Enteral Nutritional Support and Use of Diabetes-Speciﬁc ... · formulas. CONCLUSIONS— This systematic review shows that short- and long-term use of diabetes-speciﬁc formulas

Enteral Nutritional Support and Use ofDiabetes-Specific Formulas for PatientsWith DiabetesA systematic review and meta-analysis

MARINOS ELIA, MD, BSC(HONS), FRCP1

ANTONIO CERIELLO, MD2

HEINER LAUBE, MD, PHD3

ALAN J. SINCLAIR, MD, PHD4

MEIKE ENGFER, PHD5

REBECCA J. STRATTON, BSC(HONS), PHD,

SRD1

OBJECTIVE — The aim of this systematic review was to determine the benefits of nutritionalsupport in patients with type 1 or type 2 diabetes.

RESEARCH DESIGN AND METHODS — Studies utilizing an enteral nutritional sup-port intervention (oral supplements or tube feeding) were identified using electronic databasesand bibliography searches. Comparisons of interest were nutritional support versus routine careand standard versus diabetes-specific formulas (containing high proportions of monounsatu-rated fatty acids, fructose, and fiber). Outcomes of interest were measures of glycemia and lipidstatus, medication requirements, nutritional status, quality of life, complications, and mortality.Meta-analyses were performed where possible.

RESULTS — A total of 23 studies (comprising 784 patients) of oral supplements (16 studies)and tube feeding (7 studies) were included in the review, and the majority compared diabetes-specific with standard formulas. Compared with standard formulas, diabetes-specific formulassignificantly reduced postprandial rise in blood glucose (by 1.03 mmol/l [95% CI 0.58–1.47];six randomized controlled trials [RCTs]), peak blood glucose concentration (by 1.59 mmol/l[86–2.32]; two RCTs), and glucose area under curve (by 7.96 mmol � l�1 � min�1 [2.25–13.66];four RCTs, i.e., by 35%) with no significant effect on HDL, total cholesterol, or triglycerideconcentrations. In addition, individual studies reported a reduced requirement for insulin (26–71% lower) and fewer complications with diabetes-specific compared with standard nutritionalformulas.

CONCLUSIONS — This systematic review shows that short- and long-term use of diabetes-specific formulas as oral supplements and tube feeds are associated with improved glycemic controlcompared with standard formulas. If such nutritional support is given long term, this may haveimplications for reducing chronic complications of diabetes, such as cardiovascular events.

Diabetes Care 28:2267–2279, 2005

The impact of better glycemic controlon long-term clinical outcome iswell recognized in both type 1 (1)

and type 2 (2) diabetes, where hypergly-cemia may result in life-threatening com-plications and numerous comorbidities.In addition, many conditions, includingaccidental injury, stroke, and critical ill-ness, show a worse outcome in the pres-ence of hyperglycemia (3).

In the U.K., the costs associated withmajor hyperglycemic complicationsrange from £872 (€1,256 or $1,607 forblindness in one eye) to £8,459 (€12,178or $15,591 for amputation) per patient(4), and the U.S. has reported annual di-abetes health care costs of $11,157(€8,710) per patient (5). This large eco-nomic burden is unsurprising given thatpatients with diabetes are known to beadmitted to the hospital more often thanother patient groups, accounting for up to25% of intensive care admissions (3,6).Many of these hospitalized patients willrequire nutritional support. In addition,an increasing number of patients receivelong-term home enteral tube feeding(ETF), including those with diabetes (7).

Standard enteral (oral or tube) nutri-tional formulas are high in carbohydrate(mostly low–molecular weight sources),low in fat, and low in fiber. Standard formu-las may compromise glycemic control in pa-tients with diabetes, due to a rapid gastricemptying rate and rapid nutrient assimila-tion (8,9). For this reason, diabetes-specificformulas have been developed.

Diabetes-specific formulas contain adefined nutrient composition designed toenable better glycemic control. Such nu-trients include fructose (10), fiber (11),monounsaturated fatty acids (MUFAs)(12), soy protein (13,14), and antioxi-dants (15). Although general guidelinesexist regarding the composition of thediet for those with diabetes (16 –18),there are no specific guidelines for pa-tients with diabetes who are at risk of mal-nutrition, requiring nutritional support.

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

From the 1Instutite of Human Nutrition, University of Southampton, Southampton, U.K.; the 2Departmentof Pathology and Medicine (Experimental and Clinical), University of Udine, Udine, Italy; the 3Departmentof Internal Medicine, University of Giessen, Giessen, Germany; the 4Section of Geriatric Medicine andGerontology, Diabetes Research Unit, Centre for Health Services, University of Warwick, Warwick, U.K.;and the 5Clinical Nutrition Division, Royal Numico, Amsterdam, the Netherlands.

Address correspondence and reprint requests to Prof. Marinos Elia, Institute of Human Nutrition, Uni-versity of Southampton, Southampton General Hospital, MP 113 F Level, Tremona Road, Southampton,SO16 6YD, U.K. E-mail: [email protected].

Received for publication 3 March 2005 and accepted in revised form 23 May 2005.Additional information for this article can be found in an online appendix available at http://

care.diabetesjournals.org.Abbreviations: AUC, area under the curve; CCT, controlled CT; CT, clinical trial; ETF, enteral tube

feeding; MUFA, monounsaturated fatty acid; ONS, oral nutritional supplement; RCT, randomized con-trolled trial.

A table elsewhere in this issue shows conventional and Systeme International (SI) units and conversionfactors for many substances.

© 2005 by the American Diabetes Association.The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby

marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

R e v i e w s / C o m m e n t a r i e s / A D A S t a t e m e n t sR E V I E W A R T I C L E

DIABETES CARE, VOLUME 28, NUMBER 9, SEPTEMBER 2005 2267

For example, although general guidelinessuggest that a high intake of MUFAs/totalfat may be disadvantageous for the well-nourished patient with diabetes, this maybe advantageous for the treatment of amalnourished patient. Malnutrition isseen in several patient groups with diabe-tes, especially in the elderly (19,20) andthose with complications such as renalfailure or neurological dysfunctions. Inthese patients, an impaired nutritionalstatus is associated with increased suscep-tibility to and recovery from infectiouscomplications, the development of pres-sure sores (and their failure to heal), andgeneral functional decline (21). Nutri-tional support using diabetes-specific for-mulas in these patient groups mayprevent such complications.

There have been no systematic re-views or meta-analyses regarding the useof enteral nutritional support in patientswith diabetes, although a few clinical re-views have been published (8,9,22).Therefore, a systematic review and meta-analysis was undertaken with the follow-ing aims: 1) to examine the impact of

enteral (oral or tube) nutritional supportversus routine care on the nutritional sta-tus and clinical outcome of patients withdiabetes and, more specifically, 2) to in-vestigate whether diabetes-specific en-teral formulas are superior to standardenteral formulas by assessing the effects ofthese on glycemia, lipidemia, nutritionalstatus, medication requirements, qualityof life, complications, and mortality.

RESEARCH DESIGN ANDMETHODS — The rev i ew wasplanned, conducted, and reported fol-lowing published guidelines. These in-clude those issued by the CochraneCollaboration (23), the U.K. NationalHealth Service Centre for Reviews andDissemination (24,25), and the QUO-RUM guidelines (26). A flow chart (Fig. 1)illustrates the principle stages and pro-cesses undertaken.

Identification and retrieval ofstudiesPotentially relevant studies were identi-fied by searching electronic databases.

These included PubMed (27), accessed10 August 2004; Cochrane (28), accessed10 August 2004; Turning Research IntoPractice (29), accessed 19 August 2004;Clinical Evidence (30), accessed 19 Au-gust 2004; National Electronic Library forHealth Guidelines finder (31), accessed19 August 2004; and National ServiceFrameworks (32), accessed 19 August2004. The search terms included: diabe-tes mellitus, diabetic, monounsaturat*,mono-unsaturat*, MUFA, mono unsat-urat*, soy, soya, fructose, fiber, fiber, nu-trit ion*, nutrie*, enteral*, oral*,supplement*, sip, feed, formula*, liquid,tube, nasogastric, nasojejunal, nasoduo-denal, gastrostomy, jejunostomy, clinicaltrial. Bibliographies of identified trialswere checked and experts consulted forany additional studies.

Study selection criteria, dataextraction, and outcome measuresStudies were deemed eligible for inclu-sion in the review if they conformed topredetermined inclusion and exclusioncriteria. Subjects eligible for inclusion

Fig. 1—Summary of study methodology stages and process of the systematic review

Nutritional support for diabetes

2268 DIABETES CARE, VOLUME 28, NUMBER 9, SEPTEMBER 2005

were adults (aged �18 years) with type 1or type 2 diabetes or stress diabetescaused by acute illness, of any nutritionalstatus (well nourished, malnourished),and based in any setting (e.g., hospital,outpatient, home). Studies using hypoca-loric feeding regimens in obese subjectswith the intention of inducing weight losswere excluded. Eligible interventionswere formulas given enterally, eitherorally (oral nutritional supplements[ONSs]]) or by tube (ETF) that containedat least two macronutrients as well as mi-cronutrients. The intervention could pro-vide either a portion of, or the completedaily requirement for energy and could benutritionally complete or incomplete.Studies using concurrent parenteral nu-trition or dietary advice were admissible,but those utilizing only parenteral nutri-tion or only dietary counseling were ex-cluded. The comparisons of interest forboth ONSs and ETF were nutritional sup-port versus routine care, diabetes-specificformula versus standard formula, and acomparison of ETF versus parenteral nu-trition. As the definition and compositionof diabetes-specific formulas can be vari-able, for the purpose of this review formu-las containing a high proportion (e.g.,�60%) of fat, such as MUFAs and fruc-tose and fiber, were designated “diabetesspecific,” and all other formulations weredesignated “standard formula.” Where astudy provided three or more eligible in-tervention arms, the two interventionsused in the analysis were selected accord-ing to the compositions closest to a typicalstandard and a diabetes-specific feed.

Outcome measures sought were gly-cemia, lipidemia, nutritional status, med-ication requirements, quality of life,complications, and mortality. Wheremultiple measurements were provided(e.g., multiple blood samples taken post-prandially), the last in each series wasused. No other restrictions were placedon studies with regard to type of compar-ator (e.g., no nutritional support, dietaryadvice, parenteral nutrition), year of pub-lication, language (providing an Englishtranslation of the report or its abstract wasavailable), and source. Priority was givento randomized controlled trials (RCTs);however, nonrandomized controlledclinical trials (CCTs) and before-afterclinical trials (CTs) were admissible. Ob-servational study designs (e.g., cohort,case study) were excluded. Following theidentification of potentially relevant stud-

ies based on titles and abstracts, full arti-cles were obtained and evaluated by oneresearcher; a second assessor verified in-clusion/exclusion decisions. A predeter-mined data extraction table was designedto capture study characteristics and out-come data and allow the assimilation ofdata from differing study designs.

Quality assessmentThe quality of individual studies was as-sessed using two scales (33,34) by oneresearcher and verified by a second asses-sor. The first method was a six-point scaleadapted from the Quality of EvidenceQuality Assessment Scale (Agency forHealth Care and Policy Research) (33),and the second method was that used byJadad et al. (34), which was previouslyreported.

Synthesis of data and statisticalmethodsFollowing extraction of data, where ap-propriate and feasible, the results of com-parable groups of trials were combinedand meta-analysis undertaken on relevantoutcome measures. The comparisons ofinterest were nutritional support versusroutine care, diabetes-specific formulaversus standard formula, and ETF versusparenteral nutrition. Subanalyses wereplanned for studies of different length fol-low-up (�1 day vs. �1day), and separateanalyses were intended for diabetes type(type 1 versus type 2) and nutritional sta-tus (malnourished versus well nourishedversus obese).

Hedges’ unbiased estimator of thestandardized mean difference for relevanttreatment comparisons was calculated(35). The mean treatment difference wasconsidered statistically significant if the95% CI did not span the value zero. For-est plots were used to present each study’sstandardized difference and the meta-analysis estimate. Heterogeneity was in-ves t iga ted f rom the Q tes t o fheterogeneity derived by the Mantel-Haenszel method (36). Due to the smallnumber of studies included in the meta-analyses, it was deemed inappropriate toinvestigate publication bias through theuse of funnel plots (37). A fixed-effectsmodel was used to combine the treatmentestimates, which assumes no heterogene-ity between the study results. A meta-analysis estimate of the treatment effectsize was calculated as a weighted sum ofthe effect size for each study, where the

weight was calculated as the reciprocal ofHedges’ estimated variance of the effectsize for each individual study.

The selection of data for analysis wasconducted as follows: change from base-line data were used, except if one or morestudies in the meta-analysis failed to re-port baseline data, in which case postin-tervention data were used. The correla-tion between baseline and postinterven-tion data for measures of glycemia andlipidemia was assumed to be zero (r � 0),which results in the most conservativemethod of analysis. The sensitivity of thisassumption was tested by additionallyconducting the analyses using r � 0.5. Allstatistical analyses were conducted usingSAS version 8.2 (SAS Institute, Cary, NC).All data are presented as means � SD,unless otherwise stated.

RESULTS

Overall search findingsA total of 4,141 studies were identified bythe search strategy, of which 23 compliedwith the inclusion criteria and were in-cluded in the review (Fig. 1). Study detailsare provided in appendix 1 (online appen-dix[availableathttp://care.diabetesjournals.org]), and reasons for study exclusion areprovided in appendix 2 (online appendix).

Of the 23 studies included in the re-view, 19 were RCTs (11,14,38–54) scor-ing the highest grade of one, according tothe Quality of Evidence Scale (33). How-ever, the methodology of individual RCTswas often poorly described (with regardto methods of randomization, blinding,and recording number of drop-outs),with only three studies (45,47,48) scor-ing the top grade of five on the Jadad scale(34). The remaining RCTs scored four(46,49), three (38,54), two (11,14,39–44,50,52), or one (51,53). The reviewalso included three CCTs (55–57), scor-ing two on the Quality of Evidence Scale(33) and one CT (58) scoring four.

Most trials consisted of patients withtype 2 diabetes (n � 16), with fewer stud-ies of patients with type 1 diabetes (n �4). A minority of studies did not specifythe type of diabetes (n � 1), included pa-tients with type 1 or type 2 diabetesand/or stress diabetes caused by acute ill-ness (n � 1), or included only patientswith stress diabetes caused by acute ill-ness (n � 1).

Elia and Associates


Tab

le1—

Dat

aus

edin

the

met

a-an

alys

es

Var

iabl

eC

itat

ion

Tim

ing

ofsa

mpl

esIn

terv

enti

onn

Base

line

Post

inte

rven

tion

Cha

nge

from

base

line

Post

pran

dial

rise

ingl

ucos

ePe

ters

etal

.,19

89ba

selin

e(f

aste

d)�

0m

ins,

post

inte

rven

tion

�po

stpr

andi

al24

0m

ins

afte

rm

eals

tart

Stan

dard

form

ula

10N

RN

R1,

900

�32

0m

g�l

�1

�4h�

1*

Dia

bete

s-sp

ecifi

cfo

rmul

a10

NR

NR

�20

�33

0m

g�l

�1

�4h�

1*

Sanz

-Par

iset

al.,

1998

(ins

ulin

)ba

selin

e�

0m

ins

(fas

ted)

,po

stin

terv

enti

on�

post

pran

dial

120

min

saf

ter

mea

lsta

rt

Stan

dard

form

ula

201,

750

�48

0m

g/l*

2,56

0�

540

mg/

l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a20

1,94

0�

560

mg/

l*2,

160

�72

0m

g/l*

NR

†

Sanz

-Par

iset

al.,

1998

(sul

phon

yure

a)St

anda

rdfo

rmul

a5

1,51

0�

200

mg/

l*1,

900

�46

0m

g/l*

NR

†D

iabe

tes-

spec

ific

form

ula

71,

580

�29

0m

g/l*

2,03

0�

480

mg/

l*N

R†

delC

arm

enC

resp

illo

etal

.,20

03ba

selin

e�

0m

ins

(fas

ted)

,po

stin

terv

enti

on�

post

pran

dial

150

min

saf

ter

mea

lsta

rt

Stan

dard

form

ula

11N

RN

R18

4�

133

mg/

lD

iabe

tes-

spec

ific

form

ula

11N

RN

R�

8�

99m

g/l

Cra

iget

al.,

1998

Base

line

�fa

sted

,po

stin

terv

enti

on�

post

pran

dial

,bot

haf

ter

3m

onth

sin

terv

enti

on

Stan

dard

form

ula

136.

9�

0.6

mm

ol/l*

10.4

�1

mm

ol/l*

NR

†D

iabe

tes-

spec

ific

form

ula

147.

3�

0.4

mm

ol/l*

9�

1.3

mm

ol/l*

NR

†

McC

arga

ret

al.,

1998

Base

line

�fa

sted

(pre

mea

l),

post

inte

rven

tion

�2

hpo

stpr

andi

alaf

ter

28da

ysin

terv

enti

on

Stan

dard

form

ula

16N

RN

R1.

6�

0.4

mm

ol/l*

Dia

bete

s-sp

ecifi

cfo

rmul

a16

NR

NR

1.1

�0.

2m

mol

/l*

Mes

ejo

etal

.,20

03Ba

selin

e�

onad

mis

sion

(stu

dyst

art)

,po

stin

terv

enti

on�

post

pran

dial

afte

r14

days

inte

rven

tion

Stan

dard

form

ula

242,

103

�63

0m

g/l

2,22

8�

471.

2m

g/l

NR

†

Dia

bete

s-sp

ecifi

cfo

rmul

a26

1,90

9�

450

mg/

l1,

768

�44

0.1

mg/

lN

R†

Peak

bloo

dgl

ucos

eH

ofm

anet

al.,

2004

Base

line

�0

min

s(f

aste

d),

post

inte

rven

tion

�6

haf

ter

mea

lsta

rt

Stan

dard

form

ula

12N

RN

R4

�1.

4m

mol

/lD

iabe

tes-

spec

ific

form

ula

12N

RN

R2.

5�

1m

mol

/l

Hof

man

etal

.,20

04Ba

selin

e�

0m

ins

(fas

ted)

,po

stin

terv

enti

on�

120

min

saf

ter

mea

lsta

rt

Stan

dard

form

ula

10N

RN

R4

�0.

4m

mol

/l*D

iabe

tes-

spec

ific

form

ula

10N

RN

R2.

3�

0.4

mm

ol/l*

Bloo

dgl

ucos

eA

UC

Gol

ayet

al.,

1995

IAU

C(m

mol

/Lxh

),4

hSt

anda

rdfo

rmul

a6

NR

NR

15.9

�2.

3m

mol

�l�

1�h

�1*

Dia

bete

s-sp

ecifi

cfo

rmul

a6

NR

NR

8.9

�1.

8m

mol

�l�

1�h

�1*

Prin

tzet

al.,

1997

AU

C(m

mol

/l*m

in),

over

basa

l(0–

180

min

s)St

anda

rdfo

rmul

a10

NR

NR

893

�77

mm

ol�l

�1

�h�

1*

Dia

bete

s-sp

ecifi

cfo

rmul

a10

NR

NR

620

�64

mm

ol�l

�1

�h�

1*

Hof

man

etal

.,20

04iA

UC

(mm

ol/L

xm

in),

6h

Stan

dard

form

ula

12N

RN

R94

5�

474

mm

ol�l

�1

�min

�1

Dia

bete

s-sp

ecifi

cfo

rmul

a12

NR

NR

584

�32

2m

mol

�l�

1�h

�1

Hof

man

etal

.,20

04A

UC

(mm

ol/l/

120m

ins)

abov

eba

selin

ele

vel

Stan

dard

form

ula

10N

RN

R30

7�

29m

mol

�l�

1�1

20m

ins�

1*

Dia

bete

s-sp

ecifi

cfo

rmul

a10

NR

NR

168

�31

mm

ol�l

�1

�120

min

s�1*

Fast

ing

bloo

dgl

ucos

eC

raig

etal

.,19

98ba

selin

e�

stud

yst

art,

post

inte

rven

tion

�af

ter

3m

onth

sin

terv

enti

on,a

llfa

sted

Stan

dard

form

ula

136.

9�

0.6

mm

ol/l*

8.3

�1.

7m

mol

/l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a14

7.3

�0.

4m

mol

/l*6.

7�

0.7

mm

ol/l*

NR

†

McC

arga

ret

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r28

days

inte

rven

tion

,all

fast

ed

Stan

dard

form

ula

168.

73�

0.46

mm

ol/l*

7.97

�0.

53m

mol

/l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a16

9.16

�0.

59m

mol

/l*7.

54�

0.37

mm

ol/l*

NR

†



Nutritional support versus routinecareThere were no long-term studies of ONSsor ETF compared with routine care in pa-tients with diabetes. Two shorter RCTs(39,41) compared nutritional supportversus routine care. Both of these studiesalso provided data for the diabetes-specific versus standard formula compar-ison (below).

Diabetes-specific formula versusstandard formulaMost studies (14,38–56) compared dia-betes-specific formulas with standard for-mulas. Eighteen of these were RCTs(14,38–54) and two were CCTs (55,56).Most (14,40–44,46–48,51–53,55) wereshort-term single-meal studies with �24h follow-up; only seven were longer-termstudies (38,39,45,49,50,54,56) with fol-low-up of 6 days to 3 months. In all butone (47) of the short-term studies, anONS was used (14,40 – 44,46,48,51–53,55). In contrast, all but two (38,39) ofthe longer-term studies used ETF(45,49,50,54,56).

Other studiesTwo studies, including one RCT (11) andone CCT (57), compared different formu-las. One was a single-meal study compar-ing three high-fat formulas (57), and inthe other study, two standard formulaswere given as a sole source of nutrition for5 days (11). One additional CT (58) in-volved the follow-up of elderly patientswith diabetes receiving a diabetes-specificformula via tube for 42 days.

No trials comparing tube feeding withparenteral nutrition were identified, andthere were insufficient data for separateanalyses according to diabetes type (type1 versus type 2) and nutritional status(malnourished versus well nourished ver-sus obese).

OutcomesTable 1 provides details of absolute dataused in meta-analyses.

Nutritional support versus routinecareGlycemia and lipidemia. In one short-term RCT (41), a diabetes-specific for-mula given as an ONS producedsignificantly smaller rises in postprandialglucose and insulin concentrations andglucose area under the curve (AUC) com-pared with both routine care and a stan-H

DL

Cra

iget

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r3

mon

ths

inte

rven

tion

,all

fast

ed

Stan

dard

form

ula

130.

98�

0.05

mm

ol/l*

0.83

�0.

05m

mol

/l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a14

1.01

�0.

05m

mol

/l*0.

98�

0.05

mm

ol/l*

NR

†

McC

arga

ret

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r28

days

inte

rven

tion

,pr

obab

lyfa

sted

Stan

dard

form

ula

161.

37�

0.19

mm

ol/l*

1.08

�0.

07m

mol

/l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a16

1.5

�0.

16m

mol

/l*1.

2�

0.1

mm

ol/l*

NR

†

Tri

glyc

erid

esPe

ters

etal

.,19

89ba

selin

e�

0m

ins

(fas

ted)

,po

stin

terv

enti

on�

240

min

saf

ter

mea

lsta

rt(p

ostp

rand

ial)

Stan

dard

form

ula

100.

54�

0.1

g/l*

0.75

�0.

14g/

l*N

R†

Dia

bete

s-sp

ecifi

cfo

rmul

a10

0.54

�0.

17g/

l*1

�0.

22g/

l*N

R†

delC

arm

enC

resp

illo

etal

.,20

03ba

selin

e�

0m

ins

(fas

ted)

,po

stin

terv

enti

on�

150

min

saf

ter

mea

lsta

rt(p

ostp

rand

ial)

Stan

dard

form

ula

Dia

bete

s-sp

ecifi

cfo

rmul

a11 11

NR

NR

NR

NR

26�

42m

g/l

9�

88m

g/l

Cra

iget

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r3

mon

ths

inte

rven

tion

,all

fast

ed:m

easu

red

astr

iacy

lgly

cero

l.

Stan

dard

form

ula

Dia

bete

s-sp

ecifi

cfo

rmul

a13 14

0.9

�0.

07g/

l†0.

97�

0.13

g/l†

1.06

�0.

12g/

l†0.

91�

0.17

g/l†

NR

†N

R†

McC

arga

ret

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r28

days

inte

rven

tion

,pr

obab

lyfa

sted

Stan

dard

form

ula

Dia

bete

s-sp

ecifi

cfo

rmul

a16 16

1.81

�0.

27m

mol

/l*1.

46�

0.29

mm

ol/l*

1.86

�0.

28m

mol

/l*1.

33�

0.24

mm

ol/l*

NR

†N

R†

Cha

nge

into

tal

Cho

lest

erol

Cra

iget

al.,

1998

base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r3

mon

ths

inte

rven

tion

,all

fast

ed

Stan

dard

form

ula

Dia

bete

s-sp

ecifi

cfo

rmul

a13 14

4.21

�0.

18m

mol

/l*4.

16�

0.31

mm

ol/l*

3.96

�0.

23m

mol

/l*3.

95�

0.31

mm

ol/l*

NR

†N

R†

McC

arga

ret

al.,

1998

Base

line

�st

udy

star

t,po

stin

terv

enti

on�

afte

r28

days

inte

rven

tion

,pr

obab

lyfa

sted

Stan

dard

form

ula

Dia

bete

s-sp

ecifi

cfo

rmul

a16 16

5.06

�0.

21m

mol

/l*5.

24�

0.19

mm

ol/l*

4.58

�0.

27m

mol

/l*5.

05�

0.28

mm

ol/l*

NR

†N

R†

Dat

aar

em

eans

�SD

or*S

E.†

Cha

nge

from

base

line

data

calc

ulat

edby

subt

ract

ing

base

line

from

post

inte

rven

tion

valu

ean

dSD

ofth

edi

ffer

ence

calc

ulat

edby

stan

dard

form

ulas

SD(p

ost-

pre)

��

�Var

(pos

t)�

Var

(pre

)�

Cov

(pre

,pos

t).

NR

,not

reco

rded

.

Elia and Associates


dard formula. A further RCT reported thatthe use of diabetes-specific ONSs as anafternoon snack resulted in similar post-prandial blood glucose concentrations(1,950 mg/l) to an isocaloric food snack(1,960 mg/l) after a standard test meal(supper), which in both cases was signif-icantly lower than that produced by astandard formula (2,430 mg/l) (39). Nochanges in HbA1c (A1C) or in lipid pro-files (undefined in the study report) werefound in this short-term study.

Other outcomesNo studies investigated the impact ofONSs or ETF versus routine care on otherclinically relevant outcomes in diabeticpatients, including changes in nutritionalstatus, requirement for medication, qual-ity of life, complication rates, or mortality.

Diabetes-specific formula versusstandard formulaGlycemia (postprandial rise in glucose,peak glucose, glucose AUC, insulinAUC, A1C, and fasting glucose). Fig-ure 2, a meta-analysis of six RCTs(14,38,44,45,51,54), demonstrated thatdiabetes-specific formulas result in signif-icantly lower postprandial rise in bloodglucose concentrations (by 1.03 mmol/l[95% CI 0.58 –1.47]) compared withstandard formulas (effect size �0.52[�0.81 to �0.24]) (Fig. 2 and Table 1).Both the short-term (effect size �0.71[�1.14 to �0.27]) and longer-term (ef-fect size �0.38 [�0.76 to 0.0]) studiessupported this overall effect (Fig. 2). Ex-clusion of one RCT (54) in hyperglycemiccritically ill patients from the analysis didnot alter the result (effect size �0.57[�0.91 to �0.24]). These meta-analysesassumed zero correlation (r � 0) betweenbaseline and postintervention results, butwhen this assumption was relaxed (r �0.5), all meta-analyses remained signifi-cant (overall effect size �0.59 [�0.87 to�0.3]). Four RCTs provided incomplete(39,46) or graphically presented (41,47)data that could not be included in themeta-analysis. However, all four studiessuggested a lower postprandial rise in glu-cose concentrations with diabetes-specific formulas versus standardformulas; these were statistically signifi-cant in two studies (41,46).

A meta-analysis of two RCTs (47,48)demonstrated that diabetes-specific for-mulas result in significantly lower peakblood glucose concentrations (by 1.59

mmol/l [95% CI 0.86–2.32]) than stan-dard formulas (effect size �1.28 [�1.94to �0.63], assuming r � 0) (Fig. 3 andTable 1).

A meta-analysis of four RCTs (46–48,40) demonstrated that diabetes-specific formulas result in significantlysmaller (31– 45% lower) glucose AUCthan standard formulas (effect size �1.19[95% CI �1.69 to �0.7], assuming r �0) (Fig. 4 and Table 1). Six further RCTs(14,41,42,50,52,53) reported smallerglucose AUC with diabetes-specific orlow-carbohydrate versus standard formu-las. In three of these (61,73,74), the dif-ference was reported to be statisticallysignificant, while the remaining threestudies (14,56,59) did not report any sta-tistical analysis. However, these were notmeta-analyzable due to incomplete databeing reported (41,52,53) or incompati-ble data presentation (14,50) or study de-sign (42).

Three RCTs (42,46,52) reported sig-nificantly smaller insulin AUC followingdiabetes-specific or low-carbohydratecompared with standard formulas. Theresults of two other RCTs (40,50) wereless conclusive. The data were not suffi-ciently comparable to permit meta-analysis of this outcome measure.

Three long-term RCTs involvingONSs (38) and ETF (45,49) reported fa-vorable effects of diabetes-specific formu-las on A1C or fructosamine concentra-tions. One of these studies (49) demon-strated statistical significance, reporting areduction from baseline of A1C by�0.8% in the diabetes-specific group andno change from baseline in the standardgroup. The other two studies (38,45)showed reductions of A1C by 0.6% andof fructosamine by 3%, respectively,whereas increases were noted in thegroup receiving standard food. Twoshorter studies (39,58) reported nochanges in A1C with diabetes-specificformulas given orally or by tube. The datawere not sufficiently comparable to per-mit meta-analysis of this outcome mea-sure.

A meta-analysis was conducted to ex-amine the effect of diabetes-specific ver-sus standard formulas on fasting bloodglucose concentrations, following a com-bination of data from two RCTs involvingONSs (38) and ETF (45). Although inboth studies fasting blood glucose con-centrations were reduced by the use ofdiabetes-specific formulas, there was no

significant difference compared with thestandard formula when r was assumed tobe zero (effect size �0.35 [95% CI �0.86to 0.17]). Another RCT (49) showed thatdiabetes-specific ETF was associated witha significantly greater reduction in fastingblood glucose (�28.6 g/l) than standardformulas (�1.4 g/l) compared with base-line. This study was not meta-analyzablebecause it provided no measure of vari-ability. In contrast, in a further RCT,where an ONS was given as the solesource of nutrition for 6 days, no differ-ences were found between standard anddiabetes-specific formulas, although nonumerical data were reported (50).

Lipidemia (total cholesterol, HDL,and triglycerides)A meta-analysis was conducted to exam-ine the effect of diabetes-specific versusstandard formulas on total serum choles-terol concentrations. Following combina-tion of the data from two RCTs involvingONSs (38) and ETF (45), no significanteffect on cholesterol was found (effect size0.13 [95% CI �0.38 to 0.64], assumedr � 0; effect size 0.18 [�0.33 to 0.69],assumed r � 0.5, respectively). Threeother longer-term RCTs (two ETF andone ONS) and one short-term RCT(ONS), which did not provide suitabledata for meta-analysis, also reported nosignificant difference in total cholesterolof those fed diabetes-specific and stan-dard formulas (49,54).

There was inadequate information toaddress the effects on LDL in the meta-analysis. Four RCTs (38,45,49,54) re-ported no significant differences in LDL/VLDL in patients receiving diabetes-specific and standard formulas.

Meta-analysis of two RCTs involvingONSs (38) and ETF (45) found no signif-icant effect of diabetes-specific versusstandard formula on HDL (effect size 0.2[95% CI �0.31 to 0.72], r � 0; effect size0.28 [�0.24 to 0.8], r � 0.5), although inboth studies the diabetes-specific formu-las showed higher HDL concentrationsthan the standard formulas (Table 1).Other long-term RCTs (49,54) reportedno significant difference in HDL concen-tration following diabetes-specific versusstandard formulas, although they pro-vided no suitable data for meta-analysis.

A meta-analysis was conducted to in-vestigate the effect of diabetes-specificversus standard formulas on blood tri-glyceride concentrations. Although in the



Fig.

2—

The

effectof

diabetes-specificversus

standardform

ulason

thepostprandial

risein

bloodglucose

concentration:a

meta-analysis

offive

RC

Ts

(14,38,44,45,51,54).

Elia and Associates


majority of the studies the diabetes-specific formulas showed lower triglycer-ide concentrations than the standardformulas, the combined data from fourRCTs (14,38,44,45) indicated no signifi-cant effect (effect size �0.11 [95%CI�0.5 to 0.28], r � 0; effect size �0.13[�0.53 to 0.26], r � 0.5) (Table 1). Fur-thermore, two other long-term RCTs(49,54) reported no significant effect ofdiabetes-specific versus standard formu-las on triglycerides, whereas the findingsof a short-term RCT were unclear (48).None of these studies provided any de-tailed data.

Requirement for medicationThree RCTs (45,49,54) and one CCT (56)in patients with type 2 diabetes reportedreduced insulin requirements in those re-ceiving diabetes-specific formulas versusstandard formulas; two RCTs demon-strated statistical significance. In one RCT(49), patients fed diabetes-specific ETFhad significantly reduced insulin require-ments (from 38.7 units/day at study startto 32.7 units/day) compared with thosewho received a standard formula (44units/day throughout study); a differenceof 26% between the two groups. In theRCT (54) of critically ill hyperglycemicpatients, those receiving a diabetes-specific ETF had a significantly lower totalinsulin requirement compared with thosereceiving standard formula (median 8.73vs. 30.2 IU/day, respectively; a differenceof 71%), requirement per gram carbohy-drate ingested (median 0.07 vs. 0.18 IU/day) , and requirement per gramcarbohydrate ingested per kilogram bodyweight (median 0.98 vs. 2.13 IU/day). Ina further RCT (45), 25% of the patientsreceiving standard formulas needed tostart with regular insulin treatment, com-pared with none in the group receiving di-abetes-specific ETF. Nevertheless, thesefour studies provided insufficient compara-ble data to allow meta-analysis of the effectof diabetes-specific versus standard formu-las on the requirement for hypoglycemicmedication.

ComplicationsTwo RCTs (45,54) of ETF reported thisoutcome, and neither showed a signifi-cant difference in overall complicationrates between diabetes-specific and stan-dard formulas. However, post hoc 2

analysis of the data from one of these trials(45) demonstrated a tendency for a lower

Fig

.3—

The

effe

ctof

diab

etes

-spe

cific

vers

usst

anda

rdfo

rmul

ason

peak

bloo

dgl

ucos

eco

ncen

trat

ion

insh

ort-

term

stud

ies:

am

eta-

anal

ysis

oftw

oR

CT

s(4

7,48

).



incidence of urinary tract infections,pneumonia, and episodes of fever in thediabetes-specific versus standard group.The higher rate of skin infections in thediabetes-specific group was influenced byhigher rates at baseline. The data were notsufficiently comparable to permit meta-analysis of this outcome measure.

MortalityOnly one RCT of ETF in critically ill pa-tients (54) reported data on mortality. Nosignificant differences between patientsreceiving diabetes-specific and standardformulas were found in the 2-week studyperiod.

Other outcomesNo studies reported assessments of qual-ity of life or other functional measures.One RCT in critically ill patients (69) re-ported dietary intake. No significant dif-ferences in total dietary energy andnitrogen intakes were found betweenthose given diabetes-specific versus stan-dard formulas. One RCT (68) reportedanthropometric data. In this study, ONSsprovided 80% of total energy intake, andno significant differences in body weight,BMI, total body fat, or waist-to-hip ratiobetween those fed diabetes-specific ver-sus standard formulas were found.

CONCLUSIONS — This systematicreview (19 RCTs, 3 CCTs, and 1 CT)shows that the use of diabetes-specificcompared with standard formulas, givenas ONSs or ETF, consistently results insignificantly lower postprandial rise inblood glucose, peak blood glucose con-centrations, and glucose AUC in patientswith diabetes (Figs. 2– 4). This wasachieved without evidence of hypoglyce-mia; this suggests that glycemic controlmay be facilitated by the use of diabetes-specific enteral formulas compared withstandard formulas in patients with diabe-tes.

Compared with standard formulas,diabetes-specific formulas are typicallyhigher in fat (40–50% of energy, with alarge contribution from MUFAs, e.g.,�60% of fat), with a lower carbohydratecontent (�35–40% of energy) and up to15% of energy from fructose. These nu-trients could facilitate glycemic manage-ment by delaying gastric emptying (fatand fiber), delaying the intestinal absorp-tion of carbohydrate (fiber), and produc-ing smaller glycemic responses (fructose).

Fig.4

—T

heeffectofdiabetes-specific

versusstandard

formulas

onblood

glucoseA

UC

inshort-term

studies:am

eta-analysisoffour

RC

Ts

(40,46–

48).

Elia and Associates


A high proportion of MUFAs may alsohave beneficial effects on lipid profiles,but no significant effects were noted inour review. Due to the multinutrient na-ture of the formulation, it is difficult toassess which components of the diabetes-specific formulas were responsible for theeffects observed.

The impact of improved glycemiccontrol on long-term clinical outcomes iswell recognized in both type 1 (1) andtype 2 (2) diabetes. The current meta-analyses found that postprandial rise inglucose concentration was lower by 1.03mmol/l (95% CI 0.58 –1.47), and thepeak glucose concentration was reducedby 1.59 mmol/l (0.86–2.32), followingdiabetes-specific compared with standardformulas (Figs. 2 and 4). Recent studieshave demonstrated a strong correlationbetween postprandial glucose regulationand cardiovascular complications in pa-tients with diabetes (1,2,59 – 62), im-paired glucose tolerance (63,64), and all-cause mortality, whereas no suchcorrelation was demonstrated for fastingglucose control (62). Furthermore, a largeepidemiological study (65) has demon-strated that postprandial hyperglycemiais a better predictor of cardiovascular dis-ease than fasting glucose. This suggeststhat by improving glycemic control, thelong-term use of diabetes-specific versusstandard enteral formulas may reducecardiovascular complications in patientswith diabetes, although this was not as-sessed by the studies reviewed.

Patients with diabetes who are likelyto receive specific nutritional support on alonger term may include nursing homepatients, frail patients with infectiouscomplications, patients with slow-healingulcers or a history of falls and associatedfractures, and those in the pre- and post-operative period who are assessed to havepoor nutritional status. Improved glyce-mic control may also be important inacute care (e.g., stroke, intensive care),where hyperglycemia is associated with aworse outcome (3).

Intensive insulin therapy in criticallyill patients to maintain a glucose concen-tration of 4.4 – 6.1 mmol/l (comparedwith 10.0 –11.1 mmol/l in the controlgroup) improved mortality, blood streaminfections, requirement for transfusion,and critical illness polyneuropathy (66).The improved outcome was mainly due tothe lower blood glucose concentration(67,68) rather than insulin therapy (69),

with mechanisms likely to include os-motic effects and those involving genera-tion of free radicals and the immunesystem. In the study of critically ill pa-tients (54) included in this analysis, thediabetes-specific formula reduced bothinsulin dosage and circulating glucoseconcentrations; there were no significantdifferences in morbidity or mortality;however, the sample size and studylength may have been too small to detectsignificance.

In some studies, (one RCT and oneCCT), diabetes-specific formulas reducedthe quantity of hypoglycemic medicationand in some cases prevented the need forinsulin injections (45,56). Next to poten-tial health economic savings, these re-duced medication requirements may helpattenuate fluctuations in blood glucoseconcentrations and improve the quality oflife of these patients.

There are few long-term studies ex-amining clinical outcomes. One study ofETF (45) found that the diabetes-specificformula was associated with a trend to-ward reduced incidence of pneumonia,fever, and urinary tract infection relativeto the standard formula, which may haveclinical relevance for hyperglycemic pa-tients who are at increased risk of infec-tions. Further common comorbidities inpatients with diabetes include cardiovas-cular disease and hyperlipidemia. Al-though diabetes-specific feeds had ahigher fat content than standard feeds,this review suggests that diabetes-specificformulas had no detrimental effect on to-tal cholesterol, HDL, or triglycerides.There was inadequate information to ad-dress the effects on LDL in a meta-analysis.

For ETF studies, as details on theroute of administration or tube position-ing and the choice between continuous orbolus feeding regimens were insuffi-ciently reported, it was impossible to eval-uate how far the administration of thefeeds might have influenced the meta-bolic effects. A further consideration is theamount of feed administered, since pa-tients receiving ONSs may obtain only�25% of daily energy from this sourcecompared with up to 100% in tube-fedpatients.

National organizations (16,17) gener-ally recommend low-fat (25–35% of en-ergy) and high-carbohydrate diets (45–60%), rich in complex carbohydrates forthose with diabetes. The situation for

MUFAs is less clear, with the AmericanDiabetes Association reporting that thereis lack of evidence that MUFAs exert long-term effects on glucose control or othermetabolic parameters (16). In addition,formulas that have a particularly highproportion of fructose should probablybe given with some caution to critically illpatients, who are at risk of lactic acidosis.However, dietary therapy, including theuse of ONSs and ETF, given under med-ical supervision, can be individualized toinclude more liberal use of fat (such asMUFAs) (70). This may be particularlyimportant in the treatment of the mal-nourished patient, where an increased di-etary energy density may be important(71). Ultimately, there is a need to beguided by the desired clinical outcome forthe individual patient (45).

Due to the absence of RCT data com-paring the effects of nutritional supportwith routine care in patients with diabe-tes, this review primarily focused on theeffects of diabetes-specific versus stan-dard formula on metabolic control. Someof them were short-term studies in well-nourished individuals, although a num-ber of studies were in patients in need ofnutritional support (39,45,49,54,56,58).Many of the studies, however, scoredpoorly in methodology, but in some casesit may be difficult or unethical to under-take double-blinded placebo-controlledtrials with some nutritional support inter-ventions (e.g., tube feeding; oral supple-ment versus regular meal), and this mayaccount for lower RCT quality scores.There was insufficient data available toaddress the efficacy of nutritional sup-port, including diabetes-specific formu-las, according to diabetes type (type 1 ortype 2) or nutritional status.

There is clearly a need for further re-search in the form of well-designed, ade-quately powered trials that aim todetermine the role of enteral nutritionalsupport and diabetes-specific formulason the management, clinical outcome,and quality of life of malnourished pa-tients with diabetes. Furthermore, itwould be useful to establish the optimalcomposition of nutritional feeds designedto assist metabolic control, improve im-mune function, and achieve satisfactorynutritional status.

This systematic review shows that theuse of diabetes-specific oral and tube for-mulas (containing high proportions ofMUFAs, fructose, and fiber) are associ-



ated with improved glycemic controlcompared with standard formulas. In thelong term, this may aid the managementand outcome of patients with diabetes. Inparticular, cardiovascular complicationsmay be reduced, although research spe-cifically designed to examine these out-comes is warranted.

Acknowledgments— This study was con-ducted using an educational grant supplied byNumico.

Thanks to Abacus International, U.K., forresearch support and to Statwood, U.K., forstatistical support.

References1. Diabetes Control and Complications Trial

Research Group: The effect of intensivetreatment of diabetes on the developmentand progression of long-term complica-tion in insulin-dependent diabetes melli-tus. N Engl J Med 329:977–986, 1993

2. UK Prospective Diabetes Study (UKPDS)Group: Intensive blood-glucose controlwith sulphonylureas or insulin comparedwith conventional treatment and risk ofcomplications in patients with type 2 di-abetes (UKPDS 33). Lancet 352:837–853,1998

3. Coursin DB, Connery LE, Ketzler JT: Peri-operative diabetic and hyperglycemicmanagement issues. Crit Care Med 32(Suppl. 4):S116–S125, 2004

4. Clarke P, Gray A, Legood R, Briggs A, Hol-man R: The impact of diabetes-relatedcomplications on healthcare costs: resultsfrom the United Kingdom Prospective Di-abetes Study (UKPDS Study No. 65). Dia-bet Med 20:442–250, 2003

5. Logminiene Z, Norkus A, Valius L: Directand indirect diabetes costs in the world.Medicina (Kaunas) 40:16–26, 2004

6. Umpierrez GE, Isaacs SD, Bazargan N,You X, Thaler LM, Kitabchi AE: Hypergly-cemia: an independent marker of in-hos-pital mortality in patients with undiag-nosed diabetes. J Endocrinol Metab87:978–982, 2002

7. Jones BM, Stratton RJ, Holden C, RussellC, Glencorse G, Micklewright A: Trends inArtificial Nutritional Support in the UK2000–2003: Annual Report of the British Ar-tificial Nutrition Survey (BANS). Redditch,U.K., BAPEN, 2005

8. Campbell S, Schiller M: Considerationsfor enteral nutrition support of patientswith diabetes. Top Clin Nutr 7:23–32,1991

9. Coulston AM: Clinical experience withmodified enteral formulas for patientswith diabetes. Clin Nutr 17 (Suppl. 2):46–56, 1998

10. Koivisto VA, Yki-Jarvinen H: Fructoseand insulin sensitivity in patients withtype 2 diabetes. J Intern Med 233:145–153, 1993

11. Druetzler A, Bowen P, Cashmere K, Hor-witz D: Acute and chronic response ofglucose tolerance to a soy polysaccharideenriched liquid formula diet (Abstract).Fed Proc 44:1499, 1985 [abstract no.6369]

12. Garg A: High-monounsaturated-fat dietsfor patients with diabetes mellitus: ameta-analysis. Am J Clin Nutr 67 (Suppl.3):577S–582S, 1998

13. Hermansen K, Sondergaard M, Hoie L,Carstensen M, Brock B: Beneficial effectsof a soy-based dietary supplement onlipid levels and cardiovascular risk mark-ers in type 2 diabetic subjects. DiabetesCare 24:228–233, 2001

14. del Carmen Crespillo M, Olveira G, deAdana MS, Rojo-Martinez G, Garcia-Ale-man J, Olvera P, Soriguer F, Munoz A:Metabolic effects of an enteral nutritionformula for diabetes: comparison withstandard formulas in patients with type 1diabetes. Clin Nutr 22:483–487, 2003

15. Maxwell SR, Thomason H, Sandler D, Le-guen C, Baxter MA, Thorpe GH, Jones AF,Barnett AH: Antioxidant status in patientswith uncomplicated insulin-dependentand non-insulin-dependent diabetes mel-litus. Eur J Clin Invest 27:484–490, 1997

16. American Diabetes Association TaskForce for Writing Nutrition Principlesand Recommendations for the Manage-ment of Diabetes and Related Complica-tions: American Diabetes Associationposition statement: evidence-based nutri-tion principles and recommendations forthe treatment and prevention of diabetesand related complications. J Am Diet Assoc102:109–118, 2002

17. Recommendations for the nutritionalmanagement of patients with diabetesmellitus. Eur J Clin Nutr 54:353–355,2000

18. Mann JI, De Leeuw I, Hermansen K, Kara-manos B, Karlstrom B, Katsilambros N,Riccardi G, Rivellese AA, Rizkalla S, SlamaG, Toeller M, Uusitupa M, Vessby B, theDiabetes and Nutrition Study Group(DNSG) of the European Association: Ev-idence-based nutritional approaches tothe treatment and prevention of diabetesmellitus. Nutr Metab Cardiovasc Dis14:373–394, 2004

19. Benbow SJ, Hoyte R, Gill GV: Institutionaldietary provision for diabetic patients.QJM 94:27–30, 2001

20. Sinclair AJ, Gadsby R, Penfold S, CroxsonSC, Bayer AJ: Prevalence of diabetes incare home residents. Diabetes Care 24:1066–1068, 2001

21. Turnbull PJ, Sinclair AJ: Evaluation of nu-tritional status and its relationship with

functional status in older citizens with di-abetes mellitus using the mini nutritionalassessment (MNA) tool: a preliminary in-vestigation. J Nutr Health Aging 6:185–189,2002

22. Pastors JG, Franz MJ, Warshaw H, Daly A,Arnold MS: How effective is medical nu-trition therapy in diabetes care? J Am DietAssoc 103:827–831, 2003

23. The Cochrane Collaboration: CochraneReviewers Handbook 4.2.2 [handbookonline], 2004. Available at www.cochrane.org/resources/handbook/hbook.htm. Ac-cessed 22 July 2004

24. Centre for Reviews and Dissemination:Finding studies for systematic reviews: abasic checklist for researchers [article on-line], 2000. Available from www.york.ac.uk/inst/crd/revs.htm. Accessed 28 May2004

25. Centre for Reviews and Dissemination:Undertaking systematic reviews of re-search on effectiveness: CRD’s guidancefor those carrying out or commissioningreviews [article online]. Report no. 4. 2nded. Available from www.york.ac.uk/inst/crd/report4.htm. Accessed 28 May 2004

26. Moher D, Cook DJ, Eastwood S, Olkin I,Rennie D, Stroup DF: Improving the qual-ity of reports of meta-analyses of random-ised controlled trials: The QUORUMstatement. Quality of reporting meta-analyses. Lancet 354:1896–1900, 1999

27. PubMed [homepage on the Internet].Available from http://www.ncbi.nlm.nih.gov/entrez/query.fcgi. Accessed 10 Au-gust 2004

28. The Cochrane Library [homepage on theInternet]. Available from http://www.nelh.nhs.uk/cochrane.asp. Accessed 28July 2004

29. TRIP Database [homepage on the Inter-net]. Available from www.tripdatabase.com. Accessed 12 August 2004

30. Clinical Evidence Database [homepageon the Internet]. Available from www.clinicalevidence.com. Accessed 12 Au-gust 2004

31. National Electronic Library for Health[homepage on the Internet]. Available fromhttp://rms.nelh.nhs.uk/guidelinesfinder/.Accessed 12 August 2004

32. U.K. Department of Health: Nationalservice farmeworks [article online],2004. Available from http://www.dh.gov.uk/PolicyAndGuidance/HealthAndSocialCareTopics/HealthAndSocialCareArticle/fs/en?CONTENT_ID�4070951&chk�W3ar/W. Accessed 19 August 2004

33. Agency for Health Care Policy and Re-search: Acute pain management: opera-tive or medical procedures and trauma[article online], 1992. Available from http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid�hstat6.chapter.8991. Accessed 24 June2004

Elia and Associates


34. Jadad A, Moore R, Carroll D, Jenkinson C,Reynolds DJ, Gavaghan DJ, McQuay HJ:Assessing the quality of reports of ran-domized clinical trials: is blinding neces-sary? Control Clin Trials 17:1–12, 1996

35. Morgan WC, Bushman BJ: Integrating Re-sults Through Meta-Analytic Review UsingSAS Software. Cary, NC, SAS Inst., 1999

36. Mantel N, Haenszel W: Statistical aspectsof the analysis of data from retrospectivestudies of disease. J Natl Cancer Inst 22:719–748, 1959

37. Egger M, Davey Smith G, Schneider M,Minder C: Bias in meta-analysis detectedby a simple, graphical test. BMJ 315:629–634, 1997

38. McCargar LJ, Innis SM, Bowron E, Leich-ter J, Dawson K, Toth E, Wall K: Effect ofenteral nutritional products differing incarbohydrate and fat on indices of carbo-hydrate and lipid metabolism in patientswith NIDDM. Mol Cell Biochem 188:81–89, 1998

39. Galkowski J, Silverstone FA, Brad M: Useof a low carbohydrate with fiber enteralformula as a snack for elderly patientswith type II diabetes (Abstract). Clin Res37:89A, 1989

40. Printz H, Recke B, Fehmann HC, Goke B:No apparent benefit of liquid formula dietin NIDDM. Exp Clin Endocrinol Diabetes105:134–139, 1997

41. Sturmer W, Kramer E, Kasper H, Schre-zenmeir J: Favourable glycaemic effects ofa new balanced liquid diet for enteral nu-trition: results of a short term study in 30type II diabetic patients. Clin Nutr13:221–227, 1994

42. Thomas BL, Laine DC, Goetz FC: Glucoseand insulin response in diabetic subjects:acute effect of carbohydrate level and theaddition of soy polysaccharide in defined-formula diets. Am J Clin Nutr 48:1048–1052, 1988

43. Peters AL, Davidson MB: Effects of vari-ous enteral feeding products on postpran-dial blood glucose response in patientswith type I diabetes. JPEN J Parenter En-teral Nutr 16:69–74, 1992

44. Peters AL, Davidson MB, Isaac RM: Lackof glucose elevation after simulated tubefeeding with a low-carbohydrate, high-fatenteral formula in patients with type I di-abetes. Am J Med 87:178–182, 1989

45. Craig LD, Nicholson S, Silverstone FA,Kennedy RD: Use of a reduced-carbohy-drate, modified-fat enteral formula for im-proving metabolic control and clinicaloutcomes in long-term care residents withtype 2 diabetes: results of a pilot trial. Nu-trition 14:529–534, 1998

46. Golay A, Schneider H, Bloise D, Vadas L,Assai JP: The effect of a liquid Suppl. con-taining guar gum and fructose on glucosetolerance in non-insulin dependent dia-betic patients. Nutr Metab Cardiovasc Dis

5:141–148, 199547. Hofman Z, Rouws C, van Drunen J, De

Later C, Kuipers H: The effect of enteralnutrition on glucose and triglyceride con-centrations during 6h continuous feedingin diabetic patients. In 26th ESPEN (Euro-pean Society Clinical Nutrition and Metabo-lism) Congress, 2004. Lisbon, Portugal,11–14 September 2004

48. Hofman Z, van Drunen JDE, de Later C,Kulpers H: The effect of different nutri-tional feeds on the postprandial glucoseresponse in healthy volunteers and in pa-tients with type II diabetes. Eur J Clin Nutr58:1553–1556, 2004

49. Mayr P, Mertl-Roetzer M, Lauster F, PohlM, Haslbeck M, Eriksen J, Rahlfs VW:Metabolic control in type 2 diabetes tubefed patients after brain damage duringlong-term treatment with a new low car-bohydrate, high monounsaturated fattyacid containing enteral formula versus astandard-like formula: a randomised,prospective controlled, double blindmulti centre trial. Clin Nutr 23:1497–1498, 2005

50. Wang H, Jiang Z: The impact of slow re-lease starch containing enteral nutritionon blood glucose/insulin responses intype II diabetic patients: randomised con-trolled multi centre clinical study (Ab-stract). Clin Nutr 22 (Suppl. 1):S15, 2003[abstract 056-O]

51. Sanz-Paris A, Calvo L, Guallard A, SalazarI, Albero R: High-fat versus high-carbohy-drate enteral formulas: effect on bloodglucose, C-peptide, and ketones in pa-tients with type 2 diabetes treated withinsulin or sulfonylurea. Nutrition 14:840–845, 1998

52. Fix BM, Lowe W, Cockram DB, Craig LD:Effect of a liquid nutritional Suppl. con-taining a novel carbohydrate system onglucose tolerance in subjects with type IIdiabetes (Abstract). Ann Nutr Metab 45(Suppl. 1): 2004 [abstract 3.07.115]

53. Sanz A, Albero R, Playan J, Acha FJ, Casa-mayor L, Celaya S: Comparison of a highcomplex carbohydrate enteral formulawith a high mono-unsaturated fat formulain patients with type I diabetes mellitustreated with insulin or sulphonylurea(Abstract). JPEN J Parenter Enteral Nutr 18(Suppl. 1):31S, 1994 [A90]

54. Mesejo A, Acosta JA, Ortega C, Vila J, Fer-nandez M, Ferreres J, Sanchis JC, Lopez F:Comparison of a high-protein disease-specific enteral formula with a high-pro-tein enteral formula in hyperglycemiccritically ill patients. Clin Nutr 22:295–305, 2003

55. Harley JR, Pohl SL, Isaac RM: Low carbo-hydrate with fiber versus high carbohy-drate without fiber enteral formulas:effect on blood glucose excursion in pa-tients with type II diabetes (Abstract). Clin

Res 37:141A, 198956. Grahm TW, Harrington TR, Isaac RM:

Low carbohydrate (CHO) with fiber en-teral formula impedes development of hy-perglycaemia in patients with acute headinjury (Abstract). Clin Res 37:138A, 1989

57. Peters AL, Davidson MB: Addition of hy-drolyzed guar to enteral feeding productsin type I diabetic patients. Diabetes Care19:899–900, 1996

58. Abbruzzese B: Effect of a low-CHO for-mula on circadian glycaemic control andlipids in elderly patients with non insulindependent diabetes mellitus (NIDDM) re-ceiving total enteral nutrition support(Abstract). FASEB J 7:A847, 1993 [ab-stract 4891]

59. Ceriello A, Hanefeld M, Leiter L, MonnierL, Moses A, Owens D, Tajima N, Tuom-ilehto J: Postprandial glucose regulationand diabetic complications. Arch InternMed 164:2090–2095, 2004

60. Fonseca V: Clinical significance of target-ing postpranidal and fasting hyperglyce-mia in managing type 2 diabetes mellitus.Curr Med Research Opin 19:635–641,2003

61. Heine RJ, Balkau B, Ceriello A, Del PratoS, Horton ES, Taskinen MR: What doespostprandial hyperglycaemia mean? Dia-bet Med 21:208–213, 2004

62. Hanefeld M: STOP-NIDDM: a new para-digm for diabetes prevention? Nutr MetabCardiovasc Dis 12:253–258, 2002

63. Chiasson JL, Josse RG, Gomis R, HanefeldM, Karasik A, Laakso M: Acarbose for pre-vention of type 2 diabetes mellitus: theSTOP-NIDDM randomised trial. Lancet359:2072–2077, 2002

64. Chiasson JL, Josse RG, Gomis R, HanefeldM, Karasik A, Laakso M: Acarbose treat-ment and the risk of cardiovascular dis-ease and hypertension in patients withimpaired glucose tolerance: the STOP-NIDDM trial. JAMA 290:486–494, 2003

65. Balkau B: The DECODE study. Diabetesepidemiology: collaborative analysis ofdiagnostic criteria in Europe. DiabetesMetab 26:282–286, 2000

66. van den Berghe G, Wouters P, Weekers F,Verwaest C, Bruyninckx F, Schetz M,Vlasselaers D, Ferdinande P, Lauwers P,Bouillon R: Intensive insulin therapy inthe critically ill patients. N Engl J Med 345:1359–1367, 2001

67. Van den Berghe G, Bouillon R, Weekers F,Verwaest C, Schetz M, Vlasselaers D, Fer-dinande P, Lauwers P: Outcome benefit ofintensive insulin therapy in the criticallyill: insulin dose versus glycemic control.Crit Care Med 31:359–366, 2003

68. Finney SJ, Zekveld C, Elia A, ET: Glucosecontrol and mortality in critically ill pa-tients. JAMA 290:2041–2047, 2003

69. Annane D, Melchior JC: Hormone re-placement therapy for the critically ill.



Crit Care Med 31:634–635, 200370. Consensus roundtable on nutrition sup-

port of tube-fed patients with diabetes(Consensus Statement). Clin Nutr 17

(Suppl. 2):63–65, 199871. Stratton RJ, Green CJ, Elia M: Prevalence

of disease-related malnutrition. In Dis-ease-Related Malnutrition: An Evidence-

Based Approach to Treatment. Stratton RJ,Green CJ, Elia M, Eds. Wallingford,Oxon, U.K., CABI Publishing, 2003, p.35–92

Elia and Associates


Enteral Nutritional Support and Use of Diabetes-Speciﬁc ... · formulas. CONCLUSIONS— This systematic review shows that short- and long-term use of diabetes-speciﬁc formulas

Documents