- Email: [email protected]
Increased fetal adiposity: A very sensitive marker of abnormal in utero development
Increased fetal adiposity: A very sensitive marker of abnormal in utero development Patrick M Catalano, MD,a,b Alicia Thomas, RD,a Larraine Huston-Presley, MS,a and Saeid B. Amini, PhD, MBA, JDa Cleveland, Ohio OBJECTIVE: Because offspring of women with gestational diabetes mellitus have an increased risk of obesity and diabetes mellitus as young adults, our purpose was to characterize body composition at birth in infants of women with gestational diabetes mellitus and normal glucose tolerance. STUDY DESIGN: One hundred ninety-five infants of women with gestational diabetes mellitus and 220 infants of women with normal glucose tolerance had anthropometric measurements and total body electrical conductivity body composition evaluations at birth. Parental demographic, anthropometric, medical and family history data, and diagnostic glucose values were used to develop a stepwise regression model that related to fetal growth and body composition. RESULTS: There was no significant difference in birth weight (gestational diabetes mellitus [3398 ± 550 g] vs normal glucose tolerance [3337 ± 549 g], P = .26) or fat-free mass (gestational diabetes mellitus [2962 ± 405 g] vs normal glucose tolerance [2975 ± 408 g], P = .74) between groups. However, infants of women with gestational diabetes mellitus had significantly greater skinfold measures (P = .0001) and fat mass (gestational diabetes mellitus [436 ± 206 g] vs normal glucose tolerance [362 ± 198 g], P = .0002) compared with infants of women with normal glucose tolerance. In the gestational diabetes mellitus group, although gestational age had the strongest correlation with birth weight and fat-free mass, fasting glucose level had the strongest correlation with neonatal adiposity. CONCLUSION: Infants of women with gestational diabetes mellitus, even when they are average weight for gestational age, have increased body fat compared with infants of women with normal glucose tolerance. Maternal fasting glucose level was the strongest predictor of fat mass in infants of women with gestational diabetes mellitus. This increase in body fat may be a significant risk factor for obesity in early childhood and possibly in later life. (Am J Obstet Gynecol 2003;189:1698-704.)
Key words: Pregnancy, gestational diabetes mellitus, body composition, fetal growth
Fetal overgrowth is common during gestation in women with glucose intolerance. The Pedersen1 hypothesis states that fetal overgrowth or macrosomia is a consequence of increased maternal glucose, which stimulates fetal insulin production and possibly other growth factors. Increased fetal beta-cell mass in infants of diabetic mothers has been identified as early as the second trimester of pregnancy,2 and increased maternal amino acid and free fatty acids have been related to fetal overgrowth. Macrosomia complicates as many as 50% of pregnancies with gestational From the Department of Reproductive Biologya and the Schwartz Center for Nutrition and Metabolism,b Case Western Reserve University at MetroHealth Medical Center. Supported by National Institutes of Health grants No. HD-22965, PERC-11089, General Clinical Research Center MO1-RR-80, and the Weight Watchers Foundation. Received for publication January 8, 2003; revised March 24, 2003; accepted June 20, 2003. Reprint requests: Patrick M. Catalano, MD, Dept. of Reproductive Biology, Case Western Reserve University, MetroHealth Medical Center, Cleveland, OH 44109. E-mail: [email protected] Ó 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1016/S0002-9378(03)00828-7
1698
diabetes mellitus (GDM), which includes women with optimal glycemic control. There is further evidence for obesity and diabetes mellitus in children of women who were glucose intolerant during pregnancy compared with their siblings when these women had normal glucose tolerance (NGT).3 The classification of macrosomia is controversial. It has been defined as a birth weight of >4000 g, regardless of gestational age. Macrosomia has also been described as birth weight being >90th percentile for gestational age or large for gestational age. Factors such as ethnicity, altitude above sea level, sex, and parity must be considered in the assessment of fetal growth. Hence, the classification of macrosomia remains problematic. Our purpose was to conduct a prospective evaluation of neonatal body composition and anthropometric measurements in infants of women with NGT and GDM. GDM is defined as carbohydrate intolerance of varying degrees of severity, with first recognition during pregnancy.4 We elected to assess the body composition of infants on the basis of a review by Sparks,5 whereby he hypothesized that genetic factors have a stronger relationship with fat-free
Catalano et al 1699
Volume 189, Number 6 Am J Obstet Gynecol
mass (FFM), whereas in utero environment may correlate better with fetal fat mass. There has been a 33% increase in the incidence of type 2 diabetes mellitus over the last decade in the United States, particularly in young adults who are obese.6 Because the offspring of women with GDM have an increased risk of obesity and type 2 diabetes,3,7 we hypothesized that, consistent with the fetal origin of adult disease theory, infants of women with GDM have increased fat mass and not FFM compared with infants of women with NGT. Material and methods The protocol was approved by the hospital Institutional Review Board. Written informed consent was obtained from each subject before evaluation in the General Clinical Research Center. Subjects were recruited on the basis of availability to participate in this prospective protocol that evaluated fetal growth. Women with NGT were recruited from the general population, and women with GDM were recruited from our Pregnancy Diabetes Clinic from 1990 through 2000. All women in these clinics were eligible to participate and were recruited before delivery. The obstetric record was reviewed, and each mother was interviewed. Exclusion criteria included birth weight of < 2000 g, multifetal gestations, and neonates with anomalies. All subjects had a screening 1-hour 50-g glucose challenge test at 26 to 28 weeks of gestation. A glucose level of $135 mg/dL was considered positive. All subjects with positive screening results had a 3-hour 100-g oral glucose tolerance test. The subjects with NGT had either a 1-hour 50-g glucose challenge of < 135 mg/dL or, if positive, a normal 100-g 3-hour oral glucose tolerance test. The diagnosis of GDM was made according to the National Diabetes Data Group criteria.8 All women with GDM had dietary instruction by a registered dietician. The diet consisted of 50% complex carbohydrates, 30% fat, and 20% protein. Caloric intake was calculated to ensure weight gain according to the American Institute of Medicine criteria.9 All subjects with GDM performed home glucose monitoring. Target fasting and preprandial glucose values were < 100 mg/dL and 2-hour postprandial values were < 120 mg/dL. All women with GDM were encouraged to increase physical activity by walking 30 minutes each day after meals. Insulin was instituted when subjects had persistent glucose concentrations greater than target levels. Eighty percent of the anthropometric measurements and total body electrical conductivity (TOBEC) estimates of body composition were completed within 24 hours of birth; 93% of the measurements were completed within 48 hours, and the remaining 7% of the measurements were completed within 72 hours. Weight was obtained with a calibrated scale (Scale-Tronix, Wheaton, Ill), a measuring board for length, Harpenden calipers (British
Indicators, Sussex, UK) for skinfold measurements, a tape measure for circumference, and an anthropometer for limb lengths. Normative birth weight for gestational age was based on published standards from this institution.10 Details of the methods for skinfold measurements, bone lengths, and circumferences of trunk and extremities have been described previously.11 TOBEC estimates of body composition were obtained on the pediatric model HP-2 (EM-SCAN, Inc, Springfield, Ill) at birth. A description of the TOBEC instrument, theory, and validation studies are in a review by Fiorotto et al.12 Briefly, the instrument consists of a cylindric measurement chamber that contains a solenoid coil that creates an electromagnetic field that is generated by an oscillating electric current. Any conductive object placed within the coil will disturb the electromagnetic field and result in a dissipation of a quantity of the field’s energy. TOBEC estimates of neonatal body composition are based on the premise that the FFM contains electrolytes and represents a conductive component of the body, in contrast to fat mass, which is relatively anhydrous. Therefore, the FFM disturbs the electromagnetic field in a precise manner. To perform TOBEC measurements, the infant was swaddled and placed on the carriage of the instrument. Each infant was measured when the carriage was pushed into the cylinder a set distance. A computergenerated estimate of FFM was calculated from the disturbance of the electromagnetic field. Each infant was measured 10 times, and the mean estimate of FFM was used as ‘‘estimated FFM.’’ Fat mass was estimated by the subtraction of the FFM from the body weight. Anthropometric and TOBEC measurements were performed by one of three examiners. The coefficient of variation among the examiners was 3% for the anthropometrics and 7% for the skinfolds. The coefficient of variation for repetitive TOBEC measures is < 2%. Statistical analysis were performed with the use of twosample t tests, the Mann-Whitney U test, v2 analysis, analysis of variance, analysis of covariance that was adjusted for confounding covariables, and stepwise multiple regression analysis on SAS (Statistix Analysis System, Cary, NC). In the stepwise multiple regression analysis, smoking was coded with 0 (nonsmoker) and 1 (smoker), the rest of the variables were continuous. Data are expressed as mean ± SD, and a probability value of < .05 was considered significant. Results One hundred ninety-five women with GDM and 220 women with NGT were recruited. Demographics are shown in Table I. Of the 195 women with GDM, 67 women (34%) required insulin to maintain fasting and preprandial and postprandial glucose values within the target range. One hundred twenty-eight women with GDM (66%) were able to maintain glucose values within
1700 Catalano et al
December 2003 Am J Obstet Gynecol
Table I. Parental demographics Demographic Maternal Age (y) Height (cm) Pregravid weight (kg) Weight at last antenatal visit (kg) Weight gain (kg) Paternal Height (cm) Weight (kg) 1 h 50 g glucose screen (mg/dL) Maternal smoking status (%) Yes No Family history of diabetes mellitus (%) Yes No Obstetric problems(%)* Yes No Medical problems(%)y Yes No Maternal race (%) White African American Hispanic/other Parity 0 $1 Education level years (%) 0-12 13-16 >16 Route of delivery (%) Cesarean Vaginal
GDM (n = 195)
NGT (n = 220)
P value
29.0 162.4 81.0 94.9 13.9
± 6.1 ± 6.3 ± 23.2 ± 22.8 ± 9.4
28.0 164.5 65.5 80.5 14.7
± 5.8 ± 7.9 ± 15.3 ± 15.3 ± 5.9
.12 .003 .0001 .0001 .29
177.8 87.8 170
± 8.6 ± 18.0 ± 32
178.9 82.8 112
± 8.2 ± 14.2 ± 28
.18 .003 .0001
29 71
15 85
.0004
74 26
47 53
.0001
28 72
20 80
.08
20 80
16 84
.22
58 27 15
75 17 8
121 74
137 83
63 32 5
40 39 21
.0001
37 63
25 75
.01
.002 .96
Data are presented as mean ± SD. *History of first trimester spotting, endometriosis, history of neonatal death, uterine anomaly, herpes virus, chlamydia, infertility, abnormal Papanicolaou smears, and ovarian cysts. yHistory of bronchitis, asthma, colitis, seizures, hypertension, obesity, migraine headaches, anemia, and mitral valve prolapse.
the normal range using diet and exercise. Women with GDM were older, shorter, and smoked more frequently and had greater pregravid weight compared with the NGT group. Paternal weight was greater in the partners of the GDM compared with the NGT group. Thirty-seven percent of the subjects with GDM and 25% of the subjects with NGT had cesarean deliveries. Neonatal anthropometrics are shown on Table II. The average gestational age at delivery was earlier in infants of women with GDM by 4 days. There was no significant difference in length or birth weight. The infants of the women with GDM had significantly greater ponderal indexes (weight/length3 3 100) and skinfold measures at all five sites compared with the NGT group. Although there was no significant difference in the proportion of infants with weight >4000 g between groups, there were significantly more infants whose weights were >90th percentile for gestational age in the GDM group compared with the NGT group (27% vs 14%, P = .004).
The TOBEC estimates of body composition are shown on Table III. The infants of women with GDM had significantly greater body fat and percentage of body fat but no significant difference in average FFM in comparison with the NGT group. Additionally, the distribution of fat was not significant in the circumferences of the abdominal/thigh and chest/thigh ratios between the groups. Furthermore, there were no significant differences in the average skinfold abdominal + flank + subscapular/triceps + thigh ratio or abdominal + flank + subscapular/thigh ratio (data not shown). After adjustment for gestational age, maternal pregravid weight, weight at last antenatal visit, race, smoking status, and maternal and paternal height, there was no significant difference in birth weight (P = .15) or FFM (P = .60); however, a significant difference persisted in the percentage of body fat (P = .0005) between GDM and NGT. After an adjustment for the aforementioned factors plus a family history of diabetes mellitus and maternal weight
Catalano et al 1701
Volume 189, Number 6 Am J Obstet Gynecol
Table II. Neonatal anthropometrics Measure
GDM (n = 195)
Gestational age at delivery (wk) Gestiational age dating criteria (%) Ultrasound < 20 wk Ultrasound >20 wk and Dubowitz dates Dubowitz dates Neonatal Sex (%) Male Female Weight (g) Length (cm) Extremity length (cm) Upper arm Lower arm Upper leg Lower leg Circumferences (cm) Head Chest Umbilical Upper arm Lower arm Upper leg Lower leg Ponderal index (weight/length3) 3 100 Skinfolds (mm) Triceps Subscapular Flank Thigh Abdominal wall Weight for gestational age (%) < 10% 10%-90% >90% Weight (%) >4000 g < 4000 g Data are presented as mean
38.4
NGT (n = 220)
± 1.3
P value
± 1.4
.0001
73 22 5
71 20 9
.57
51 49
54 46
39.0
.57
3398 49.9
± 550 ± 2.1
3337 50.0
± 549 ± 2.2
.26 .68
8.2 7.3 9.7 8.3
± 0.6 ± 0.6 ± 0.9 ± 0.7
8.4 7.4 9.9 8.5
± 0.6 ± 0.5 ± 0.9 ± 0.7
.002 .03 .09 .02
34.7 31.8 31.3 10.6 9.7 15.2 11.1 2.72
± 1.4 ± 2.0 ± 2.3 ± 1.3 ± 1.0 ± 1.6 ± 1.1 ± 0.28
34.5 31.8 31.0 10.4 9.6 15.0 11.0 2.66
± 1.5 ± 2.1 ± 2.2 ± 1.0 ± 0.9 ± 1.6 ± 1.0 ± 0.26
.18 .88 .16 .18 .16 .16 .25 .02
4.7 5.4 4.2 6.0 3.5
± 1.1 ± 1.4 ± 1.2 ± 1.4 ± 0.9
4.2 4.6 3.8 5.4 3.0
± 1.0 ± 1.2 ± 1.0 ± 1.5 ± 0.8
.0001 .0001 .0001 .0001 .0001
5 68 27
6 79 14
.007
12 88
9 91
.37
± SD.
gain, there still remained a significant difference (P = .04) in fat mass between groups. Because more of the infants of women with GDM were macrosomic (ie, >90th percentile; 27% vs 14%) compared with the neonates of women with NGT , we next evaluated only neonates whose weights were average for gestational age (ie, between the 10th and 90th percentiles). There were 132 infants of women with GDM and 175 infants of women with NGT for analysis. The subjects with GDM were significantly shorter and had greater pregravid weight and weight at last antenatal visit as compared with the NGT group. The subjects with GDM also had a greater incidence of positive smoking history, family history of diabetes mellitus, and less post-high school education compared with the NGT group. The infants of the women with GDM were delivered more often by cesarean delivery (36% vs 25%, P = .03) compared with the NGT group. The average gestational age at delivery was again significantly earlier in infants of women with GDM (38.5 ± 1.4 weeks of gestation vs 39.0 ± 1.4 weeks of gestation,
Table III. Neonatal body composition GDM (n = 195) FFM (g) Fat mass (g) Body fat (%)
2962 436 12.4
± 405 ± 206 ± 4.6
Data are presented as mean
NGT (n = 220) 2975 362 10.4
± 408 ± 198 ± 4.6
P value .74 .0002 .0001
± SD.
P = .002), but there was no significant difference in the mean birth weight (3202 ± 357 g vs 3249 ± 372 g, P = .27) between groups. Infants of women with GDM had significantly greater skinfold measures at all five sites in comparison with the infants of women with NGT (P = .007 .0001). TOBEC estimates of body composition showed that the infants of the women with GDM had significantly less FFM (2832 ± 286 g vs 2919 ± 287 g, P = .008) but had significantly greater fat mass (371 ± 163 g vs 329 ± 150 g, P = .02) percent body fat (11.4% ± 4.6% vs 9.9% ± 4.0%, P = .002) in comparison
1702 Catalano et al
December 2003 Am J Obstet Gynecol
Table IV. Multiple stepwise regression analysis for factors that affected fetal growth and body composition in 220 women with NGT Factor Weight Gestational age Maternal height Maternal weight gain Glucose screen Smoking status (–) FFM Gestational age Maternal height Maternal weight gain Fat mass Gestational age Weight at first antenatal visit Maternal weight gain Glucose screen Body fat (%) Gestational age Glucose screen Maternal weight gain Pregravid weight
r2 0.22 0.25 0.28 0.30 0.31 0.20 0.23 0.24
Dr2
0.03 0.03 0.02 0.01
Table V. Multiple stepwise regression analysis for factors that affected fetal growth and body composition in 195 women with GDM
P value
.007 .01 .008 .05
0.03 0.01
.0001 .002 .05
0.16 0.21 0.24 0.26
0.05 0.03 0.02
.0001 .004 .006 .01
0.14 0.18 0.20 0.23
0.04 0.02 0.03
.0001 .01 .001 .01
with the infants of women with NGT. There was no significant difference in the distribution of body fat between groups. After an adjustment was made for maternal height and pregravid weight, gestational age, smoking status, race, and paternal weight, there was no significant difference in birth weight (P = .42) or FFM (P = .59). Adjustment for the aforementioned factors plus weight at the last antenatal visit and a family history of diabetes mellitus in fat mass (P = .02) and percent body fat (P = .02) continued to be significant between groups (detailed data available on request). We next compared the body composition estimates between the infants of the women with GDM who were treated with diet alone (group A1, n = 128 infants) and infants who were treated with diet plus insulin (group A2, n = 67 infants) on the average. The A2 GDM group had greater pregravid weight (87.9 ± 25.5 kg vs 77.4 ± 21.1 kg, P = .0002). The A2 GDM group also had greater parity (49% > 1.0 vs 32% > 1.0, P = .02) and cesarean delivery (52% vs 29%, P = .001) compared with the A1 GDM group. The A2 GDM group was delivered earlier (37.9 ± 1.1 weeks of gestation vs 38.7 ± 1.4 weeks of gestation, P = .0001) than the A1 GDM group. There was no significant difference in birth weight (3497 ± 556 g vs 3346 ± 542 g, P = .07) between groups. The skinfold measures were significantly greater in the A2 GDM group: subscapular (5.9 ± 1.6 mm vs 5.1 ± 1.2 mm, P = .0001), triceps (5.0 ± 1.1 mm vs 4.5 ± 1.0 mm, P = .003), flank (4.5 ± 1.2 mm vs 4.0 ± 1.1 mm, P = .01), thigh (6.3 ± 1.5 mm vs 5.8 ± 1.3 mm, P = .02), and abdomen (4.0 ± 1.4 mm vs 3.3 ± 0.9 mm, P = .0001) compared with the NGT group. There were no significant differences in the
Factor Weight Gestational age Fasting glucose Smoking status (–) FFM Gestational age Smoking status (–) Fasting glucose Maternal height Fat mass Fasting glucose Gestational age Body fat (%) Fasting glucose Gestational age
r2
Dr2
P value
0.07 0.17 0.21
0.10 0.04
.0003 .0001 .003
0.07 0.14 0.19 0.22
0.07 0.05 0.03
.0003 .0006 .0005 .02
0.11 0.17
0.06
.0001 .0007
0.09 0.13
0.04
.0001 .007
TOBEC estimates of FFM (3005 ± 397 g vs 2939 ± 409 g, P = .28), but there were significantly greater fat mass (492 ± 215 g vs 407 ± 196 g, P = .006) and percent body fat (13.6% ± 4.6% vs 11.7% ± 4.5%, P = .007) in the A2 GDM group compared with the A1 GDM group. After the adjustment for gestational age, maternal pregravid weight, parity, and smoking status, there were significant differences in birth weight (P = .0006) and FFM (P = .03). Because smoking affects FFM and not fat mass, when we adjusted for gestational age, maternal pregravid weight, and parity, the A2 GDM group persisted in having greater fat mass (P = .002) and percent body fat (P = .003) as compared with the A1 GDM group. We performed a multiple stepwise regression analysis to assess the factors and relative strength of each variable in regard to birth weight and body composition in the NGT group (Table IV) and GDM group (Table V). The purpose of this analysis was to identify factors that are available in clinical practice and potentially amenable to intervention to decrease the risk of adiposity. The independent variables for the NGT group included maternal age, height, pregravid weight, weight at last visit, weight gain, parity, paternal height and weight, gestational age, neonatal gender, presence of family history of diabetes mellitus, race, smoking status, and glucose screen. In the GDM group, we included all the variables for the NGT group with the exception of the glucose screen, but included the fasting, 1-, 2-, and 3-hour glucose values from the diagnostic oral glucose tolerance test. In the NGT group, gestational age had the strongest correlation with birth weight and body composition. Maternal height was a significant variable with respect to birth weight and FFM, whereas maternal pregravid weight, weight gain, and glucose screen were included in the regression models of fat mass and percent body fat. In the GDM group, gestational age had the strongest correlation with birth
Volume 189, Number 6 Am J Obstet Gynecol
weight and FFM, although fasting glucose had the strongest correlation with fat mass and percent body fat. Comment The most significant finding of our study was that of a significant increase in adiposity of average for gestational aged infants of women with GDM. Previous studies by Pettitt et al3 and Silverman et al7 have shown that infants of women with GDM have an increased risk of adolescent obesity and glucose intolerance. However, detailed body composition data at the time of birth was not available. Earlier studies ascribed the increases in birth weight of infants of diabetic women to either fat mass and/or FFM. The carcass analysis studies of Fee and Weil13 in stillbirths and cases of neonatal death described normal body water in FFM but significant increases in body fat in comparison with control subjects. In 40 infants of women with GDM, Whitelaw14 reported that neonatal fat mass, as estimated from skinfolds, was significantly greater in infants of women with diabetes mellitus compared with a nondiabetic reference population. Additionally, Enzi et al15 evaluated body composition using skinfolds in 17 infants of women with GDM and 17 nondiabetic control subjects. The mean ± SE percent body fat in the infants of the GDM was 17% ± 1.7% compared with 12.2% ± 0.5% in the control subjects. However, body weight was also greater in the GDM group (3508 ± 200 g) compared with the control group (3103 ± 100 g). In contrast, Naeye16 reported that there was an increase in FFM, in contrast with fat mass, in overgrown stillborn fetuses or neonatal deaths for women with GDM. These infants also had evidence of both hyperplasia and hypertrophy in cellular tissue of various organs. All these studies were conducted on a small number of infants, and the cause of the fetal/ neonatal death was often unknown. Additional body composition studies were conducted recently by Nasiat et al.17 Although the investigators reported an increase in subscapular fat in the infants of the women with diabetes mellitus (both pregestational and gestational), the birth weight of the infants of the women with diabetes mellitus was significantly greater than the birth weight of the nondiabetic control subjects. Furthermore, fat mass was estimated as the sum of four skinfolds. In a study by McFarland et al,18 although the authors found greater fat mass using an anthropometric formula in infants of women with diabetes mellitus, the number of study subjects was small and limited to macrosomic infants. The results of our study extend these results on the basis of method and number and analysis of study subjects. Thus, the current goal of achieving appropriate weight for gestational age may not be adequate given these results. The summary of the Fourth International WorkshopConference on Gestational Diabetes Mellitus19 noted that, although treatment of hyperglycemia can decrease
Catalano et al 1703
the risk of macrosomia markedly, optimal treatment has yet to be defined. Furthermore, maternal glucose concentrations, less than those used in the diagnosis of GDM, have also been related to increased fetal growth.20 Various treatment strategies, which include intensive treatment protocols to normalize maternal glucose concentration, are used currently in the treatment of GDM. However, such protocols may result in as many as 66% of women requiring insulin, and the overly aggressive treatment of maternal glycemia may actually increase the risk of small for gestational age neonates.21,22 Thirty-four percent of the patients with GDM (n = 67) were treated with insulin (A2 GDM group) in addition to diet and exercise. Thus, even in GDM patients who were treated with insulin, the neonates were significantly heavier and had increased adiposity in comparison with the infants of the A1 GDM group. We speculate that the A2 GDM group had greater pregravid insulin resistance as compared with the A1 GDM group. This is based on the A2 GDM group having significantly greater pregravid weight and parity compared with the A1 GDM group. These two factors have been associated with increased maternal insulin resistance and neonatal adiposity.23 Hence, the criteria for the diagnosis of GDM earlier in gestation may be warranted. The issue of whether the optimization of the control of fasting as compared with either preprandial or postprandial glucose concentrations is more important for the control of fetal overgrowth continues to be debated, as does the importance of glucose control in early, compared with late, gestation.24 Of interest, in our study, fasting glucose levels at the time of the oral glucose tolerance test was the factor with the strongest correlation with fetal fat accretion. These data support a study by UvenaCelebrezze et al25 that reported that fasting glucose had the strongest correlation with fetal fat mass in comparison with premeal or postmeal glucose values, based on glucose self-monitoring seven times per day. In summary, infants of women with mild glucose intolerance (ie, GDM) have increased body fat compared with infants of women with NGT; this is independent of birth weight and may be regarded as very early onset obesity. We propose that the use of specific outcome measurements such as fetal adiposity may be a more sensitive and specific outcome measure at birth. The increase in body fat at birth may be the harbinger of an increased risk of obesity and other metabolic complications in adolescence for these infants. Although the difference in fat mass between the two groups was only 74 g (Table III), this represents a 20% difference between the NGT and GDM infants. Should this 20% difference in fat mass persist into adolescence, it will become a significant problem. Therefore, we are in the process of following this cohort to evaluate long-term outcome that is relative to adolescent obesity, glucose tolerance, and the insulin resistance syndrome.
1704 Catalano et al
REFERENCES 1. Pedersen J. The pregnant diabetic and her newborn: problems and management. Baltimore: Williams & Wilkens; 1967. 2. Reiher H, Fahrmann K, Noack S, Woltanski KP, Jutzi E, Hahn UM, et al. Age dependent insulin secretion of the endocrine pancreas in vitro from fetuses of diabetic and nondiabetic patients. Diabetes Care 1983;5:446-51. 3. Pettitt DJ, Baird HR, Aleck KA, Bennett PH, Knowler WC. Excessive obesity in offspring of Pima Indian women with diabetes during pregnancy. N Engl J Med 1983;308:242-5. 4. Metzger BE. The organizing committee: summary and recommendations of the third international workshop: conference on gestational diabetes mellitus. Diabetes 1991;40(Suppl):197-201. 5. Sparks J. Human intrauterine growth and nutrient accretion. Semin Perinatol 1984;18:74-93. 6. Mokdad AH, Engelgau MN, Ford ES, Vinicor F, Bowman BA, Marks JS, et al. Diabetes trends in the US: 1990-1998. Diabetes Care 2000; 23:1278-83. 7. Silverman BL, Rizzo TA, Cho NH, Metzger BE. Long-term effects of the intrauterine environment: the Northwestern University Diabetes in Pregnancy Center. Diabetes Care 1998;21(Suppl): B142-9. 8. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:1039-57. 9. Institute of Medicine, National Academy of Sciences, Food and Nutrition Board. Nutrition during pregnancy, 1: weight gain. Washington (DC): National Academy Press; 1990. 10. Amini SB, Catalano PM, Hirsch V, Mann LI. An analysis of birth weight by gestational age: using a computerized perinatal data base. Obstet Gynecol 1993;83:342-52. 11. Catalano PM, Thomas AJ, Avallone DA, Amini SB. Anthropometric estimation of neonatal body composition. Am J Obstet Gynecol 1995;173:1176-81. 12. Fiorotto ML, Cochran WJ, Runk RC, Sheng J-P, Klish WJ. Total body electrical conductivity measurements: effects of body composition and geometry. Am J Physiol 1987;252:R798-800. 13. Fee BA, Weil WB. Body composition of infants of diabetic mothers by direct analysis. Ann N Y Acad Sci 1963;110:869-97.
December 2003 Am J Obstet Gynecol
14. Whitelaw A. Subcutaneous fat in newborn infants of diabetic mothers: an indication of quality of diabetic control. Lancet 1977; 1:15-8. 15. Enzi G, Inelman EM, Caretta F, Villani F, Zanardo V, DeBiasi F. Development of adipose tissue in newborns of gestational diabetes and insulin diabetic mothers. Diabetes 1980;29:100-4. 16. Naeye RL. Infants of diabetic mothers: a quantitative morphologic study. Pediatrics 1965;35:980-8. 17. Nasiat H, Abalkhail B, Fageeh W, Shabat A, Zahrawy FE. Anthropometric measurements of newborns of gestational diabetic mothers: does it indicate disproportionate fetal growth? J Matern Fetal Med 1997;6:291-5. 18. McFarland MB, Trylovich CG, Langer O. Anthropometric differences in macrosomic infant of diabetic and non-diabetic mothers. J Matern Fetal Med 1998;6:292-5. 19. Metzger BE, Coustan D. The organizing committee summary and recommendations of the fourth international workshop: conference on gestational diabetes mellitus. Diabetes Care 1998;21(Suppl):B161-7. 20. Tallarigo L, Giampietro O, Penno G, Milcoli R, Gregore G, Navalesi R. Relation of glucose tolerance to complications of pregnancy in nondiabetic women. N Engl J Med 1986;315:989-92. 21. Langer O, Rodriguez DA, Xenaris EM, McFarland MB, Berkus MD, Arrendondo F. Intensified versus conventional management of gestational diabetes. Am J Obstet Gynecol 1994;170:1036-47. 22. Langer O, Levy J, Brustman L, Anyaegbunam A, Merkatz R, Divon M. Glycemic control in gestational diabetes mellitus: how tight is tight enough? Small for gestational age versus large for gestational age. Am J Obstet Gynecol 1989;161:645-53. 23. Catalano PM, Drago NM, Amini SB. Maternal carbohydrate metabolism and its relationship to fetal growth and body composition. Am J Obstet Gynecol 1995;172:1464-70. 24. Jovanovic-Peterson L, Peterson CM, Reed GF, Metzger BE, Mills JL, Knopp RH, et al. Maternal postprandial glucose levels and infant birth weight: the Diabetes in Early Pregnancy Study. Am J Obstet Gynecol 1991;164:103-11. 25. Uvena-Celebrezze J, Fung C, Thomas AJ, Huston-Presley L, Amini SB, Catalano PM. Relationship of neonatal body composition to maternal glucose control in women with gestational diabetes mellitus. J Matern Fetal Neonat Med 2002;12:396-401.
Key words: Pregnancy, gestational diabetes mellitus, body composition, fetal growth
Fetal overgrowth is common during gestation in women with glucose intolerance. The Pedersen1 hypothesis states that fetal overgrowth or macrosomia is a consequence of increased maternal glucose, which stimulates fetal insulin production and possibly other growth factors. Increased fetal beta-cell mass in infants of diabetic mothers has been identified as early as the second trimester of pregnancy,2 and increased maternal amino acid and free fatty acids have been related to fetal overgrowth. Macrosomia complicates as many as 50% of pregnancies with gestational From the Department of Reproductive Biologya and the Schwartz Center for Nutrition and Metabolism,b Case Western Reserve University at MetroHealth Medical Center. Supported by National Institutes of Health grants No. HD-22965, PERC-11089, General Clinical Research Center MO1-RR-80, and the Weight Watchers Foundation. Received for publication January 8, 2003; revised March 24, 2003; accepted June 20, 2003. Reprint requests: Patrick M. Catalano, MD, Dept. of Reproductive Biology, Case Western Reserve University, MetroHealth Medical Center, Cleveland, OH 44109. E-mail: [email protected] Ó 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1016/S0002-9378(03)00828-7
1698
diabetes mellitus (GDM), which includes women with optimal glycemic control. There is further evidence for obesity and diabetes mellitus in children of women who were glucose intolerant during pregnancy compared with their siblings when these women had normal glucose tolerance (NGT).3 The classification of macrosomia is controversial. It has been defined as a birth weight of >4000 g, regardless of gestational age. Macrosomia has also been described as birth weight being >90th percentile for gestational age or large for gestational age. Factors such as ethnicity, altitude above sea level, sex, and parity must be considered in the assessment of fetal growth. Hence, the classification of macrosomia remains problematic. Our purpose was to conduct a prospective evaluation of neonatal body composition and anthropometric measurements in infants of women with NGT and GDM. GDM is defined as carbohydrate intolerance of varying degrees of severity, with first recognition during pregnancy.4 We elected to assess the body composition of infants on the basis of a review by Sparks,5 whereby he hypothesized that genetic factors have a stronger relationship with fat-free
Catalano et al 1699
Volume 189, Number 6 Am J Obstet Gynecol
mass (FFM), whereas in utero environment may correlate better with fetal fat mass. There has been a 33% increase in the incidence of type 2 diabetes mellitus over the last decade in the United States, particularly in young adults who are obese.6 Because the offspring of women with GDM have an increased risk of obesity and type 2 diabetes,3,7 we hypothesized that, consistent with the fetal origin of adult disease theory, infants of women with GDM have increased fat mass and not FFM compared with infants of women with NGT. Material and methods The protocol was approved by the hospital Institutional Review Board. Written informed consent was obtained from each subject before evaluation in the General Clinical Research Center. Subjects were recruited on the basis of availability to participate in this prospective protocol that evaluated fetal growth. Women with NGT were recruited from the general population, and women with GDM were recruited from our Pregnancy Diabetes Clinic from 1990 through 2000. All women in these clinics were eligible to participate and were recruited before delivery. The obstetric record was reviewed, and each mother was interviewed. Exclusion criteria included birth weight of < 2000 g, multifetal gestations, and neonates with anomalies. All subjects had a screening 1-hour 50-g glucose challenge test at 26 to 28 weeks of gestation. A glucose level of $135 mg/dL was considered positive. All subjects with positive screening results had a 3-hour 100-g oral glucose tolerance test. The subjects with NGT had either a 1-hour 50-g glucose challenge of < 135 mg/dL or, if positive, a normal 100-g 3-hour oral glucose tolerance test. The diagnosis of GDM was made according to the National Diabetes Data Group criteria.8 All women with GDM had dietary instruction by a registered dietician. The diet consisted of 50% complex carbohydrates, 30% fat, and 20% protein. Caloric intake was calculated to ensure weight gain according to the American Institute of Medicine criteria.9 All subjects with GDM performed home glucose monitoring. Target fasting and preprandial glucose values were < 100 mg/dL and 2-hour postprandial values were < 120 mg/dL. All women with GDM were encouraged to increase physical activity by walking 30 minutes each day after meals. Insulin was instituted when subjects had persistent glucose concentrations greater than target levels. Eighty percent of the anthropometric measurements and total body electrical conductivity (TOBEC) estimates of body composition were completed within 24 hours of birth; 93% of the measurements were completed within 48 hours, and the remaining 7% of the measurements were completed within 72 hours. Weight was obtained with a calibrated scale (Scale-Tronix, Wheaton, Ill), a measuring board for length, Harpenden calipers (British
Indicators, Sussex, UK) for skinfold measurements, a tape measure for circumference, and an anthropometer for limb lengths. Normative birth weight for gestational age was based on published standards from this institution.10 Details of the methods for skinfold measurements, bone lengths, and circumferences of trunk and extremities have been described previously.11 TOBEC estimates of body composition were obtained on the pediatric model HP-2 (EM-SCAN, Inc, Springfield, Ill) at birth. A description of the TOBEC instrument, theory, and validation studies are in a review by Fiorotto et al.12 Briefly, the instrument consists of a cylindric measurement chamber that contains a solenoid coil that creates an electromagnetic field that is generated by an oscillating electric current. Any conductive object placed within the coil will disturb the electromagnetic field and result in a dissipation of a quantity of the field’s energy. TOBEC estimates of neonatal body composition are based on the premise that the FFM contains electrolytes and represents a conductive component of the body, in contrast to fat mass, which is relatively anhydrous. Therefore, the FFM disturbs the electromagnetic field in a precise manner. To perform TOBEC measurements, the infant was swaddled and placed on the carriage of the instrument. Each infant was measured when the carriage was pushed into the cylinder a set distance. A computergenerated estimate of FFM was calculated from the disturbance of the electromagnetic field. Each infant was measured 10 times, and the mean estimate of FFM was used as ‘‘estimated FFM.’’ Fat mass was estimated by the subtraction of the FFM from the body weight. Anthropometric and TOBEC measurements were performed by one of three examiners. The coefficient of variation among the examiners was 3% for the anthropometrics and 7% for the skinfolds. The coefficient of variation for repetitive TOBEC measures is < 2%. Statistical analysis were performed with the use of twosample t tests, the Mann-Whitney U test, v2 analysis, analysis of variance, analysis of covariance that was adjusted for confounding covariables, and stepwise multiple regression analysis on SAS (Statistix Analysis System, Cary, NC). In the stepwise multiple regression analysis, smoking was coded with 0 (nonsmoker) and 1 (smoker), the rest of the variables were continuous. Data are expressed as mean ± SD, and a probability value of < .05 was considered significant. Results One hundred ninety-five women with GDM and 220 women with NGT were recruited. Demographics are shown in Table I. Of the 195 women with GDM, 67 women (34%) required insulin to maintain fasting and preprandial and postprandial glucose values within the target range. One hundred twenty-eight women with GDM (66%) were able to maintain glucose values within
1700 Catalano et al
December 2003 Am J Obstet Gynecol
Table I. Parental demographics Demographic Maternal Age (y) Height (cm) Pregravid weight (kg) Weight at last antenatal visit (kg) Weight gain (kg) Paternal Height (cm) Weight (kg) 1 h 50 g glucose screen (mg/dL) Maternal smoking status (%) Yes No Family history of diabetes mellitus (%) Yes No Obstetric problems(%)* Yes No Medical problems(%)y Yes No Maternal race (%) White African American Hispanic/other Parity 0 $1 Education level years (%) 0-12 13-16 >16 Route of delivery (%) Cesarean Vaginal
GDM (n = 195)
NGT (n = 220)
P value
29.0 162.4 81.0 94.9 13.9
± 6.1 ± 6.3 ± 23.2 ± 22.8 ± 9.4
28.0 164.5 65.5 80.5 14.7
± 5.8 ± 7.9 ± 15.3 ± 15.3 ± 5.9
.12 .003 .0001 .0001 .29
177.8 87.8 170
± 8.6 ± 18.0 ± 32
178.9 82.8 112
± 8.2 ± 14.2 ± 28
.18 .003 .0001
29 71
15 85
.0004
74 26
47 53
.0001
28 72
20 80
.08
20 80
16 84
.22
58 27 15
75 17 8
121 74
137 83
63 32 5
40 39 21
.0001
37 63
25 75
.01
.002 .96
Data are presented as mean ± SD. *History of first trimester spotting, endometriosis, history of neonatal death, uterine anomaly, herpes virus, chlamydia, infertility, abnormal Papanicolaou smears, and ovarian cysts. yHistory of bronchitis, asthma, colitis, seizures, hypertension, obesity, migraine headaches, anemia, and mitral valve prolapse.
the normal range using diet and exercise. Women with GDM were older, shorter, and smoked more frequently and had greater pregravid weight compared with the NGT group. Paternal weight was greater in the partners of the GDM compared with the NGT group. Thirty-seven percent of the subjects with GDM and 25% of the subjects with NGT had cesarean deliveries. Neonatal anthropometrics are shown on Table II. The average gestational age at delivery was earlier in infants of women with GDM by 4 days. There was no significant difference in length or birth weight. The infants of the women with GDM had significantly greater ponderal indexes (weight/length3 3 100) and skinfold measures at all five sites compared with the NGT group. Although there was no significant difference in the proportion of infants with weight >4000 g between groups, there were significantly more infants whose weights were >90th percentile for gestational age in the GDM group compared with the NGT group (27% vs 14%, P = .004).
The TOBEC estimates of body composition are shown on Table III. The infants of women with GDM had significantly greater body fat and percentage of body fat but no significant difference in average FFM in comparison with the NGT group. Additionally, the distribution of fat was not significant in the circumferences of the abdominal/thigh and chest/thigh ratios between the groups. Furthermore, there were no significant differences in the average skinfold abdominal + flank + subscapular/triceps + thigh ratio or abdominal + flank + subscapular/thigh ratio (data not shown). After adjustment for gestational age, maternal pregravid weight, weight at last antenatal visit, race, smoking status, and maternal and paternal height, there was no significant difference in birth weight (P = .15) or FFM (P = .60); however, a significant difference persisted in the percentage of body fat (P = .0005) between GDM and NGT. After an adjustment for the aforementioned factors plus a family history of diabetes mellitus and maternal weight
Catalano et al 1701
Volume 189, Number 6 Am J Obstet Gynecol
Table II. Neonatal anthropometrics Measure
GDM (n = 195)
Gestational age at delivery (wk) Gestiational age dating criteria (%) Ultrasound < 20 wk Ultrasound >20 wk and Dubowitz dates Dubowitz dates Neonatal Sex (%) Male Female Weight (g) Length (cm) Extremity length (cm) Upper arm Lower arm Upper leg Lower leg Circumferences (cm) Head Chest Umbilical Upper arm Lower arm Upper leg Lower leg Ponderal index (weight/length3) 3 100 Skinfolds (mm) Triceps Subscapular Flank Thigh Abdominal wall Weight for gestational age (%) < 10% 10%-90% >90% Weight (%) >4000 g < 4000 g Data are presented as mean
38.4
NGT (n = 220)
± 1.3
P value
± 1.4
.0001
73 22 5
71 20 9
.57
51 49
54 46
39.0
.57
3398 49.9
± 550 ± 2.1
3337 50.0
± 549 ± 2.2
.26 .68
8.2 7.3 9.7 8.3
± 0.6 ± 0.6 ± 0.9 ± 0.7
8.4 7.4 9.9 8.5
± 0.6 ± 0.5 ± 0.9 ± 0.7
.002 .03 .09 .02
34.7 31.8 31.3 10.6 9.7 15.2 11.1 2.72
± 1.4 ± 2.0 ± 2.3 ± 1.3 ± 1.0 ± 1.6 ± 1.1 ± 0.28
34.5 31.8 31.0 10.4 9.6 15.0 11.0 2.66
± 1.5 ± 2.1 ± 2.2 ± 1.0 ± 0.9 ± 1.6 ± 1.0 ± 0.26
.18 .88 .16 .18 .16 .16 .25 .02
4.7 5.4 4.2 6.0 3.5
± 1.1 ± 1.4 ± 1.2 ± 1.4 ± 0.9
4.2 4.6 3.8 5.4 3.0
± 1.0 ± 1.2 ± 1.0 ± 1.5 ± 0.8
.0001 .0001 .0001 .0001 .0001
5 68 27
6 79 14
.007
12 88
9 91
.37
± SD.
gain, there still remained a significant difference (P = .04) in fat mass between groups. Because more of the infants of women with GDM were macrosomic (ie, >90th percentile; 27% vs 14%) compared with the neonates of women with NGT , we next evaluated only neonates whose weights were average for gestational age (ie, between the 10th and 90th percentiles). There were 132 infants of women with GDM and 175 infants of women with NGT for analysis. The subjects with GDM were significantly shorter and had greater pregravid weight and weight at last antenatal visit as compared with the NGT group. The subjects with GDM also had a greater incidence of positive smoking history, family history of diabetes mellitus, and less post-high school education compared with the NGT group. The infants of the women with GDM were delivered more often by cesarean delivery (36% vs 25%, P = .03) compared with the NGT group. The average gestational age at delivery was again significantly earlier in infants of women with GDM (38.5 ± 1.4 weeks of gestation vs 39.0 ± 1.4 weeks of gestation,
Table III. Neonatal body composition GDM (n = 195) FFM (g) Fat mass (g) Body fat (%)
2962 436 12.4
± 405 ± 206 ± 4.6
Data are presented as mean
NGT (n = 220) 2975 362 10.4
± 408 ± 198 ± 4.6
P value .74 .0002 .0001
± SD.
P = .002), but there was no significant difference in the mean birth weight (3202 ± 357 g vs 3249 ± 372 g, P = .27) between groups. Infants of women with GDM had significantly greater skinfold measures at all five sites in comparison with the infants of women with NGT (P = .007 .0001). TOBEC estimates of body composition showed that the infants of the women with GDM had significantly less FFM (2832 ± 286 g vs 2919 ± 287 g, P = .008) but had significantly greater fat mass (371 ± 163 g vs 329 ± 150 g, P = .02) percent body fat (11.4% ± 4.6% vs 9.9% ± 4.0%, P = .002) in comparison
1702 Catalano et al
December 2003 Am J Obstet Gynecol
Table IV. Multiple stepwise regression analysis for factors that affected fetal growth and body composition in 220 women with NGT Factor Weight Gestational age Maternal height Maternal weight gain Glucose screen Smoking status (–) FFM Gestational age Maternal height Maternal weight gain Fat mass Gestational age Weight at first antenatal visit Maternal weight gain Glucose screen Body fat (%) Gestational age Glucose screen Maternal weight gain Pregravid weight
r2 0.22 0.25 0.28 0.30 0.31 0.20 0.23 0.24
Dr2
0.03 0.03 0.02 0.01
Table V. Multiple stepwise regression analysis for factors that affected fetal growth and body composition in 195 women with GDM
P value
.007 .01 .008 .05
0.03 0.01
.0001 .002 .05
0.16 0.21 0.24 0.26
0.05 0.03 0.02
.0001 .004 .006 .01
0.14 0.18 0.20 0.23
0.04 0.02 0.03
.0001 .01 .001 .01
with the infants of women with NGT. There was no significant difference in the distribution of body fat between groups. After an adjustment was made for maternal height and pregravid weight, gestational age, smoking status, race, and paternal weight, there was no significant difference in birth weight (P = .42) or FFM (P = .59). Adjustment for the aforementioned factors plus weight at the last antenatal visit and a family history of diabetes mellitus in fat mass (P = .02) and percent body fat (P = .02) continued to be significant between groups (detailed data available on request). We next compared the body composition estimates between the infants of the women with GDM who were treated with diet alone (group A1, n = 128 infants) and infants who were treated with diet plus insulin (group A2, n = 67 infants) on the average. The A2 GDM group had greater pregravid weight (87.9 ± 25.5 kg vs 77.4 ± 21.1 kg, P = .0002). The A2 GDM group also had greater parity (49% > 1.0 vs 32% > 1.0, P = .02) and cesarean delivery (52% vs 29%, P = .001) compared with the A1 GDM group. The A2 GDM group was delivered earlier (37.9 ± 1.1 weeks of gestation vs 38.7 ± 1.4 weeks of gestation, P = .0001) than the A1 GDM group. There was no significant difference in birth weight (3497 ± 556 g vs 3346 ± 542 g, P = .07) between groups. The skinfold measures were significantly greater in the A2 GDM group: subscapular (5.9 ± 1.6 mm vs 5.1 ± 1.2 mm, P = .0001), triceps (5.0 ± 1.1 mm vs 4.5 ± 1.0 mm, P = .003), flank (4.5 ± 1.2 mm vs 4.0 ± 1.1 mm, P = .01), thigh (6.3 ± 1.5 mm vs 5.8 ± 1.3 mm, P = .02), and abdomen (4.0 ± 1.4 mm vs 3.3 ± 0.9 mm, P = .0001) compared with the NGT group. There were no significant differences in the
Factor Weight Gestational age Fasting glucose Smoking status (–) FFM Gestational age Smoking status (–) Fasting glucose Maternal height Fat mass Fasting glucose Gestational age Body fat (%) Fasting glucose Gestational age
r2
Dr2
P value
0.07 0.17 0.21
0.10 0.04
.0003 .0001 .003
0.07 0.14 0.19 0.22
0.07 0.05 0.03
.0003 .0006 .0005 .02
0.11 0.17
0.06
.0001 .0007
0.09 0.13
0.04
.0001 .007
TOBEC estimates of FFM (3005 ± 397 g vs 2939 ± 409 g, P = .28), but there were significantly greater fat mass (492 ± 215 g vs 407 ± 196 g, P = .006) and percent body fat (13.6% ± 4.6% vs 11.7% ± 4.5%, P = .007) in the A2 GDM group compared with the A1 GDM group. After the adjustment for gestational age, maternal pregravid weight, parity, and smoking status, there were significant differences in birth weight (P = .0006) and FFM (P = .03). Because smoking affects FFM and not fat mass, when we adjusted for gestational age, maternal pregravid weight, and parity, the A2 GDM group persisted in having greater fat mass (P = .002) and percent body fat (P = .003) as compared with the A1 GDM group. We performed a multiple stepwise regression analysis to assess the factors and relative strength of each variable in regard to birth weight and body composition in the NGT group (Table IV) and GDM group (Table V). The purpose of this analysis was to identify factors that are available in clinical practice and potentially amenable to intervention to decrease the risk of adiposity. The independent variables for the NGT group included maternal age, height, pregravid weight, weight at last visit, weight gain, parity, paternal height and weight, gestational age, neonatal gender, presence of family history of diabetes mellitus, race, smoking status, and glucose screen. In the GDM group, we included all the variables for the NGT group with the exception of the glucose screen, but included the fasting, 1-, 2-, and 3-hour glucose values from the diagnostic oral glucose tolerance test. In the NGT group, gestational age had the strongest correlation with birth weight and body composition. Maternal height was a significant variable with respect to birth weight and FFM, whereas maternal pregravid weight, weight gain, and glucose screen were included in the regression models of fat mass and percent body fat. In the GDM group, gestational age had the strongest correlation with birth
Volume 189, Number 6 Am J Obstet Gynecol
weight and FFM, although fasting glucose had the strongest correlation with fat mass and percent body fat. Comment The most significant finding of our study was that of a significant increase in adiposity of average for gestational aged infants of women with GDM. Previous studies by Pettitt et al3 and Silverman et al7 have shown that infants of women with GDM have an increased risk of adolescent obesity and glucose intolerance. However, detailed body composition data at the time of birth was not available. Earlier studies ascribed the increases in birth weight of infants of diabetic women to either fat mass and/or FFM. The carcass analysis studies of Fee and Weil13 in stillbirths and cases of neonatal death described normal body water in FFM but significant increases in body fat in comparison with control subjects. In 40 infants of women with GDM, Whitelaw14 reported that neonatal fat mass, as estimated from skinfolds, was significantly greater in infants of women with diabetes mellitus compared with a nondiabetic reference population. Additionally, Enzi et al15 evaluated body composition using skinfolds in 17 infants of women with GDM and 17 nondiabetic control subjects. The mean ± SE percent body fat in the infants of the GDM was 17% ± 1.7% compared with 12.2% ± 0.5% in the control subjects. However, body weight was also greater in the GDM group (3508 ± 200 g) compared with the control group (3103 ± 100 g). In contrast, Naeye16 reported that there was an increase in FFM, in contrast with fat mass, in overgrown stillborn fetuses or neonatal deaths for women with GDM. These infants also had evidence of both hyperplasia and hypertrophy in cellular tissue of various organs. All these studies were conducted on a small number of infants, and the cause of the fetal/ neonatal death was often unknown. Additional body composition studies were conducted recently by Nasiat et al.17 Although the investigators reported an increase in subscapular fat in the infants of the women with diabetes mellitus (both pregestational and gestational), the birth weight of the infants of the women with diabetes mellitus was significantly greater than the birth weight of the nondiabetic control subjects. Furthermore, fat mass was estimated as the sum of four skinfolds. In a study by McFarland et al,18 although the authors found greater fat mass using an anthropometric formula in infants of women with diabetes mellitus, the number of study subjects was small and limited to macrosomic infants. The results of our study extend these results on the basis of method and number and analysis of study subjects. Thus, the current goal of achieving appropriate weight for gestational age may not be adequate given these results. The summary of the Fourth International WorkshopConference on Gestational Diabetes Mellitus19 noted that, although treatment of hyperglycemia can decrease
Catalano et al 1703
the risk of macrosomia markedly, optimal treatment has yet to be defined. Furthermore, maternal glucose concentrations, less than those used in the diagnosis of GDM, have also been related to increased fetal growth.20 Various treatment strategies, which include intensive treatment protocols to normalize maternal glucose concentration, are used currently in the treatment of GDM. However, such protocols may result in as many as 66% of women requiring insulin, and the overly aggressive treatment of maternal glycemia may actually increase the risk of small for gestational age neonates.21,22 Thirty-four percent of the patients with GDM (n = 67) were treated with insulin (A2 GDM group) in addition to diet and exercise. Thus, even in GDM patients who were treated with insulin, the neonates were significantly heavier and had increased adiposity in comparison with the infants of the A1 GDM group. We speculate that the A2 GDM group had greater pregravid insulin resistance as compared with the A1 GDM group. This is based on the A2 GDM group having significantly greater pregravid weight and parity compared with the A1 GDM group. These two factors have been associated with increased maternal insulin resistance and neonatal adiposity.23 Hence, the criteria for the diagnosis of GDM earlier in gestation may be warranted. The issue of whether the optimization of the control of fasting as compared with either preprandial or postprandial glucose concentrations is more important for the control of fetal overgrowth continues to be debated, as does the importance of glucose control in early, compared with late, gestation.24 Of interest, in our study, fasting glucose levels at the time of the oral glucose tolerance test was the factor with the strongest correlation with fetal fat accretion. These data support a study by UvenaCelebrezze et al25 that reported that fasting glucose had the strongest correlation with fetal fat mass in comparison with premeal or postmeal glucose values, based on glucose self-monitoring seven times per day. In summary, infants of women with mild glucose intolerance (ie, GDM) have increased body fat compared with infants of women with NGT; this is independent of birth weight and may be regarded as very early onset obesity. We propose that the use of specific outcome measurements such as fetal adiposity may be a more sensitive and specific outcome measure at birth. The increase in body fat at birth may be the harbinger of an increased risk of obesity and other metabolic complications in adolescence for these infants. Although the difference in fat mass between the two groups was only 74 g (Table III), this represents a 20% difference between the NGT and GDM infants. Should this 20% difference in fat mass persist into adolescence, it will become a significant problem. Therefore, we are in the process of following this cohort to evaluate long-term outcome that is relative to adolescent obesity, glucose tolerance, and the insulin resistance syndrome.
1704 Catalano et al
REFERENCES 1. Pedersen J. The pregnant diabetic and her newborn: problems and management. Baltimore: Williams & Wilkens; 1967. 2. Reiher H, Fahrmann K, Noack S, Woltanski KP, Jutzi E, Hahn UM, et al. Age dependent insulin secretion of the endocrine pancreas in vitro from fetuses of diabetic and nondiabetic patients. Diabetes Care 1983;5:446-51. 3. Pettitt DJ, Baird HR, Aleck KA, Bennett PH, Knowler WC. Excessive obesity in offspring of Pima Indian women with diabetes during pregnancy. N Engl J Med 1983;308:242-5. 4. Metzger BE. The organizing committee: summary and recommendations of the third international workshop: conference on gestational diabetes mellitus. Diabetes 1991;40(Suppl):197-201. 5. Sparks J. Human intrauterine growth and nutrient accretion. Semin Perinatol 1984;18:74-93. 6. Mokdad AH, Engelgau MN, Ford ES, Vinicor F, Bowman BA, Marks JS, et al. Diabetes trends in the US: 1990-1998. Diabetes Care 2000; 23:1278-83. 7. Silverman BL, Rizzo TA, Cho NH, Metzger BE. Long-term effects of the intrauterine environment: the Northwestern University Diabetes in Pregnancy Center. Diabetes Care 1998;21(Suppl): B142-9. 8. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:1039-57. 9. Institute of Medicine, National Academy of Sciences, Food and Nutrition Board. Nutrition during pregnancy, 1: weight gain. Washington (DC): National Academy Press; 1990. 10. Amini SB, Catalano PM, Hirsch V, Mann LI. An analysis of birth weight by gestational age: using a computerized perinatal data base. Obstet Gynecol 1993;83:342-52. 11. Catalano PM, Thomas AJ, Avallone DA, Amini SB. Anthropometric estimation of neonatal body composition. Am J Obstet Gynecol 1995;173:1176-81. 12. Fiorotto ML, Cochran WJ, Runk RC, Sheng J-P, Klish WJ. Total body electrical conductivity measurements: effects of body composition and geometry. Am J Physiol 1987;252:R798-800. 13. Fee BA, Weil WB. Body composition of infants of diabetic mothers by direct analysis. Ann N Y Acad Sci 1963;110:869-97.
December 2003 Am J Obstet Gynecol
14. Whitelaw A. Subcutaneous fat in newborn infants of diabetic mothers: an indication of quality of diabetic control. Lancet 1977; 1:15-8. 15. Enzi G, Inelman EM, Caretta F, Villani F, Zanardo V, DeBiasi F. Development of adipose tissue in newborns of gestational diabetes and insulin diabetic mothers. Diabetes 1980;29:100-4. 16. Naeye RL. Infants of diabetic mothers: a quantitative morphologic study. Pediatrics 1965;35:980-8. 17. Nasiat H, Abalkhail B, Fageeh W, Shabat A, Zahrawy FE. Anthropometric measurements of newborns of gestational diabetic mothers: does it indicate disproportionate fetal growth? J Matern Fetal Med 1997;6:291-5. 18. McFarland MB, Trylovich CG, Langer O. Anthropometric differences in macrosomic infant of diabetic and non-diabetic mothers. J Matern Fetal Med 1998;6:292-5. 19. Metzger BE, Coustan D. The organizing committee summary and recommendations of the fourth international workshop: conference on gestational diabetes mellitus. Diabetes Care 1998;21(Suppl):B161-7. 20. Tallarigo L, Giampietro O, Penno G, Milcoli R, Gregore G, Navalesi R. Relation of glucose tolerance to complications of pregnancy in nondiabetic women. N Engl J Med 1986;315:989-92. 21. Langer O, Rodriguez DA, Xenaris EM, McFarland MB, Berkus MD, Arrendondo F. Intensified versus conventional management of gestational diabetes. Am J Obstet Gynecol 1994;170:1036-47. 22. Langer O, Levy J, Brustman L, Anyaegbunam A, Merkatz R, Divon M. Glycemic control in gestational diabetes mellitus: how tight is tight enough? Small for gestational age versus large for gestational age. Am J Obstet Gynecol 1989;161:645-53. 23. Catalano PM, Drago NM, Amini SB. Maternal carbohydrate metabolism and its relationship to fetal growth and body composition. Am J Obstet Gynecol 1995;172:1464-70. 24. Jovanovic-Peterson L, Peterson CM, Reed GF, Metzger BE, Mills JL, Knopp RH, et al. Maternal postprandial glucose levels and infant birth weight: the Diabetes in Early Pregnancy Study. Am J Obstet Gynecol 1991;164:103-11. 25. Uvena-Celebrezze J, Fung C, Thomas AJ, Huston-Presley L, Amini SB, Catalano PM. Relationship of neonatal body composition to maternal glucose control in women with gestational diabetes mellitus. J Matern Fetal Neonat Med 2002;12:396-401.