Glycemic Control Throughout Pregnancy and Fetal Growth in Insulin-Dependent Diabetes

K. RAYCHAUDHURI, MRCS, MRCOG, AND M. J. A. MARESH, MD, FRCOG

Objective: To determine the time of growth acceleration in fetuses of insulin-dependent diabetic women who are large for gestational age (LGA) at birth and the relationship between growth acceleration and diabetic control throughout pregnancy. Methods: We studied a consecutive sample of 76 women with insulin-dependent diabetes divided by those who delivered LGA or normally grown infants. Fetal abdominal circumference (AC) was measured ultrasonically at regular intervals between 20 and 34 weeks gestation. Diabetic control was assessed by regular measurement of glycosylated hemoglobin and capillary blood glucose levels. Results: A signicant difference in fetal AC between groups developed between 20 and 24 weeks gestation, and the LGA group continued to have accelerated fetal growth. Between 18 and 24 weeks glycosylated hemoglobin and capillary blood glucose concentrations were signicantly higher in women who delivered LGA infants. After 28 weeks, blood glucose concentrations and glycosylated hemoglobin did not differ signicantly between groups. There was a nonsignicant trend toward more vaginal deliveries in the normal group (45% versus 32%). Conclusion: In insulin-dependent diabetic pregnancy, although actual growth acceleration occurred from about 20 weeks gestation, growth potential of fetuses appeared to be determined by prevailing maternal glucose concentrations before then. Excessive growth continued despite subsequent satisfactory glucose control. If strict blood glucose control is maintained during rst and second trimesters, it might reduce the incidence of LGA infants. (Obstet Gynecol 2000; 95:190 4. 2000 by The American College of Obstetricians and Gynecologists.)

Macrosomia, a major cause of fetal morbidity and mortality, occurs in a statistically signicant proportion of fetuses of pregnant women with insulin-dependent diabetes, despite relatively good glycemic control. Rates of 20 40% of women who had infants with birth weights over the 90th percentile have been quoted even
From the Department of Obstetrics and Gynaecology, St. Marys Hospital, Manchester, United Kingdom. Julie Morris provided statistical assistance.

from tertiary centers.1 Although perinatal mortality rates have decreased during the past decade and have reached a plateau2,3 and congenital malformations in fetuses of diabetic women are reduced by good glycemic control in the periconceptional period,4 the incidence of fetal macrosomia and its associated complications remains high. This might be because clinicians or pregnant, diabetic women are complacent about optimum blood glucose concentrations in diabetic pregnancies or are confused because of differing recommendations for target blood glucose concentration.2,5 8 Attempts to correlate birth weight and macrosomia with diabetic control need to allow for maternal factors such as obesity and excessive weight gain in pregnancy, which might contribute to developing LGA neonates.9,10 Strict control of blood glucose concentration is associated with small for gestational age (SGA) infants.11 Poor correlation between blood glucose concentration and birth weight might also relate to the gestational age at which tight control was achieved. Some of those factors might account for conicting results of some of the studies3,1215 that investigated the relationship between birth weight and blood glucose concentration during different trimesters. A relationship between LGA infants and fasting blood glucose concentration between 27 and 32 weeks gestation has been shown12 and also a relationship between birth weight and glycemic control before 32 weeks.13 Stubbs et al14 showed no relationship between diabetic control in the rst and second trimesters and subsequent birth weight. Peck et al3 did show a relationship between rst-trimester glycemic control and birth weight, as did Peterson et al.15 In the latter study, there was no association when allowance was made for differences in subsequent trimesters, and third-trimester nonfasting levels appeared to be the most inuential.15 Birth weight does not give the full picture of the fetal growth prole. The aims of the current study were to determine fetal growth rates using ultrasound and to determine when growth acceleration commenced in

190 0029-7844/00/$20.00
PII S0029-7844(99)00485-8

Obstetrics & Gynecology

LGA fetuses. Detailed data on diabetic control throughout pregnancy was necessary to determine relationships between growth acceleration and diabetic control.

Table 1. Clinical Characteristics

Characteristic Maternal age (y) Married Smokers Primigravid Initial weight (kg) Weight at delivery (kg) White class F/R Delivery gestation (days) AGA (n 42) 28.3 5.3 30 (71) 7 (29) 26 (62) 65.2 9.6 76.6 11.3 9 (21) 249 23 LGA (n 34) 28.5 4.7 27 (84) 3 (10) 15 (47) 66.4 7.7 80.3 7.7 6 (18) 253 12

Methods
Pregnant, insulin-dependent diabetic women were cared for by one physician and one obstetrician, who worked in a tertiary center clinic with diabetes specialist midwives and a dietician. Women were treated using standardized protocols with set series of investigations at regular intervals. The case notes of insulin-dependent diabetic women who delivered at 24 or more weeks gestation between 1990 and July 1995 were reviewed. There were no multiple pregnancies during the study period. For women who had more than one pregnancy, only the rst was used, which resulted in 82 deliveries. Each mother was classied according to her neonates birth weight. Neonates were grouped as LGA (above the 90th percentile), appropriate for gestational age (AGA) (10 90th percentile), and small for gestational age (SGA) (below the 10th percentile), corrected for gestational age and sex.16 Women were booked by 12 weeks gestation and each had a dating ultrasonographic scan in the rst trimester. After that examination, women had anomaly scans between 18 and 20 weeks, followed by serial growth scans from about 24 weeks until 36 38 weeks gestation. Despite a rigid protocol, some women had their scans 1 or more weeks before or after the desired gestational age. For those cases, measurements were estimated so that if scans were done at 19 and 21 weeks, those measurements were plotted on a modied standard abdominal circumference growth chart17 and the measurement at 20 weeks gestation was extrapolated from the chart, assuming linear growth. Using that method, it was possible to calculate an abdominal circumference (AC) measurement of all fetuses at 20, 24, 28, 32, and 34 completed weeks. More scans were done between weeks 24 37 in the LGA group than in the AGA group (mean 6.5 versus 4.5, P .05). Glycosylated hemoglobin concentrations were measured in each woman at the initial visit and subsequently every 4 weeks until 36 37 weeks gestation. Automated liquid chromotography was used with a reference range in a normal population of 4.84 0.46% (mean standard deviation [SD]). The mean values at 18 (16 19), 24 (2225), 28 (26 29), and 32 (30 33) completed weeks were used. Women were assessed by diabetes specialist midwives to ensure that they could correctly use meters to self measure their capillary blood glucose concentrations. Women took the measurements about half an hour before eating three times per day (early morning,

AGA appropriate for gestational age, LGA large for gestational age. There were no statistically signicant differences between groups (t test, 2 test). Data are given as mean standard deviation or n (%).

midday, and early evening) and also at bedtime. No data stored on the meters were used. The mean daily capillary blood glucose concentrations were calculated from the four daily recordings on the home blood glucose monitoring charts. Those recordings were made between 16 and 33 completed weeks for each woman, and mean values at 18 (mean 16 19 weeks), 24 (mean 2225 weeks), 28 (mean 26 29 weeks), and 32 (mean 30 33 weeks) completed weeks gestation were calculated and used for analysis. Information about mode of delivery, gestational age at delivery, perinatal mortality, and admission to the neonatal intensive care unit were obtained from chart reviews. For statistical analysis, two-way repeated-measures analysis of variance was used for comparing glycosylated hemoglobin, blood glucose, and fetal AC. In addition, the unpaired t test and 2 test were used.

Results
There were 34 LGA (41%), 42 AGA (51%), and six SGA (7%) infants. The SGA group were not considered further. Characteristics of mothers of LGA and AGA infants are shown in Table 1. All women were white. There was no signicant difference between groups with regard to age, marital status, smoking habits, parity, weight at initial visit, weight at delivery, and gestation at delivery. There were also similar numbers of women in both groups who had diabetic vascular problems of nephropathy, proliferative retinopathy, or both (White class F/R). (Three of six women who had SGA infants also were in class F/R). There were two infants with congenital cardiac abnormalities, one in each group. During the study, one other woman from the clinic had a second-trimester termination for neural tube defect. Mean ultrasonographic fetal AC did not differ significantly between LGA and AGA groups at 20 weeks

VOL. 95, NO. 2, FEBRUARY 2000

Raychaudhuri and Maresh

Fetal Growth and Diabetes 191

Table 2. Fetal Abdominal Circumference

Ultrasound AC (cm) Gestation (wk) 20 24 28 32 34

Table 4. Capillary Blood Glucose Concentration

Gestation (wk) Blood glucose (mmol/L) LGA (n 34) Mean 95% condence interval AGA (n 42) Mean 95% condence interval Abbreviations as in Table 1. 18 7.3 6.6, 7.9 6.1 5.6, 6.6 24 7.2 6.5, 8.0 5.9 5.5, 6.3 28 6.8 6.3, 7.3 6.4 6.0, 6.8 32 6.5 6.0, 7.1 6.0 5.6, 6.4

LGA (n 34) Mean 15.5 21.3 26.3 31.4 33.7 95% condence 15.2, 15.8 20.8, 21.8 25.8, 26.8 30.9, 31.9 33.1, 34.3 interval AGA (n 42) Mean 15.2 19.6 24.7 29.0 31.4 95% condence 15.0, 15.4 19.3, 20.0 24.3, 25.1 28.6, 29.4 30.9, 31.9 interval AC abdominal circumference; LGA large for gestational age; AGA appropriate for gestational age.

gestation. However, repeated-measures analysis of variance showed a signicant difference between groups at 20 24 weeks and 24 34 weeks (P .001). There was a signicant difference in interaction between groups over those time periods (P .001) suggesting that growth acceleration in the LGA group that started at 20 weeks continued after 24 weeks. The mean fetal AC values, with condence intervals, are shown in Table 2. Mean maternal glycosylated hemoglobin concentrations were signicantly higher in the LGA compared with the AGA group during weeks 18 24 (P .02). The signicant rate of reduction of glycosylated hemoglobin concentration over the period (P .001) was similar between groups. When we analyzed data over the 18 32-week period, there was no signicant difference between groups. Both groups showed signicant decreases in glycosylated hemoglobin concentration between 18 32 weeks (P .001), but analysis of the interaction showed that there was a signicant difference (P .001) between groups. Examination of mean glycosylated hemoglobin concentrations (Table 3) showed why there was a difference. The groups differed signicantly at 20 weeks, but the AGA group had no difference with time, whereas the LGA group showed a reduction in concentration of glycosylated hemoglobin that resulted in a signicant difference in the interaction.
Table 3. Glycosylated Hemoglobin Concentration
Glycosylated hemoglobin (%) LGA (n 34) Mean 95% condence interval AGA (n 42) Mean 95% condence interval Abbreviations as in Table 1. Gestation (wk) 18 7.5 7.2, 7.8 6.9 6.5, 7.2 24 7.0 6.7, 7.4 6.5 6.2, 6.9 28 6.8 6.5, 7.2 6.8 6.5, 7.2 32 6.8 6.4, 7.2 6.9 6.5, 7.3

The mean maternal capillary blood glucose concentration was also signicantly higher in the LGA group between 18 24 weeks compared with the AGA group (P .003), but there was no signicant change with time. Between 28 32 weeks there was no signicant difference between groups; however, between 18 32 weeks there was a signicant difference between the two groups (P .01). Analysis of interaction again showed a signicant difference in blood glucose concentrations between groups during the 18 32-week period, similar to that of glycosylated hemoglobin concentration. The mean capillary blood glucose measurements (Table 4) showed that initial values at 18 weeks were signicantly different, but that those mean values in the AGA group did not change subsequently, whereas in the LGA group they decreased. Modes of delivery did not differ signicantly between groups, although there was a trend toward more normal vaginal deliveries in the AGA (n 19, 45%) compared with the LGA group (n 11, 32%). There was a similar number of cesarean deliveries in each group (AGA n 17, 40%; LGA n 15, 44%) with a trend toward more cesareans in the LGA group being done during labor, the indication usually being slow progress (AGA n 8, 20%; LGA n 12, 35%). There were no stillbirths in either group. In the AGA group, six (14%) infants were admitted to the neonatal intensive care unit for up to 2 days, and 5 (15%) in the LGA group were admitted. Indications included mild hypoglycemia, prematurity, and observation; those cases are not discussed further. There were six (14%) infants in the AGA group and eight (24%) in the LGA group admitted for more than 2 days. Those included the two infants with congenital cardiac problems, one in each group. One infant in the LGA group was admitted after signicant shoulder dystocia. One infant in each group had respiratory problems. Hypoglycemia was a major problem in four infants in the LGA group and three in the AGA group. All infants were discharged alive from the neonatal unit.

192 Raychaudhuri and Maresh

Fetal Growth and Diabetes

Obstetrics & Gynecology

Discussion
Elevated blood glucose concentration in diabetic pregnancy is associated with worse maternal and perinatal outcomes. Of all perinatal outcomes, congenital abnormalities and unexplained fetal deaths are the most serious, and fetal macrosomia is the most common. The current study showed a high incidence of fetal macrosomia, in keeping with recent reviews.1 The maternal consequence of fetal macrosomia is increased maternal morbidity resulting from instrumental and operative deliveries. The current study showed a trend toward that when infants were LGA. Although there were no perinatal deaths, morbidity secondary to fetal macrosomia might occur with fetal trauma at vaginal delivery, and increased neonatal respiratory morbidity might be due to the higher cesarean rate. Although there was one case of shoulder dystocia in an LGA infant, respiratory problems in that group were not more frequent, which agrees with the nding of no signicant differences in gestational age at delivery or cesarean rate. Although one might expect hypoglycemia to be more common in macrosomic infants, similar numbers in both groups had signicant hypoglycemic problems. All mothers and infants were treated by staff with considerable experience, and whether the results can be extrapolated to smaller centers is not known. Whether there is any long-term morbidity from being born macrosomic is also unknown. Factors that might affect birth weight include maternal body mass index,9 excessive weight gain in pregnancy,9,10 ethnic background, and age, but both groups in the current study had similar characteristics. The only signicant differences between groups were that the glycosylated hemoglobin and blood glucose concentrations were higher in the LGA group than in the AGA group early in the second trimester. Although both groups had comparable blood glucose and glycosylated hemoglobin concentrations after 24 weeks gestation, growth acceleration continued in the LGA group, which implies that diabetic control in the rst half of pregnancy inuences birth weight. Previous studies that compared birth weight and diabetic control throughout pregnancy have produced conicting results.3,1215 Some studies showed an association between higher blood glucose concentrations in the rst half of pregnancy and delivery of LGA infants,3,15 although one reported that the association disappeared when differences in each trimester were considered and that the only persistent signicant relationship was third-trimester diabetic control.15 The advantage of the current study was that it documented fetal growth by using frequent ultrasonographic scans so that the effects of diabetic control could be correlated

with growth. Growth acceleration (assessed using fetal AC) in LGA infants had commenced by 20 24 weeks if not earlier. Because glycosylated hemoglobin indicates the previous 6 8 weeks blood glucose concentration, our ndings suggest that diabetic control before 24 weeks or possibly 20 weeks is critical for determining birth weight. After 24 weeks, fetal growth rates continued to differ signicantly between groups, although by 28 weeks there was no signicant difference between groups in maternal glucose control. The similarity in control probably indicated attempts to tighten diabetic control from 26 28 weeks, when accelerated fetal growth was suspected. Thus, it appeared that fetal growth acceleration was determined in the rst half of pregnancy and continued despite improvements in diabetic control. Our results are in accord with Pedersens hyperglycemic hyperinsulinemic theory18 that maternal hyperglycemia causes fetal hyperglycemia and resultant fetal pancreatic B cell hyperplasia. That causes fetal hyperinsulinemia and results in hyperplasia of the adiopocytes. In that case it was postulated that exposure to high blood glucose concentration in early gestation could lead to accelerated fetal growth later in gestation through an irreversible insult on the fetal pancreas. Increased cells in the adipose tissue might account for increased growth in later gestation, although glycemic control has stabilized. To test that hypothesis further a randomized, controlled trial of an intervention in the late rst trimester or early second trimester would be needed to obtain good diabetic control and observe whether that reduced the incidence of macrosomia compared with women who were treated by current strategies.

References
1. Fraser R. Diabetic control in pregnancy and intrauterine growth of the fetus. Br J Obstet Gynaecol 1995;102:35. 2. Garner P. Type I diabetes mellitus and pregnancy. Lancet 1995; 346:157 61. 3. Peck RW, Price DE, Lang GD, MacVicar J, Hearnshaw JR. Birthweight of babies born to mothers with type I diabetes: Is it related to blood glucose control in the rst trimester? Diabetes Med 1991;8:258 62. 4. Kitzmiller JL, Gavin LA, Gin GD, Peterson LJ, Main EK, Zigrang WD. Preconception care of diabetes glycemic control prevents congenital anomalies. JAMA 1991;265:731 6. 5. Combs CA, Gunderson E, Kitzmiller JL, Gavin LA, Main EK. Relationship of fetal macrosomia to maternal postprandial glucose control during pregnancy. Diabetes Care 1992;15:12517. 6. Langer O. Prevention of macrosomia. In: Oats JN, ed. Diabetes in pregnancy. Baillieres clinical obstet and gynecol. London: Bailliere Tindall, 1991;5.2:333 47. 7. Karlsson K, Kjellmer I. The outcome of diabetic pregnancies in relation to the mothers blood sugar level. Am J Obstet Gynecol 1972;112:21320.

VOL. 95, NO. 2, FEBRUARY 2000

Raychaudhuri and Maresh

Fetal Growth and Diabetes 193

8. Farrag OAM. Prospective study of 3 metabolic regimens in pregnant diabetics. Aust N Z J Obstet Gynaecol 1987;27:6 9. 9. Boyd ME, Usher RH, McLean FH. Fetal macrosomia: Prediction, risks, proposed management. Obstet Gynecol 1983;61:71522. 10. Madsen H, Ditzel J. The inuence of maternal weight, smoking, vascular complications and glucose regulation on the birthweight of infants of type I diabetic women. Eur J Obstet Gynaecol Reprod Biol 1991;39:1759. 11. Roversi GD, Gargiulo M, Nicoloni U, Pedretti E, Marini A, Barbarani V, et al. A new approach to the treatment of diabetic pregnant women. Am J Obstet Gynecol 1979;135:56776. 12. Persson B, Hanson U. Fetal size at birth in relation to quality of blood glucose control in pregnancies complicated by pregestational diabetes mellitus. Br J Obstet Gynaecol 1996;103:42733. 13. Lin CC, River J, River P, Blix PM, Moawad AH. Good diabetic control early in pregnancy and favourable fetal outcome. Obstet Gynecol 1986;67:51 6. 14. Stubbs SM, Leslie RDG, John PN. Fetal macrosomia and maternal diabetic control in pregnancy. BMJ 1981;282:439 40. 15. Peterson LJ, 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;164:10311. 16. Thomson AM, Billewicz WZ, Hytten FE. The assessment of fetal growth. J Obstet Gynaecol Br Commonwealth 1968;75:90316.

17. Campbell S, Wilkin D. Ultrasonic measurement of fetal abdomen circumference in the estimation of fetal weight. Br J Obstet Gynaecol 1975;82:689 97. 18. Pederson J. The pregnant diabetic and her newbornproblems and management. Copenhagen, Denmark: Munkgaard, 1977.

Address reprint requests to:

Dr. M. J. A. Maresh, MD, FRCOG Department of Obstetrics and Gynaecology Saint Marys Hospital Whitworth Park Manchester M13 OJH United Kingdom

Received February 11, 1999. Received in revised form June 18, 1999. Accepted July 15, 1999.

Copyright 2000 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.

STANDARDS

FOR REPORTING TRIALS

The CONSORT listing of standards for reporting randomized trials has been adopted as policy by Obstetrics & Gynecology. Investigators who are planning, conducting, or reporting randomized trials should be thoroughly familiar with these standards. A copy can be obtained by contacting: Obstetrics & Gynecology, 10921 Wilshire Boulevard, Suite 403, Los Angeles, CA 90024-3908; FAX (310) 208-2838.

194 Raychaudhuri and Maresh

Fetal Growth and Diabetes

Obstetrics & Gynecology