal Illness, and 2-Year Outcomes in Very Preterm Infants. J Pediatr. 2017;188:115-21.

113. Vyssoulis G, Pietri P, Vlachopoulos C, Alexopoulos N, Kyvelou SM, Terentes-Printzios D, et al. Early adverse effect of abnormal glucose metabolism on arterial stiffness in drug naive hypertensive patients. Diab Vasc Dis Res. 2012;9(1):18-24.

114. Miller JA, Floras JS, Zinman B, Skorecki KL, Logan AG. Effect of hyperglycaemia on arterial pressure, plasma renin activity and renal function in early diabetes. Clin Sci (Lond). 1996;90(3):189-95.

115. Miller JA. Impact of hyperglycemia on the renin angiotensin system in early human type 1 diabetes mellitus. J Am Soc Nephrol. 1999;10(8):1778-85.

116. Villafana S, Huang F, Hong E. Role of the sympathetic and renin angiotensin systems in the glucose-induced increase of blood pressure in rats. Eur J Pharmacol. 2004;506(2):143-50.

117. Callaway DA, McGill-Vargas LL, Quinn A, Jordan JL, Winter LA, Anzueto D, et al. Prematurity disrupts glomeruli development, whereas prematurity and hyperglycemia lead to altered nephron maturation and increased oxidative stress in newborn baboons. Pediatr Res. 2018;83(3):702-11.

118. Morgan C. The potential risks and benefits of insulin treatment in hyperglycaemic preterm neonates. Early Hum Dev. 2015;91(11):655-9.

76

119. Bottino M, Cowett RM, Sinclair JC. Interventions for treatment of neonatal hyperglycemia in very low birth weight infants. Cochrane Database Syst Rev. 2011(10):CD007453.

120. Sinclair JC, Bottino M, Cowett RM. Interventions for prevention of neonatal hyperglycemia in very low birth weight infants. Cochrane Database Syst Rev. 2011(10):CD007615.

121. Burgess L, Morgan C, Mayes K, Tan M. Plasma arginine levels and blood glucose control in very preterm infants receiving 2 different parenteral nutrition regimens. JPEN J Parenter Enteral Nutr. 2014;38(2):243-53.

122. Andronikou S, Hanning I. Parenteral nutrition effect on serum insulin in the preterm infant. Pediatrics. 1987;80(5):693-7.

123. Osborn DA, Schindler T, Jones LJ, Sinn JK, Bolisetty S. Higher versus lower amino acid intake in parenteral nutrition for newborn infants. Cochrane Database Syst Rev. 2018;3:CD005949.

124. Thureen PJ, Melara D, Fennessey PV, Hay WW, Jr. Effect of low versus high intravenous amino acid intake on very low birth weight infants in the early neonatal period. Pediatr Res. 2003;53(1):24-32.

125. Burattini I, Bellagamba MP, Spagnoli C, D'Ascenzo R, Mazzoni N, Peretti A, et al. Targeting 2.5 versus 4 g/kg/day of amino acids for extremely low birth weight infants: a randomized clinical trial. J Pediatr. 2013;163(5):1278-82 e1.

126. Mahaveer A, Grime C, Morgan C. Increasing early protein intake is associated with a reduction in insulin-treated hyperglycemia in very preterm infants. Nutr Clin Pract. 2012;27(3):399-405.

127. Gilbertson N, Kovar IZ, Cox DJ, Crowe L, Palmer NT. Introduction of intravenous lipid administration on the first day of life in the very low birth weight neonate. J Pediatr. 1991;119(4):615-23.

128. Thabet F, Bourgeois J, Guy B, Putet G. Continuous insulin infusion in hyperglycaemic very-low-birth-weight infants receiving parenteral nutrition. Clin Nutr. 2003;22(6):545-7.

129. Pappoe TA, Wu SY, Pyati S. A randomized controlled trial comparing an aggressive and a conventional parenteral nutrition regimen in very low birth weight infants. Journal of Neonatal-Perinatal Medicine. 2009;2(3):149-56.

130. Vaucher YE, Walson PD, Morrow G, 3rd. Continuous insulin infusion in hyperglycemic, very low birth weight infants. J Pediatr Gastroenterol Nutr. 1982;1(2):211-7.

131. Ostertag SG, Jovanovic L, Lewis B, Auld PA. Insulin pump therapy in the very low birth weight infant. Pediatrics. 1986;78(4):625-30.

132. Heron P, Bourchier D. Insulin infusions in infants of birthweight less than 1250 g and with glucose intolerance. Aust Paediatr J. 1988;24(6):362-5.

133. Tottman AC, Alsweiler JM, Bloomfield FH, Gamble G, Jiang Y, Leung M, et al. Long-Term Outcomes of Hyperglycemic Preterm Infants Randomized to Tight Glycemic Control. J Pediatr. 2018;193:68-75 e1.

134. Ng SM, May JE, Emmerson AJ. Continuous insulin infusion in hyperglycaemic extremely-low- birth-weight neonates. Biol Neonate. 2005;87(4):269-72.

77

135. Collins JW, Jr., Hoppe M, Brown K, Edidin DV, Padbury J, Ogata ES. A controlled trial of insulin infusion and parenteral nutrition in extremely low birth weight infants with glucose intolerance. J Pediatr. 1991;118(6):921-7.

136. Binder ND, Raschko PK, Benda GI, Reynolds JW. Insulin infusion with parenteral nutrition in extremely low birth weight infants with hyperglycemia. J Pediatr. 1989;114(2):273-80.

137. Beardsall K, Ogilvy-Stuart AL, Frystyk J, Chen JW, Thompson M, Ahluwalia J, et al. Early elective insulin therapy can reduce hyperglycemia and increase insulin-like growth factor-I levels in very low birth weight infants. J Pediatr. 2007;151(6):611-7, 7 e1.

138. Beardsall K, Vanhaesebrouck S, Ogilvy-Stuart AL, Vanhole C, Palmer CR, van Weissenbruch M, et al. Early insulin therapy in very-low-birth-weight infants. N Engl J Med. 2008;359(18):1873-84.

139. Heald A, Abdel-Latif ME, Kent AL. Insulin infusion for hyperglycaemia in very preterm infants appears safe with no effect on morbidity, mortality and long-term neurodevelopmental outcome. J Matern Fetal Neonatal Med. 2012;25(11):2415-8.

140. Poindexter BB, Karn CA, Denne SC. Exogenous insulin reduces proteolysis and protein synthesis in extremely low birth weight infants. J Pediatr. 1998;132(6):948-53.

141. Petty C, Cunningham NL. Insulin adsorption by glass infusion bottles, polyvinylchloride infusion containers, and intravenous tubing. Anesthesiology. 1974;40(4):400-4.

142. Fuloria M, Friedberg MA, DuRant RH, Aschner JL. Effect of flow rate and insulin priming on the recovery of insulin from microbore infusion tubing. Pediatrics. 1998;102(6):1401-6.

143. Koh TH, Eyre JA, Aynsley-Green A. Neonatal hypoglycaemia--the controversy regarding definition. Arch Dis Child. 1988;63(11):1386-8.

144. Alkalay AL, Sarnat HB, Flores-Sarnat L, Elashoff JD, Farber SJ, Simmons CF. Population meta-analysis of low plasma glucose thresholds in full-term normal newborns. Am J Perinatol. 2006;23(2):115-9.

145. Committee on Fetus and Newborn, Adamkin DH. Postnatal glucose homeostasis in late-preterm and term infants. Pediatrics. 2011;127(3):575-9.

146. Harris D, Weston P, Harding J. Glucose Profiles in Healthy Term Babies in the First Five Days: The Glucose in Well Babies (GLOW) Study. Journal of Paediatrics and Child Health. 2019;55:23.

147. Fernandez Martinez MDM, Llorente JLG, de Cabo JM, Lopez MAV, Porcel M, Rubio JDD, et al. Monitoring the Frequency and Duration of Hypoglycemia in Preterm Infants and Identifying Associated Factors. Fetal Pediatr Pathol. 2019:1-11.

148. Lucas A, Morley R, Cole TJ. Adverse neurodevelopmental outcome of moderate neonatal hypoglycaemia. BMJ. 1988;297(6659):1304-8.

149. Staffler A, Klemme M, Mola-Schenzle E, Mittal R, Schulze A, Flemmer AW. Very low birth weight preterm infants are at risk for hypoglycemia once on total enteral nutrition. J Matern Fetal Neonatal Med. 2013;26(13):1337-41.

78

150. Ward Platt M, Deshpande S. Metabolic adaptation at birth. Semin Fetal Neonatal Med. 2005;10(4):341-50.

151. van Kempen AA, Romijn JA, Ruiter AF, Ackermans MT, Endert E, Hoekstra JH, et al. Adaptation of glucose production and gluconeogenesis to diminishing glucose infusion in preterm infants at varying gestational ages. Pediatr Res. 2003;53(4):628-34.

152. Hawdon JM, Weddell A, Aynsley-Green A, Ward Platt MP. Hormonal and metabolic response to hypoglycaemia in small for gestational age infants. Arch Dis Child. 1993;68(3 Spec No):269-73.

153. Menni F, de Lonlay P, Sevin C, Touati G, Peigne C, Barbier V, et al. Neurologic outcomes of 90 neonates and infants with persistent hyperinsulinemic hypoglycemia. Pediatrics. 2001;107(3):476-9.

154. Montassir H, Maegaki Y, Ogura K, Kurozawa Y, Nagata I, Kanzaki S, et al. Associated factors in neonatal hypoglycemic brain injury. Brain Dev. 2009;31(9):649-56.

155. Kinnala A, Rikalainen H, Lapinleimu H, Parkkola R, Kormano M, Kero P. Cerebral magnetic resonance imaging and ultrasonography findings after neonatal hypoglycemia. Pediatrics. 1999;103(4 Pt 1):724-9.

156. Duvanel CB, Fawer CL, Cotting J, Hohlfeld P, Matthieu JM. Long-term effects of neonatal hypoglycemia on brain growth and psychomotor development in small-for-gestational-age preterm infants. J Pediatr. 1999;134(4):492-8.

157. Filan PM, Inder TE, Cameron FJ, Kean MJ, Hunt RW. Neonatal hypoglycemia and occipital cerebral injury. J Pediatr. 2006;148(4):552-5.

158. McKinlay CJD, Alsweiler JM, Anstice NS, Burakevych N, Chakraborty A, Chase JG, et al. Association of Neonatal Glycemia With Neurodevelopmental Outcomes at 4.5 Years. JAMA Pediatr. 2017;171(10):972-83.

159. Tin W, Brunskill G, Kelly T, Fritz S. 15-year follow-up of recurrent "hypoglycemia" in preterm infants. Pediatrics. 2012;130(6):e1497-503.

160. Goode RH, Rettiganti M, Li J, Lyle RE, Whiteside-Mansell L, Barrett KW, et al. Developmental Outcomes of Preterm Infants With Neonatal Hypoglycemia. Pediatrics. 2016;138(6).

161. van Kempen A, Eskes PF, Nuytemans D, van der Lee JH, Dijksman LM, van Veenendaal NR, et al. Lower versus Traditional Treatment Threshold for Neonatal Hypoglycemia. N Engl J Med. 2020;382(6):534-44.

162. Shah R, Harding J, Brown J, McKinlay C. Neonatal Glycaemia and Neurodevelopmental Outcomes: A Systematic Review and Meta-Analysis. Neonatology. 2019;115(2):116-26.

163. Wackernagel D, Gustafsson A, Edstedt Bonamy AK, Reims A, Ahlsson F, Elfving M, et al. Swedish national guideline for prevention and treatment of neonatal hypoglycaemia in newborn infants with gestational age >/=35 weeks. Acta Paediatr. 2020;109(1):31-44.

164. Bonamy AK, Källén K, Norman M. High blood pressure in 2.5-year-old children born extremely preterm. Pediatrics. 2012;129(5):e1199-204.

165. Edstedt Bonamy AK, Mohlkert LA, Hallberg J, Liuba P, Fellman V, Domellöf M, et al. Blood Pressure in 6-Year-Old Children Born Extremely Preterm. J Am Heart Assoc. 2017;6(8).

79

166. de Jong F, Monuteaux MC, van Elburg RM, Gillman MW, Belfort MB. Systematic review and meta-analysis of preterm birth and later systolic blood pressure. Hypertension. 2012;59(2):226-34.

167. Doyle LW, Faber B, Callanan C, Morley R. Blood pressure in late adolescence and very low birth weight. Pediatrics. 2003;111(2):252-7.

168. Hack M, Schluchter M, Cartar L, Rahman M. Blood pressure among very low birth weight (<1.5 kg) young adults. Pediatr Res. 2005;58(4):677-84.

169. Duncan AF, Heyne RJ, Morgan JS, Ahmad N, Rosenfeld CR. Elevated systolic blood pressure in preterm very-low-birth-weight infants </=3 years of life. Pediatr Nephrol. 2011;26(7):1115-21.

170. Ueda P, Cnattingius S, Stephansson O, Ingelsson E, Ludvigsson JF, Bonamy AK. Cerebrovascular and ischemic heart disease in young adults born preterm: a population-based Swedish cohort study. Eur J Epidemiol. 2014;29(4):253-60.

171. Cook NR, Cohen J, Hebert PR, Taylor JO, Hennekens CH. Implications of small reductions in diastolic blood pressure for primary prevention. Arch Intern Med. 1995;155(7):701-9.

172. Johansson S, Norman M, Legnevall L, Dalmaz Y, Lagercrantz H, Vanpée M. Increased catecholamines and heart rate in children with low birth weight: perinatal contributions to sympathoadrenal overactivity. J Intern Med. 2007;261(5):480-7.

173. Rakow A, Johansson S, Legnevall L, Sevastik R, Celsi G, Norman M, et al. Renal volume and function in school-age children born preterm or small for gestational age. Pediatr Nephrol. 2008;23(8):1309-15.

174. Bonamy AK, Bendito A, Martin H, Andolf E, Sedin G, Norman M. Preterm birth contributes to increased vascular resistance and higher blood pressure in adolescent girls. Pediatr Res. 2005;58(5):845-9.

175. Watterberg KL, Hintz SR, Do B, Vohr BR, Lowe J, Newman JE, et al. Adrenal function links to early postnatal growth and blood pressure at age 6 in children born extremely preterm. Pediatr Res. 2019;86(3):339-47.

176. Barker DJ, Osmond C, Law CM. The intrauterine and early postnatal origins of cardiovascular disease and chronic bronchitis. J Epidemiol Community Health. 1989;43(3):237-40.

177. Johansson S, Iliadou A, Bergvall N, Tuvemo T, Norman M, Cnattingius S. Risk of high blood pressure among young men increases with the degree of immaturity at birth. Circulation. 2005;112(22):3430-6.

178. Whincup P, Cook D, Papacosta O, Walker M. Birth weight and blood pressure: cross sectional and longitudinal relations in childhood. BMJ. 1995;311(7008):773-6.

179. Kolacek S, Kapetanovic T, Luzar V. Early determinants of cardiovascular risk factors in adults. B. Blood pressure. Acta Paediatr. 1993;82(4):377-82.

180. Ferguson TS, Younger-Coleman NO, Tulloch-Reid MK, Knight-Madden JM, Bennett NR, Samms-Vaughan M, et al. Birth weight and maternal socioeconomic circumstances were inversely related to systolic blood pressure among Afro-Caribbean young adults. J Clin Epidemiol. 2015;68(9):1002-9.

80

181. Pereira JA, Rondo PH, Lemos JO, Pacheco de Souza JM, Dias RS. The influence of birthweight on arterial blood pressure of children. Clin Nutr. 2010;29(3):337-40.

182. Zhang Z, Kris-Etherton PM, Hartman TJ. Birth weight and risk factors for cardiovascular disease and type 2 diabetes in US children and adolescents: 10 year results from NHANES. Matern Child Health J. 2014;18(6):1423-32.

183. Zarrati M, Shidfar F, Razmpoosh E, Nezhad FN, Keivani H, Hemami MR, et al. Does low birth weight predict hypertension and obesity in schoolchildren? Ann Nutr Metab. 2013;63(1-2):69-76.

184. Mu M, Wang SF, Sheng J, Zhao Y, Li HZ, Hu CL, et al. Birth weight and subsequent blood pressure: a meta-analysis. Arch Cardiovasc Dis. 2012;105(2):99-113.

185. Taittonen L, Nuutinen M, Turtinen J, Uhari M. Prenatal and postnatal factors in predicting later blood pressure among children: cardiovascular risk in young Finns. Pediatr Res. 1996;40(4):627-32.

186. Horta BL, Barros FC, Victora CG, Cole TJ. Early and late growth and blood pressure in adolescence. J Epidemiol Community Health. 2003;57(3):226-30.

187. Jarvelin MR, Sovio U, King V, Lauren L, Xu B, McCarthy MI, et al. Early life factors and blood pressure at age 31 years in the 1966 northern Finland birth cohort. Hypertension. 2004;44(6):838-46.

188. Kowalski RR, Beare R, Doyle LW, Smolich JJ, Cheung MM, Victorian Infant Collaborative Study Group. Elevated Blood Pressure with Reduced Left Ventricular and Aortic Dimensions in Adolescents Born Extremely Preterm. J Pediatr. 2016;172:75-80 e2.

189. Keijzer-Veen MG, Finken MJ, Nauta J, Dekker FW, Hille ET, Frolich M, et al. Is blood pressure increased 19 years after intrauterine growth restriction and preterm birth? A prospective follow-up study in The Netherlands. Pediatrics. 2005;116(3):725-31.

190. Belfort MB, Gillman MW, McCormick MC. Prenatal and perinatal predictors of blood pressure at school age in former preterm, low birth weight infants. J Perinatol. 2012;32(4):265-9.

191. Belfort MB, Martin CR, Smith VC, Gillman MW, McCormick MC. Infant weight gain and school-age blood pressure and cognition in former preterm infants. Pediatrics. 2010;125(6):e1419-26.

192. Vohr BR, Allan W, Katz KH, Schneider KC, Ment LR. Early predictors of hypertension in prematurely born adolescents. Acta Paediatr. 2010;99(12):1812-8.

193. de Jong M, Lafeber HN, Cranendonk A, van Weissenbruch MM. Components of the metabolic syndrome in early childhood in very-low-birth-weight infants. Horm Res Paediatr. 2014;81(1):43-9.

194. Embleton ND, Korada M, Wood CL, Pearce MS, Swamy R, Cheetham TD. Catch-up growth and metabolic outcomes in adolescents born preterm. Arch Dis Child. 2016;101(11):1026-31.

195. Suikkanen J, Matinolli HM, Eriksson JG, Jarvenpaa AL, Andersson S, Kajantie E, et al. Early postnatal nutrition after preterm birth and cardiometabolic risk factors in young adulthood. PLoS One. 2018;13(12):e0209404.

81

196. Kappeler L, De Magalhaes Filho C, Leneuve P, Xu J, Brunel N, Chatziantoniou C, et al. Early postnatal nutrition determines somatotropic function in mice. Endocrinology. 2009;150(1):314-23.

197. Wilson AC, Forsyth JS, Greene SA, Irvine L, Hau C, Howie PW. Relation of infant diet to childhood health: seven year follow up of cohort of children in Dundee infant feeding study. BMJ. 1998;316(7124):21-5.

198. Martin RM, Gunnell D, Smith GD. Breastfeeding in infancy and blood pressure in later life: systematic review and meta-analysis. Am J Epidemiol. 2005;161(1):15-26.

199. Lawlor DA, Riddoch CJ, Page AS, Andersen LB, Wedderkopp N, Harro M, et al. Infant feeding and components of the metabolic syndrome: findings from the European Youth Heart Study. Arch Dis Child. 2005;90(6):582-8.

200. Forsyth JS, Willatts P, Agostoni C, Bissenden J, Casaer P, Boehm G. Long chain polyunsaturated fatty acid supplementation in infant formula and blood pressure in later childhood: follow up of a randomised controlled trial. BMJ. 2003;326(7396):953.

201. Brambilla P, Bedogni G, Pietrobelli A, Cianfarani S, Agostoni C. Predictors of blood pressure at 7-13 years: The "new millennium baby" study. Nutr Metab Cardiovasc Dis. 2016;26(8):706-12.

202. Normia J, Laitinen K, Isolauri E, Poussa T, Jaakkola J, Ojala T. Impact of intrauterine and post-natal nutritional determinants on blood pressure at 4 years of age. J Hum Nutr Diet. 2013;26(6):544-52.

203. Lucas A, Morley R. Does early nutrition in infants born before term programme later blood pressure? BMJ. 1994;309(6950):304-8.

204. Singhal A, Cole TJ, Lucas A. Early nutrition in preterm infants and later blood pressure: two cohorts after randomised trials. Lancet. 2001;357(9254):413-9.

205. Webb AL, Conlisk AJ, Barnhart HX, Martorell R, Grajeda R, Stein AD. Maternal and childhood nutrition and later blood pressure levels in young Guatemalan adults. Int J Epidemiol. 2005;34(4):898-904.

206. Singhal A, Cole TJ, Fewtrell M, Kennedy K, Stephenson T, Elias-Jones A, et al. Promotion of faster weight gain in infants born small for gestational age: is there an adverse effect on later blood pressure? Circulation. 2007;115(2):213-20.

207. Niklasson A, Albertsson-Wikland K. Continuous growth reference from 24th week of gestation to 24 months by gender. BMC Pediatr. 2008;8:8.

208. Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M, et al. A new and improved population-based Canadian reference for birth weight for gestational age. Pediatrics. 2001;108(2):E35.

209. National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics. 2004;114(2 Suppl 4th Report):555-76.

210. Wechsler DJSA, TX: The Psychological Corporation. Wechsler intelligence scale for children–Fourth Edition (WISC-IV). 2003.

211. Henderson SE, Sugden DA, Barnett AL. Movement Assessment Battery for Children-2: MABC-2: Pearson Assessment; 2007.

82

212. Rosenbaum PL, Palisano RJ, Bartlett DJ, Galuppi BE, Russell DJ. Development of the Gross Motor Function Classification System for cerebral palsy. Dev Med Child Neurol. 2008;50(4):249-53.

213. Beardsall K. Real time continuous glucose monitoring in neonatal intensive care. Early Hum Dev. 2019;138:104844.

214. Tiberi E, Cota F, Barone G, Perri A, Romano V, Iannotta R, et al. Continuous glucose monitoring in preterm infants: evaluation by a modified Clarke error grid. Ital J Pediatr. 2016;42:29.

215. Perri A, Giordano L, Corsello M, Priolo F, Vento G, Zecca E, et al. Continuous glucose monitoring (CGM) in very low birth weight newborns needing parenteral nutrition: validation and glycemic percentiles. Ital J Pediatr. 2018;44(1):99.

216. Thomson L, Elleri D, Bond S, Howlett J, Dunger DB, Beardsall K. Targeting glucose control in preterm infants: pilot studies of continuous glucose monitoring. Arch Dis Child Fetal Neonatal Ed. 2019;104(4):F353-F9.

217. Harris DL, Battin MR, Weston PJ, Harding JE. Continuous glucose monitoring in newborn babies at risk of hypoglycemia. J Pediatr. 2010;157(2):198-202 e1.

218. Beardsall K, Ogilvy-Stuart AL, Ahluwalia J, Thompson M, Dunger DB. The continuous glucose monitoring sensor in neonatal intensive care. Arch Dis Child Fetal Neonatal Ed. 2005;90(4):F307-10.

219. Galderisi A, Lago P, Steil GM, Ghirardo M, Cobelli C, Baraldi E, et al. Procedural Pain during Insertion of a Continuous Glucose Monitoring Device in Preterm Infants. J Pediatr. 2018;200:261-4 e1.

220. Westin V, Klevebro S, Domellöf M, Vanpée M, Hallberg B, Stoltz Sjöström E. Improved nutrition for extremely preterm infants - A population based observational study. Clin Nutr ESPEN. 2018;23:245-51.

221. Uettwiller F, Chemin A, Bonnemaison E, Favrais G, Saliba E, Labarthe F. Real-time continuous glucose monitoring reduces the duration of hypoglycemia episodes: a randomized trial in very low birth weight neonates. PLoS One. 2015;10(1):e0116255.

222. Galderisi A, Facchinetti A, Steil GM, Ortiz-Rubio P, Cavallin F, Tamborlane WV, et al. Continuous Glucose Monitoring in Very Preterm Infants: A Randomized Controlled Trial. Pediatrics. 2017;140(4).

223. Beardsall K, Thomson L, Guy C, van Weissenbruch MM, Iglesias I, Muthukumar P, et al. Protocol of a randomised controlled trial of real-time continuous glucose monitoring in neonatal intensive care 'REACT'. BMJ Open. 2018;8(6):e020816.

224. Beardsall K, Thomson L, Elleri D, Dunger DB, Hovorka R. Feasibility of automated insulin delivery guided by continuous glucose monitoring in preterm infants. Arch Dis Child Fetal Neonatal Ed. 2019.

225. Signal M, Le Compte A, Harris DL, Weston PJ, Harding JE, Chase JG, et al. Impact of retrospective calibration algorithms on hypoglycemia detection in newborn infants using continuous glucose monitoring. Diabetes Technol Ther. 2012;14(10):883-90.

226. McKinlay CJD, Chase JG, Dickson J, Harris DL, Alsweiler JM, Harding JE. Continuous glucose monitoring in neonates: a review. Matern Health Neonatol Perinatol. 2017;3:18.

83

227. Thomas F, Signal M, Harris DL, Weston PJ, Harding JE, Shaw GM, et al. Continuous glucose monitoring in newborn infants: how do errors in calibration measurements affect detected hypoglycemia? J Diabetes Sci Technol. 2014;8(3):543-50.

228. Jeong TD, Cho EJ, Ko DH, Lee W, Chun S, Hong KS, et al. Large-scale performance evaluation of Accu-Chek inform II point-of-care glucose meters. Scand J Clin Lab Invest. 2016;76(8):657-63.

229. Warner JV, Wu JY, Buckingham N, McLeod DS, Mottram B, Carter AC. Can one point-of-care glucose meter be used for all pediatric and adult hospital patients? Evaluation of three meters, including recently modified test strips. Diabetes Technol Ther. 2011;13(1):55-62.

230. Reddy VRS, M ES, Gowda YCB, Suhail SM. Comparison of Point of Care (POC) Testing of Glucose by B Braun Glucometer and Hemocue Glucose 201+ Analyser Versus Centralised Testing in Neonatal Intensive Care Unit (NICU). J Clin Diagn Res. 2014;8(7):PC10-3.

231. Goldner Pérez CP, Kushnir A, Saslow J, Quinones Cardona V. Long-Term Developmental Outcomes in Preterm Neonates Exposed to Hyperglycemia. Pediatrics. 2019;144(2 MettingAbstract):656.

232. Kanarek KS, Santeiro ML, Malone JI. Continuous infusion of insulin in hyperglycemic low-birth weight infants receiving parenteral nutrition with and without lipid emulsion. JPEN J Parenter Enteral Nutr. 1991;15(4):417-20.

233. Hovi P, Vohr B, Ment LR, Doyle LW, McGarvey L, Morrison KM, et al. Blood Pressure in Young Adults Born at Very Low Birth Weight: Adults Born Preterm International Collaboration. Hypertension. 2016;68(4):880-7.

234. Alsweiler J, Williamson K, Bloomfield F, Chase G, Harding J. Computer-determined dosage of insulin in the management of neonatal hyperglycaemia (HINT2): protocol of a randomised controlled trial. BMJ Open. 2017;7(3):e012982.