Value of homocysteine levels, troponin I, and score for neonatal acute physiology and perinatal extension II as early predictors of morbidity

JOBNAME: No Job Name PAGE: 1 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: B99FB366/v2503/blackwell/journals/ped_v0_i0/ped_3485

Original Articleped_3485 1..9

Value of homocysteine levels, troponin I, and score for neonatal acutephysiology and perinatal extension II as early predictors of morbidity

Hale Usluer,1 Gulcan Turker2 and Ayse Sevim Gokalp2

1Department of Pediatrics and 2Neonatology Division, Kocaeli University, Faculty of Medical, Kocaeli, Turkey

Abstract Background: Although the effects of cardiac troponin I (cTnI) have been documented in infants, the associations amongthe value of maternal and cord blood total homocysteine (Hcy) levels, cord blood cTnI and the score for neonatal acutephysiology and perinatal extension II (SNAPPE-II) values in infants have not been documented. The aim of this studywas to determine the value of maternal total Hcy (mtHcy) and cord blood total Hcy (ctHcy) levels, cTnI and SNAPPE-IIas predictive factors of morbidity in newborns.Methods: Maternal and cord blood samples were routinely collected for analysis from all prospective participatingvolunteers. In this case–control study, both hospitalized (n = 71) and non-hospitalized (n = 148) newborns were identifiedand followed until discharge.Results: Regression analysis revealed that pre-eclampsia, gestational age, mtHcy and SNAPPE-II values were signifi-cantly and independently associated with morbidity. Cord blood pH, ctHcy levels, SNAPPE-II values, and pre-eclampsiawere associated with mtHcy levels. A similar association was found among cTnI, cord blood pH, mtHcy levels and ctHcylevels. The specificity and sensitivity values of mtHcy in predicting newborn morbidity were 62% and 78%, respectively,while the specificity and sensitivity values of SNAPPE-II in predicting newborn morbidity were 96% and 97%,respectively.Conclusions: Elevated mtHcy levels were associated with cTnI, SNAPPE-II values, cord blood gas, and neonatalmorbidities. These results suggest that SNAPPE-II may be an early predictor of morbidity after delivery and that elevatedmtHcy levels may be an early prenatal biomarker of morbidity in newborns.

Key words cord blood gas, homocysteine, morbidity, SNAPPE-II, troponin I.

Homocysteine (Hcy) is a sulfur-containing amino acid formedfrom the essential amino acid methionine. Hcy is associated withincreased lipid peroxidation, plasma triglycerides, and serum uricacid as well as with blunted endothelial-dependent vasorelax-ation and alteration of the anticoagulant endothelium to a proco-agulant phase.1 These Hcy-mediated vascular and metabolicchanges are similar to those associated with pre-eclampsia.2

Elevated maternal total Hcy (mtHcy) levels are associated withthe severity of pre-eclampsia3,4 in mothers and with prematur-ity5 and low-birthweight4 in infants. In addition, mtHcy levelsare associated with placental abruption, early recurrent fetalloss, intrauterine growth retardation (IUGR), and neural tubedefects.6–8 Maternal levels of Hcy correlate directly with neonatalHcy levels,3,9 and studies show that cord blood total Hcy (ctHcy)levels in infants are dependent on gestational age at birth, post-natal age and neonatal diet.4,5,9,10–12 Despite these documentedeffects of tHcy on a pregnancy and the fetus, very little is known

about the role of tHcy in infants; in particular, little is knownabout the associations between tHcy and morbidity, cord bloodgas, Apgar score, respiratory distress syndrome (RDS) and tran-sient tachypnea.

Cardiac troponins (cTn) are useful markers for the detectionof myocardial ischemic damage in cases of asphyxia13,14 or incritically ill newborns.15 Brasselet et al.16 found a correlationbetween cardiac troponin I (cTnI) levels and pre-percutaneouscoronary intervention plasma tHcy concentration in cases ofmyocardial infarction after percutaneous coronary intervention.In addition, Bodi et al.17 found that cTnI and tHcy levels shouldbe combined with traditional clinical markers (age, heart failureand diabetes) to best determine risk after an acute coronary event.Despite the documented effects of cTnI on pregnancy and a fetus,very little is known about its role in newborns; we are especiallyinterested in whether the association between tHcy and cTnI canbe used to predict morbidity in newborns.

The score for neonatal acute physiology and perinatal exten-sion II (SNAPPE-II) has been used to predict mortality in new-borns.18 However, cord blood cTnI was found to be a moresensitive predictor of mortality in critically ill newborns than thescore for neonatal acute physiology.15 In addition, cTnI is an earlypredictor of fetal chronic hypoxia related to pre-eclampsia.19

Correspondence: Gulcan Turker, MD, PhD, Kocaeli ÜniversitesiTıp Fakültesi Çocuk Saglıgı ve Hastalıkları Anabilim Dalı, Umut-tepe Kampusü, 41380 Kocaeli, Turkey. Email: [email protected], [email protected]

Received 18 February 2010; revised 1 October 2010; accepted 8September 2011.

Pediatrics International (2011) ••, ••–•• doi: 10.1111/j.1442-200X.2011.03485.x

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

1

2

345

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

4647484950515253

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

JOBNAME: No Job Name PAGE: 2 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: 91315354/v2503/blackwell/journals/ped_v0_i0/ped_3485

Although the effects of cord blood cTnI have been documented ininfants, the associations among mtHcy, ctHcy, cord blood cTnIand SNAPPE-II values in infants have not been documented.

The purpose of this study was to explain this discrepancy andto determine the value of using mtHcy, ctHcy, cTnI level andSNAPPE-II to predict newborn morbidity. The results will becompared to the predictive value of SNAPPE-II alone.

Methods

Design

This prospective case–control study was performed in the Unit ofNeonatology at the Faculty of Medicine of Kocaeli University incollaboration with the Department of Obstetrics and Gynecology.The study group consisted of successive mothers admitted to ourhospital for delivery who volunteered to participate. A total of252 newborns were born in the tertiary Obstetric and GynecologyClinic of Kocaeli University Hospital during the study period.Eighty-two newborns were hospitalized during the study period.Eleven hospitalized and 18 non-hospitalized infants wereexcluded from the study because parental consent was notobtained. Four infants with congenital anomalies were notenrolled. The study was performed with the approval of theMedical Faculty’s ethics committee; in accordance with the Dec-laration of Helsinki, informed consent forms were obtained fromall participating volunteers.

All of the newborns included in this study were born at ouruniversity hospital. The gestational age of these newborns rangedfrom 24 to 41 weeks.

During the study period, cord blood samples were routinelycollected in the delivery room. Arterial cord blood samples wereobtained from all participating volunteers for blood gas analysisand venous cord blood samples were obtained to measure ctHcyand cTnI levels. Maternal blood samples were obtained from allparticipating volunteers 2 h before delivery.

Study population group

All newborns were from singleton pregnancies. Newborns werefollowed in our department until discharge or death and werecategorized as hospitalized (n = 71) or non-hospitalized (n =148). None of the mothers took folic acid supplements or vitaminB12 during either the preconception period or during the preg-nancy. In the hospitalized group, 19 of the preterm newborns hadsevere RDS, and six term newborns had mild or moderatehypoxic-ischemic encephalopathy (grade I and II according tothe Sarnat classification); three preterm newborns had septicshock and a positive blood culture, four term and seven latepreterm newborns had transient tachypnea, three term newbornshad congenital heart disease, five term newborns had hyperbiliru-binemia, and 14 term newborns with maternal pre-eclampsia hadhypoglycemia or hypocalcemia.

Data collection

Face-to-face interviews were conducted with all of the partici-pating mothers at the hospital after delivery. Interview questionsaddressed potential confounding factors, such as demographicfactors before and after pregnancy, disease during pregnancy,

maternal chronic disease, obstetric history, smoking, and the useof multivitamin supplements. Data collected for the infantsincluded gestational age, birthweight, sex, Apgar score, mode ofdelivery, intrapartum cardiotocography, cTnI, arterial cord bloodgas and SNAPPE-II.18

SNAPPE-II

SNAPPE-II scores were calculated from nine variables: meanblood pressure, temperature, PaO2/FiO2, cord blood pH, multipleseizures, urine output, birthweight, being small-for-gestationalage (SGA), and Apgar score at 5 min. The numeric value ofSNAPPE-II is defined as the arithmetic sum of the pointsassigned to each variable based on the worst value. It is calcu-lated from data collected during the first 12 h of admission.18 Ifurine was confirmed but urine output was not measured, a scoreof 0 was assumed.18

Pre-eclampsia

Obstetricians used criteria established by the American Collegeof Obstetricians and Gynecologists to identify pre-eclampsia.20

Pre-eclampsia was defined as an increase in blood pressure above140/90 in a previously normotensive woman after the 20th ges-tational week on at least two occasions 6 h apart, combined withproteinuria. Proteinuria was defined as a protein dipstick readinggreater than 1+ on more than one midstream urine sample 6 hapart, or a 24-h urine excretion of greater than 0.3 g protein in theabsence of urinary infection.

Criteria for rescue treatment of RDS

Rescue treatment with surfactant was given to preterm infantswith RDS who needed intubation and assisted ventilation andwho did not present radiological evidence of another diseaseprocess.

IUGR

Birthweight, height and head circumference percentiles wereevaluated according to Lubchenco’s criteria.21 Infants whosebirthweights were under the 10th percentile were identified withIUGR.

tHcy analysis

The umbilical cord was double clamped near the placenta and theinfant immediately after delivery of the placenta. Blood fromumbilical arteries was separated into vials containing ethylene-diaminetetraacetic acid. The sample was placed on ice immedi-ately and centrifuged at 3000 g for 30 min to separate the plasma.The tHcy concentration was measured in an Immulite 2000immunoassay analyzer (Siemens, Los Angeles, CA, USA) usingHomocysteine Bead Pack L2H012153 and HomocysteineReagent Wedge L2H0A2154 kits (Los Angeles, CA, USA) and acompetitive immunoassay. The lower and upper limits for thedetection of tHcy were 2 mmol/L and 50 mmol/L, respectively.According to the manufacturer, interference with the competitiveimmunoassay can occur in the presence of substances, such asserum bilirubin (greater than 200 mg/L), hemoglobin (less than

2 H Usluer et al.

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

1

2

3

4

5

6

789

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

JOBNAME: No Job Name PAGE: 3 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: A351B2DE/v2503/blackwell/journals/ped_v0_i0/ped_3485

3 g/dL), and cholesterol (greater than 512 mg/dL). For this assay,the within-run coefficient of variation is 8% and the run totalcoefficient of variation is 10%.

cTnI analysis

The cTnI levels were measured using an AxSYM System ana-lyzer with the Abbott cTnI microparticle enzyme immunoassay(Abbott Park, IL, USA). For this assay, the within-run coefficientof variation is 6.6%, and the manufacturer claims minimal cross-reactivity with cardiac troponin C (0.01%), cardiac troponin T(0.34%) and skeletal troponin I (0.04%) at a concentration of1000 ng/mL. The lower limit for the detection of cTnI was0.1 ng/mL.

According to the manufacturer’s instructions, a serum cTnIlevel of 2 ng/mL indicates myocardial injury in adults. Less than1 mL of cord blood was required for the cTnI assay. According tothe manufacturer, the cTnI AxSYM I assay has less than 10%interference in the presence of substances such as serum bilirubin(greater than 20 mg/dL), hemoglobin (greater than 1000 mg/dL),cholesterol (greater than 500 mg/dL), triglycerides (greater than1000 mg/dL) and protein (greater than 10 g/dL).

Statistical analysis

For many variables, the data showed wide standard deviationsand abnormal distributions. The statistical analyses were there-fore carried out by calculating medians and minimum–maximumvalues. Differences between groups were analyzed using theMann–Whitney U-test or the c2-test. The correlations betweenmtHcy and ctHcy levels, cTnI and other study parameters wereevaluated using Spearman’s rank correlation coefficient for con-tinuous variables. Variables that were significant in the univariateanalysis were entered into a backward conditional logistic regres-sion model selection process for the independent predictors ofmorbidity. Variables that were significant in the univariate analy-sis were entered into a backward conditional linear regressionmodel selection process for the independent predictors of mtHcyand ctHcy. Pre-eclampsia, prematurity, birthweight, gestational

age, sex, Apgar score at 1, 5 and 10 min, arterial cord blood basedeficit, arterial cord blood pH, cTnI levels and SNAPPE-II valueswere entered into the regression as independent variables relatedto mtHcy and ctHcy. In addition to RDS, transient tachypnea andIUGR were entered into the regression as independent variablesrelated to morbidity.

Results

Birthweight, gestational age, Apgar scores, arterial cord bloodpH, cord blood base deficit, SNAPPE-II values and mtHcy levelswere higher in the hospitalized group than in the non-hospitalizedgroup (Table 1). Sex, delivery mode, ctHcy and cTnI levels didnot differ between the two groups. The level of mtHcy wascorrelated with SNAPPE-II values (r = 0.589, P < 0.0001), Apgarscore at 1 min (r = -0.229, P = 0.001), cord blood pH (r = -0.264,P = 0.003), cord blood base excess (r = -0.225, P = 0.014) andcTnI (r = 0.166, P = 0.014). In a simple correlation, gestationalage was associated with mHcy for all patients (r = -0.19, P =0.002). Analysis revealed no correlations between ctHcy leveland sex, birthweight or mode of delivery, Apgar score at 1 min,cord blood pH or cord blood base excess (Spearman’s rank cor-relation coefficient, P > 0.05 for all). The level of ctHcy wascorrelated only with cTnI levels (r = 0.240, P < 0.0001). All of thehospitalized term newborns who had mothers with pre-eclampsiahad hypoglycemia or hypocalcemia. All of the newborns withtransient tachypnea were delivered by elective cesarean section.

In model 1, backward stepwise logistic regression analysisrevealed that after controlling for potentially confoundingfactors, pre-eclampsia, gestational age, SNAPPE-II and mtHcywere significantly and independently associated with morbidity.No similar association was found with birthweight, sex, Apgarscore at 1 and 5 min, arterial cord blood base deficit, arterial cordblood pH, and ctHcy levels (Table 2). In model 2, backwardstepwise logistic regression analysis revealed that after control-ling for potentially confounding factors, IUGR, base deficit,SNAPPE-II, mtHcy, RDS and prematurity were significantly andindependently associated with morbidity. No similar association

Table 1 Characteristics of the hospitalized and non-hospitalized newborns

Hospitalizedn = 71

Non-hospitalizedn = 148

P-value

Number (preterm/term) 39/32 22/126 <0.0001*Sex (male/female) (n) 42/29 79/69 0.5*Delivery mode (vaginal/cesarean section) 33/38 92/56 0.9*Pre-eclampsia 23 18Birthweight (g) 2450 (600–4100) 3200 (2040–4850) <0.0001**Gestational age (weeks) 36 (24–41) 38 (32–41) <0.0001**Apgar score at 1 min 8 (1–10) 9 (1–10) <0.0001**Apgar score at 5 min 9 (2–10) 9 (1–10) 0.008**Cord arterial blood base deficit (mmol/L) -1.65 (-22–2.8) -2.4 (-7.6–2.5) 0.7**Cord arterial blood pH 7.29 (6.7–7.41) 7.31 (7.16–7.45) 0.021**SNAPPE-II 5 (0–24) 0 (0–5) <0.0001**Cord venous blood cardiac troponin I (ng/ml) 0.6 (0–2.1) 0.2 (0–2) 0.06**Maternal total homocysteine levels (mmol/L) 5.84 (2.2–16.9) 4.84 (2–14.4) <0.0001**Cord venous blood total homocysteine levels (mmol/L) 6.5 (2–19.7) 5.74 (2–16.4) 0.058**

Bold values indicate statistical significance. *c2-test was used. **Expressed as median (min–max) and analyzed using the Mann–Whitney U-test.SNAPPE-II, score for neonatal acute physiology and perinatal extension II.

Homocysteine and troponin I in newborn 3

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

1

2

345

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

212223

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39404142

4344454647484950515253545556

5758

59

60

61

62

63

646566

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

JOBNAME: No Job Name PAGE: 4 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: AF7C0B53/v2503/blackwell/journals/ped_v0_i0/ped_3485

was found with arterial cord blood pH, ctHcy levels, cTnI levelsand transient tachypnea (Table 2).

After controlling for potentially confounding factors, back-ward linear regression analysis revealed that arterial cord bloodpH, SNAPPE-II, pre-eclampsia and ctHcy levels in model 1,IUGR, cTnI, ctHcy and hospitalization in model 2, and RDS inmodel 3 were significantly and independently associated withmtHcy levels (Table 3). Gestational age, arterial cord blood basedeficit, duration of hospitalization and transient tachypnea wereexcluded as independent predictors of mtHcy (Table 3).

Regression analysis also revealed that cord blood pH, cTnI,and mtHcy levels in model 1, cTnI, RDS, mtHcy, and IUGR inmodel 2, and SNAPPE-II, cTnI, RDS and mtHcy in model 3 weresignificantly and independently associated with ctHcy levels(Table 4).

Gestational age, Apgar score at 1, 5 and 10 min, arterial cordblood base deficit, duration of hospitalization, and transient tac-hypnea were excluded as independent predictors of ctHcy(Table 4).

The receiver–operator characteristic (ROC) curve analysisshowed that a mtHcy level of 4.4 mmol/L predicted morbidity.The positive and negative predictive values of mtHcy in predict-ing morbidity were 80% and 61%, respectively. The specificityand sensitivity values of mtHcy in predicting morbidity were62% and 78%, respectively (area under the curve is 0.64; stan-dard error 0.04; 95%CI: 0.56–0.72). The ROC curve analysisshowed that a SNAPPE-II value of 4 predicted morbidity. Thepositive and negative predictive values of SNAPPE-II in predict-ing morbidity were 92% and 81%, respectively. The specificity

and sensitivity values of SNAPPE-II in predicting morbiditywere 96% and 97%, respectively (area under the curve is 0.76;standard error 0.036; 95%CI: 0.68–0.83) (Fig. 1). SNAPPE-IIhad the highest sensitivity and specificity in predicting morbidityat birth.

Discussion

The purpose of this study was to identify any association betweenmorbidity and mtHcy and among mtHcy levels, cord blood cTnI,cord blood gas and SNAPPE-II in newborns.

The effects of Hcy on endothelial function include decreasedbioavailability of nitric oxide, increased oxidative stress, stimu-lation of smooth cell proliferation, and increased expression ofinflammatory cytokines by the vascular endothelium.22 Tasatargilet al.23 showed that Hcy causes a significant alteration in thevascular reactivity of pulmonary arteries. This alteration seemedto occur via oxidative stress in the pulmonary artery endotheliumwith subsequent deoxyribonucleic acid damage and activationof the polyadenosine diphosphate-ribose polymerase pathway.Increased tHcy levels and decreased serum folate levels are cor-related with the occurrence of neonatal asphyxia. In addition,significantly increased serum tHcy may be found in asphyxiatedneonates.24 However, neonatal encephalopathy in pre-eclampsiais not only dependent on severe intrauterine hypoxemia. Neuro-toxicity is also associated with the dual actions of Hcy at theN-methyl-D-aspartic acid receptor.25 The present study found thatan increased mtHcy level is also correlated with cord blood gas,Apgar scores and cTnI. Significantly elevated mtHcy and ctHcylevels may therefore be associated with chronic or acute hypoxia.

Table 2 Risk factors for morbidity

B P-value OR 95%CIModel 1

Over all constant -1 <0.0001Pre-eclampsia -2.54 0.029 0.07 0.0079–0.77Gestational age (weeks) -0.47 0.011 0.62 0.43–0.89SNAPPE-II 0.79 0.0015 2.21 1.35–3.62Maternal total homocysteine levels (mmol/L) 0.26 0.039 1.29 1.01–1.66Constant 107.7 0.015

Model 2Over all constant -7 <0.0001IUGR 2.01 0.01 7.51 1.6–35.2Base deficit 0.19 0.049 1.2 1.06–1.4SNAPPE-II 0.67 0.038 1.95 1.24–3.08Maternal total homocysteine levels (mmol/L) 0.15 0.043 1.16 1,0046–1.36RDS -3,66 0.027 0.025 0.001–0.66Prematurity 3.5 0.0019 34.5 3.69–322Constant 100,6 0.002

Bold values indicate statistical significance. Model 1: Backward conditional (stepwise) logistic regression analysis revealed that pre-eclampsia,gestational age, SNAPPE-II, and maternal total homocysteine levels were significantly and independently associated with morbidity. Othervariables: prematurity, birthweight, gestational age, sex, Apgar score at 1, 5 and 10 min, arterial cord blood base deficit, arterial cord blood pH,maternal and cord blood total homocysteine levels, cardiac troponin I levels and SNAPPE-II values were entered into the regression as theindependent variables; morbidity was the dependent variable. Constant: morbidity. Model 2: Backward conditional (stepwise) logistic regressionanalysis revealed that IUGR, gestational age, arterial cord blood base deficit, SNAPPE-II values, maternal total homocysteine levels, RDS andprematurity were significantly and independently associated with morbidity. Other variables: arterial cord blood base deficit, arterial cord blood pH,cord blood total homocysteine levels, cardiac troponin I levels and transient tachypnea. B, Estimate of the change in the dependent variable that canbe attributed to a change of one unit in the independent variable; IUGR, intrauterine growth retardation; RDS, respiratory distress syndrome;SNAPPE-II, score for neonatal acute physiology and perinatal extension II.

4 H Usluer et al.

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

123

456789

10111213141516171819

20212223242526272829

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

646566

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

JOBNAME: No Job Name PAGE: 5 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: BDD38349/v2503/blackwell/journals/ped_v0_i0/ped_3485

Table 3 Independent predictors of maternal homocysteine level

OR P-value 95%CIModel 1

(Constant) 49,9 <0.0001 23.08–76,71Cord arterial blood pH -6.52 0.001 -10.2–-2.75SNAPPE-II 0.13 0.001 -10.28–-2.75Pre-eclampsia 1,65 <0.0001 0.74–2.56Cord venous blood total homocysteine levels (mmol/L) 0.57 <0.0001 0.47–0.68

Model 2Constant 1.31 0.002 0.478–-2.151SNAPPE-II 0.11 0.005 0.036–0.2Cord venous blood cardiac troponin I (ng/ml) 0.77 0.013 0.17–1.375Hospitalization 0.79 0.047 0.011–-0.310IUGR -1.38 0.012 -2,456–-0.31Cord venous blood total homocysteine levels (mmol/L) 0.58 <0.0001 0.476–0.69

Model 3Constant 41.398 0.001 18.135–64.661Cord venous blood total homocysteine levels (mmol/L) .642 <0.0001 0.556–0.728SNAPPE-II .072 0.077 -0.008–0.153Cord arterial blood pH -5.472 0.001 -8.655–(-2.288)Cord venous blood cardiac troponin I (ng/ml) 0.449 0.047 0.007–0.891RDS 1.19 0.017 0.21–2.17

Bold values indicate statistical significance. Model 1: Backward conditional linear regression analysis revealed that cord blood pH, SNAPPE-IIvalues, pre-eclampsia and cord blood total homocysteine levels were significantly and independently associated with maternal total homocysteinelevels. Constant: maternal total homocysteine levels. Other variables: Gestational age, Apgar score at 1, 5 and 10 min, arterial cord blood basedeficit, cardiac troponin I and duration of hospitalization. Model 2: Backward conditional linear regression analysis revealed that venous cord bloodcardiac troponin I, hospitalization, IUGR and venous cord blood total homocysteine levels were significantly and independently associated withmaternal total homocysteine levels. Constant: maternal total homocysteine levels. Other variables: Gestational age, RDS, transient tachypnea andprematurity. Model 3: Backward conditional linear regression analysis revealed that venous cord blood total homocysteine levels, SNAPPE-IIvalues, arterial cord blood pH, venous cord blood cardiac troponin I, and RDS were significantly and independently associated with maternal totalhomocysteine levels. Constant: maternal total homocysteine levels. Other variables: Gestational age, transient tachypnea and prematurity. Unstand-ardized Coefficients B, regression coefficients used to compute the regression equation. IUGR, intrauterine growth retardation; RDS, respiratorydistress syndrome; SNAPPE-II, score for neonatal acute physiology and perinatal extension II.

Table 4 Independent predictors of cord blood homocysteine level

OR P-value 95%CIModel 1

Constant -41.9 0.013 -73.57–-8.99Cord arterial blood pH 5.87 0.009 1.48–10.26Cord venous blood cardiac troponin I (ng/ml) 1.19 <0.0001 0.53–1.84Maternal blood total homocysteine levels (mmol/L) 0.74 <0.0001 0.59–0.88

Model 2Constant 1.811 <0.0001 1.200Cord venous blood cardiac troponin I (ng/ml) 0.805 0.001 0.322RDS -1.627 0.001 -2.541Maternal blood total homocysteine levels (mmol/L) 0.761 0.000 0.657IUGR 1.040 0.034 0.080

Model 3Constant 1.876 <0.0001 1.268–2.485SNAPPE-II -0.040 0.381 -0.130–0.050Cord venous blood cardiac troponin I (ng/ml) 0.864 0.001 0.380–1.347RDS -1.301 0.020 -2.395–-0.208Maternal venous blood total homocysteine levels (mmol/L) 0.762 <0.0001 0.657–0.868

Bold values indicate statistical significance. Model 1: Backward conditional linear regression analysis revealed that cord blood pH, cardiactroponin I, and maternal total homocysteine levels were significantly and independently associated with venous cord blood total homocysteinelevels. Constant: cord venous blood total homocysteine levels. Other variables: Gestational age, Apgar score at 1, 5 and 10 min, arterial cord bloodbase deficit, duration of hospitalization, and SNAPPE-II. Model 2: Backward conditional linear regression analysis revealed that cord blood pH andmaternal total homocysteine levels were significantly and independently associated with venous cord blood total homocysteine levels, RDS andIUGR. Constant: venous cord blood total homocysteine levels. Other variables: Gender, hospitalization and gestational age. Model 3: Backwardconditional linear regression analysis revealed that SNAPPE-II values, venous cord blood cardiac troponin I, RDS and maternal venous blood totalhomocysteine levels were significantly and independently associated with venous cord blood total homocysteine levels. Constant: venous cordblood total homocysteine levels. Other variables: Gestational age and transient tachypnea. IUGR, intrauterine growth retardation; RDS, respiratorydistress syndrome; SNAPPE-II, score for neonatal acute physiology and perinatal extension II.

Homocysteine and troponin I in newborn 5

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

123

456789

1011121314151617181920212223

2425262728293031323334

35

363738

3940414243444546474849505152535455

56575859606162636465

JOBNAME: No Job Name PAGE: 6 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: 6B42048F/v2503/blackwell/journals/ped_v0_i0/ped_3485

Cho et al.26 found that elevated concentrations of brain natri-uretic factor, C-reactive protein, Hcy and the presence of theMTHFR C677T mutation independently contributed to theability to predict cardiovascular disease in adults.26 Previousstudies have also shown that hyperhomocysteinemia may restrictfetal growth3–5 and increase fetal mortality.3 We found an asso-ciation among elevated mtHcy levels, ctHcy, cord blood pH andcTnI. Our results confirm that elevated mtHcy may be a riskfactor for perinatal hypoxia.

High SNAPPE-II values assessed in the first day of life, theseverity of RDS, birthweight and gestational age correlated wellwith outcomes such as mortality, morbidity and the duration ofhospitalization. SNAPPE-II values can also be a used to control

for differences in the degree of illness and risk between neona-tal intensive care units in multicenter studies.18 Kahn et al.27

identified a correlation between SNAPPE-II values and thromb-ocytopenia, delivery mode, and intraventricular hemorrhage inlow-birthweight infants. Lım and Rozyckı demonstrated a corre-lation between the score and sepsis, necrotizing enterocolitis, andmortality.28 Griffin and Moorman found a positive relationshipbetween the score for neonatal acute physiology and sepsis.29 Inour study, a positive correlation was found between morbidityand SNAPPE-II values, RDS and mtHcy. However, a positivecorrelation was not found between transient tachypnea, mtHcyor ctHcy. All of the newborns with transient tachypnea weredelivered through elective cesarean section. Therefore, elective

1,00,80,60,40,20,0

1 - Specificity

1,0(a)

(b)

0,8

0,6

0,4

0,2

0,0

Sen

siti

vity

ROC Curve

Test result variables Area SEM P-value %95CIPrematurity 0.700 0.040 <0.0001 0.622-0.779

Pre-eclampsia 0.601 0.043 0.015 0.518-0.685

Maternal blood total homocysteine levels

0.644 0.040 0.001 0.566-0.722

Cord blood total homocysteine levels

0.569 0.043 0.100 0.484-0.654

SNAPPE-II 0.748 0.040 <0.0001 0.670-0.826

Cord blood cardiac troponin I

0.577 0.042 0.064 0.494-0.661

Fig. 1 (a) Receiver–operator characteristic (ROC) curve for prematurity, pre-eclampsia, maternal blood total homocysteine (tHcy), cord bloodtHcy, score for neonatal acute physiology and perinatal extension II (SNAPPE-II) and cord blood cardiac troponin I (cTnI) as predictors ofnewborn morbidity at birth. Diagonal segments are produced by ties. , prematurity; , pre-eclampsia; , perinatal hypoxia; , cord pH;

, base excess; , maternal tHcy; , cord tHcy; , SNAPPE-II; , cord cTnl; , reference line. (b) Area under the ROC. SNAPPE-II hadthe highest sensitivity and specificity in predicting morbidity at birth.

6 H Usluer et al.

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

12345

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

JOBNAME: No Job Name PAGE: 7 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: E41D0278/v2503/blackwell/journals/ped_v0_i0/ped_3485

cesarean section may be a prenatal risk factor for transient tac-hypnea in newborns.

ROC curve analysis revealed that compared with mtHcy,SNAPPE-II was a more sensitive predictor of morbidity in new-borns; however, while SNAPPE-II is a useful postnatal predictor,mtHcy is a better prenatal predictor. Based on the findings of thepresent study, the optimal mtHcy cut-off value in predictingmorbidity is 4.4. Therefore, elevated mtHcy levels may be aprenatal risk factor for morbidity in newborns. SNAPPE-IIvalues, cord cTnI and ctHcy can be quickly determined in thelaboratory of any hospital, and the material needed to determinemtHcy is easily obtained from mothers before delivery. Impor-tantly, mtHcy level is a potential biomarker of neonatal outcomesbefore delivery. There are several limitations to this study, includ-ing the small sample size and the established difference in themedian gestational age between the study and control groups. Itis difficult to adjust for gestational age in hospitalized patientsbecause premature deliveries are more common in hospitalizedpatients. It is therefore likely that this study underestimated dif-ferences between hospitalized and control patients. Mortality wasnot studied because there were few non-survivors.

Conclusion

Elevated mtHcy levels were associated with cTnI, SNAPPE-II,cord blood gas and neonatal morbidities. In addition, ctHcy levelswere significantly correlated with mtHcy, cord arterial pH andcTnI levels. These results suggest that SNAPPE-II may be anearly predictor of morbidity after delivery and that an elevatedmtHcy level in pregnant women may be an early biomarker ofnewborn morbidity before delivery.

References

1 Fodinger M, Buchmayer H, Sunder-Plassmann G. Moleculargenetics of homocysteine metabolism. Miner. Electrolyte Metab.1999; 25: 269–78.

2 Hague WM. Homocysteine and pregnancy. Best Pract. Res. Clin.Obstet. Gynaecol. 2003; 17: 459–69.

3 Ingec M, Borekcı B, Kadanalı S. Elevated plasma homocysteineconcentrations in severe preeclampsia and eclampsia. Tokohu J.Exp. Med. 2005; 206: 225–31.

4 Murphy MM, Scott JM, Arija V, Molloy AM, Fernandez-BallartJD. Maternal homocysteine before conception and throughoutpregnancy predicts fetal homocysteine and birth weight. Clin.Chem. 2004; 50: 1406–12.

5 Engel SM, Olshan AF, Siega-Riz AM, Savitz DA, Chanock SJ.Polymorphisms in folate metabolizing genes and risk for sponta-neous preterm and small-for-gestational age birth. Am. J. Obstet.Gynecol. 2006; 195: 1231. e1–11.

6 Rivard-Infante C, Rivard GE, Gauthier R, Théoret Y. Unexpectedrelationship between plasma homocysteine and intrauterine growthrestriction. Clin. Chem. 2003; 49: 1476–82.

7 Ray JC, Laskin CA. Folic acid and homocysteine metabolismdefects and risk of placental abruption, preeclampsia, and sponta-neous pregnancy loss: a systematic review. Placenta 1999; 20:519–29.

8 Rosenquist T, Ratashak A, Selhub J. Homocysteine induces con-genital heart and neural tube: effect of folic acid. Proc. Natl. Acad.Sci. U. S. A. 1996; 93: 15227–32.

9 Baksu A, Taksın M, Goker N, Baksu B, Uluocak A. Plasmahomocysteine in late pregnancies complicated with preeclampsiaand in newborns. Am. J. Perinatol. 2006; 23: 31–5.

10 Hongsprabhas P, Saboohi F, Aranda JV et al. Plasma homocys-teine concentrations of preterm infants. Biol. Neonate 1999; 76:65–71.

11 Jyothi S, Misra I, Morris G, Benton A, Griffin D, Allen S. Red cellfolate and plasma homocysteine in preterm infants. Neonatology2007; 92: 264–8.

12 Sturtz WJ, Leef KH, Mackley AB, Sharma S, Bottiglieri T, PaulDA. Homocysteine levels in preterm infants: is there an associationwith intraventricular hemorrhage? A prospective cohort study.BMC Pediatr. 2007; 28: 38–43.

13 Trevisanuto D, Picco G, Golin R et al. Cardiac troponin I inasphyxiated neonates. Biol. Neonate 2006; 89: 190–3.

14 Turker G, Babaoglu K, Gokalp AS, Sarper N, Zengin E, ArısoyAE. Cord blood cardiac troponin I as an early predictor of short-term outcome in perinatal hypoxia. Biol. Neonate 2004; 86: 131–7.

15 Türker G, Sarper N, Babaoglu K, Gökalp AS, Duman C, ArisoyAE. Early prognostic significance of umbilical cord troponin I incritically ill newborns. Prospective study with a control group. J.Perinat. Med. 2005; 33: 54–9.

16 Brasselet C, Garnotel R, Lafont A et al. Prepercutaneous coronaryintervention plasma homocysteine concentration is a useful predic-tor of angioplasty-induced myocardial damage. Clin. Chem. 2005;51: 2374–7.

17 Bodí V, Sanchis J, Llàcer A et al. Risk stratification in non-STelevation acute coronary syndromes: predictive power of troponinI, C-reactive protein, fibrinogen and homocysteine. Int. J. Cardiol.2005; 98: 277–83.

18 Richardson DK, Corcoran JD, Escobar GJ, Lee SK. SNAP-II andSNAPPE-II: simplified newborn illness severity and mortality riskscores. J. Pediatr. 2001; 138: 92–100.

19 Blackwell SC, Redman ME, Whitty JE et al. The effect of intra-partum magnesium sulfate therapy on fetal cardiac troponin I levelsat delivery. J. Matern. Fetal Neonatal Med. 2002; 12: 327–31.

20 American College of Obstetrics and Gynecology. Hypertensionpregnancy. ACOG Tech. Bull. 1996; 219: 1–8.

21 Lubchenco LO, Hansman C, Boyd E. Intrauterine growth in lengthand head circumference as estimated from live births at gestationalages from 26 to 42 weeks. Pediatrics 1966; 37: 403–8.

22 Hayashi T, Honda G, Suzuki K. An atherogenic stimulus homocys-teine inhibits cofactor activity of thrombomodulin and enhancesthrombomodulin expression in human umbilical vein endothelialcells. Blood 1992; 79: 2930–36.

23 Tasatargil A, Sadan G, Karasu E. Homocysteine-induced changesin vascular reactivity of guinea-pig pulmonary arteries: role of theoxidative stress and poly (ADP-ribose) polymerase activation.Pulm. Pharmacol. Ther. 2007; 20: 265–72.

24 Mi WY, Liu W, Liu TC et al. Serum levels of homocysteine andfolate in neonates with asphyxia. Zhongguo Dang Dai Er Ke ZaZhi 2008; 10: 130–2.

25 Lipton SA, Kim WK, Choi YB et al. Neurotoxicity associated withdual actions of homocysteine at the N-methyl-D-aspartate receptor.Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 5923–8.

26 Cho SE, Hong KS, Shin GJ, Chung WS. The methylenetetrahy-drofolate reductase C677T gene mutation is associated with hyper-homocysteinemia, cardiovascular disease and plasma B-typenatriuretic peptide levels in Korea. Clin. Chem. Lab. Med. 2006;44(9): 1070–5.

27 Kahn DJ, Richardson DK, Billett HH. Association of thrombocy-topenia and delivery method with intraventricular hemorrhageamong very-low-birth-weight infants. Am. J. Obstet. Gynecol.2002; 186: 109–16.

28 Lim L, Rozycki HJ. Postnatal SNAP-II scores in neonatal intensivecare unit patients: relationship to sepsis, necrotizing enterocolitis,and death. J. Matern. Fetal Neonatal Med. 2008; 21(6): 415–9.

29 Griffin MP, Moorman JR. Toward the early diagnosis of neonatalsepsis and sepsis-like illness using novel heart rate analysis. Pedi-atrics 2001; 107: 97–104.

Homocysteine and troponin I in newborn 7

© 2011 The AuthorsPediatrics International © 2011 Japan Pediatric Society

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

293031

3233343536373839404142434445464748495051525354555657585960

616263646566676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127

JOBNAME: No Job Name PAGE: 8 SESS: 9 OUTPUT: Fri Oct 28 15:44:24 2011 SUM: 077EC202/v2503/blackwell/journals/ped_v0_i0/ped_3485

Toppan Best-set Premedia LimitedJournal Code: PED Proofreader: EmilyArticle No: 3485 Delivery date: 28 October 2011Page Extent: 7