Fetal macrosomia is one of the major complications of diabetic pregnancy. According to Pedersen's hypothesis,1 fetal macrosomia is associated with fetal hyperglycemia and related hyperinsulinemia resulting from maternal hyperglycemia. Although diabetic mothers with poor glycemic control during pregnancy were more likely to deliver macrosomic infants compared with those who had good glycemic control,2,3 strict glycemic control sometimes failed to prevent macrosomia. It was reported that the risk of having a large for gestational age (LGA) infant was reduced if intensive glycemic control was begun before but not after 32 weeks' gestation.4 Those reports suggested that diabetic macrosomia is associated with maternal metabolic condition at a certain gestational age.
Conversely, even a minor degree of maternal glucose intolerance also represents increased risk of macrosomia.5 A significantly higher incidence of LGA infants was observed in women with abnormal diabetic screening but normal oral glucose tolerance test (GTT) in comparison with those who had negative diabetic screens.6,7 These findings suggest fetal growth is determined largely by maternal factors, including not only plasma glucose levels but also other fuels, such as lipids and amino acids,8 especially in nondiabetic women. Maternal serum lipid levels increase during mid to late gestation, which is believed to be beneficial to mother and fetus in terms of lactation and nutrition.9 A recent study reported that postprandial triglyceride but not postprandial glucose levels at diabetic screen at 24–28 weeks' gestation were significantly associated with relative birth weight (the observed birth weight/the 50th percentile birth weight for gestational age).10 However, it has not been documented conclusively whether elevated triglyceride levels are associated with the risk of fetal macrosomia. Furthermore, it is not known whether midpregnancy maternal fasting triglyceride levels are associated with birth weight, independent of maternal glucose levels and obesity.
We studied nondiabetic women with positive midpregnancy diabetic screen, a group at high risk of fetal macrosomia. Our objective was to determine whether maternal serum lipid levels, including triglyceride, free fatty acids, and total cholesterol, at 24–32 weeks' gestation are associated with newborn weight at term, and therefore, associated with a risk of developing an LGA infant and whether the association is independent of maternal obesity and plasma glucose levels.
Materials and Methods
We enrolled Japanese pregnant women who had positive diabetic screen test results (at least 135 mg/dL of plasma glucose level at 1 hour after 50-g oral glucose challenge) and a normal 75-g oral GTT at 24–32 weeks' gestation at Nagasaki University Hospital between November 1992 and October 1999. We excluded women with pregestational or gestational diabetes mellitus to eliminate therapeutic biases in the association between maternal metabolic variables and fetal growth. We also excluded women with hypertensive disorder, thyroid disorder, lupus, and antiphospholipid syndrome, because those conditions are associated with fetal growth restriction due to placental insufficiency rather than metabolic factors. Subjects who delivered before 37 weeks' gestation and cases of fetal congenital malformation or multifetal gestation were also excluded.
After an overnight fast, 75-g GTT was started between 9 and 10 AM. Fasting blood samples were drawn to measure plasma glucose, serum triglyceride, free fatty acids, and total cholesterol levels. One- and 2-hour blood samples after an oral glucose load were also drawn to measure plasma glucose concentration. Plasma glucose was measured by the oxidase method. Serum triglyceride, free fatty acids, and total cholesterol were measured using enzymatic methods. Prepregnant body mass index (BMI) was calculated from prepregnant weight and height reported by subjects. Maternal weight gain during pregnancy was defined as an increase in weight from prepregnant weight to weight at the last visit.
All subjects were followed up until delivery. Neonatal birth weight above the 90th percentile of the gender-specific Japanese birth weight curve was defined as LGA.11 Gestational age was estimated by last menstrual period and confirmed by early fetal ultrasonographic measurements in all subjects. We used the criteria of the Japan Society of Obstetrics and Gynecology using a 75-g GTT to diagnose gestational diabetes mellitus (GDM).12 The criteria included fasting plasma glucose at least 100 mg/dL, 1-hour plasma glucose at least 180 mg/dL, and 2-hour plasma glucose at least 150 mg/dL. Women with two or more abnormal values were given a diagnosis of GDM and excluded from the study.
We used univariable linear regression analysis to evaluate the association between each lipid concentration and newborn weight at term. Regarding the variables correlating with birth weight, the association was tested using analysis of covariance to control for confounding variables affecting the association. Possible confounds included maternal plasma glucose levels (fasting or postprandial), prepregnant BMI, maternal weight gain during pregnancy, fetal gender, and gestational age at birth. We also used multiple logistic regression analysis to test the hypothesis that maternal hyperlipidemia in midpregnancy is a risk factor for LGA. Maternal hyperlipidemia was defined as a value higher than the 75th percentile value of each lipid concentration. In the model, we used maternal fasting plasma glucose concentration at GTT, maternal prepregnant BMI, and weight gain during pregnancy as confounding variables. The χ2 test was used to compare the incidence of LGA infants between women with and without hyperlipidemia.
Sample size was estimated with the following assumptions: a correlation coefficient of 0.25 between fasting lipid concentration and newborn weight, alpha error of 0.05, and power of 0.9. The sample size required to detect the correlation coefficient was 164 subjects. We used StatView version 5.0 (SAS Institute Inc., Cary, NC) for statistical analysis, and P < .05 was defined as significant.
One hundred seventy-eight subjects were enrolled. We excluded 20 because they were given a diagnosis of GDM (n = 10) or pregnancy-induced hypertension (n = 10). We also excluded seven cases of fetal malformation and five cases of preterm delivery. Consequently, 146 women who had normal oral GTT results were studied.
Maternal clinical and metabolic characteristics and the correlation coefficient between each variable and newborn weight are summarized in Table 1. As well as maternal prepregnant BMI, maternal fasting serum triglyceride levels and fasting plasma glucose levels correlated significantly with newborn weight at term in univariable analysis. Maternal weight gains during pregnancy correlated positively but not significantly with newborn weight. Birth weight did not correlate significantly with postprandial plasma glucose concentration or the other fasting serum lipid levels. However, as expected, fasting triglyceride levels correlated significantly with prepregnant BMI (r = 0.32, P = .001) and fasting plasma glucose concentration (r = 0.19, P = .026). Therefore, in addition to gestational age at delivery, weight gain during pregnancy, and neonatal gender, these variables could confound the association between serum triglyceride levels and newborn birth weight. After controlling for prepregnant BMI, fasting plasma glucose levels, weight gain during pregnancy, gestational age at delivery, and fetal gender, the association between maternal fasting serum triglyceride levels and birth weight remained significant (Table 2). In this model, neither maternal fasting plasma glucose levels nor prepregnant BMI were significantly associated with birth weight.
Hypertriglyceridemia was defined as more than the 75th percentile value (259 mg/dL) of all subjects. The incidence of LGA infants was significantly higher in mothers with hypertriglyceridemia (four of 34, 12%) than in mothers who had normal triglyceride levels (one of 112, 0.9%) (P = .012). We applied a logistic regression model with confounding variables, including fasting plasma glucose levels, prepregnant BMI, and weight gain during pregnancy, to test whether maternal hypertriglyceridemia predicts risk of having an LGA infant at term, independent of fasting plasma glucose and maternal obesity. In the model, hypertriglyceridemia at 24–32 weeks' gestation was a significant predictor of having an LGA infant at term (odds ratio 11.6; 95% confidence interval 1.1, 122; P = .04), independent of maternal glucose levels and obesity (Table 3).
Maternal hyperlipidemia during pregnancy is associated with increased lipogenesis during the first two thirds of gestation.13 Maternal fat accumulation peaks in mid gestation and decreases in late gestation, whereas maternal serum lipid levels increase in mid to late pregnancy.14 It is considered a maternal adaptation to maintain stable fuel distribution to the fetus.8 In animal studies, shortage or excess of lipids in maternal circulation affected fetal growth and morbidity.15 However, the role of maternal hyperlipidemia in fetal growth is still not known.
In this study, we found that maternal fasting serum triglyceride levels at 24–32 weeks' gestation were significantly and positively associated with newborn weight at term. The association was independent of maternal prepregnant obesity, weight gain during pregnancy, or midpregnancy plasma glucose levels (either fasting or postprandial). The study by Knopp et al,10 in which similar eligibility criteria were used, found that triglyceride level after a 50-g oral glucose load at 24–28 weeks' gestation was significantly associated with adjusted newborn weight, independent of maternal obesity. The correlation coefficient was higher in our study (r = 0.22, P < .01) than in Knopp's report (r = 0.13, P < .05). Difference in maternal conditions, the fasting or the postprandial state, might have caused the variation in the studies. Although some investigators found an association between maternal free fatty acid concentration and neonatal birth weight,13 we did not find any relation of either total cholesterol or free fatty acids to newborn weight.
We also showed subjects who had fasting serum triglyceride levels above the 75th percentile at 24–32 weeks' gestation were at significant risk of having LGA infants at term, independent of maternal prepregnant BMI, maternal weight gain during pregnancy, and plasma glucose levels. Our findings regarding the effect of midpregnancy maternal hypertriglyceridemia on fetal overgrowth might be limited to women with positive diabetic screen. Recent reports suggest that women with positive results of diabetic screen, even with a normal GTT result, might be at risk of macrosomia in comparison with those who had negative results of diabetic screen.6,7 In addition, Knopp et al10 found a positive association between serum triglyceride levels after 50-g oral glucose load and adjusted newborn weight in women whose diabetic screen results were negative. Therefore, maternal triglyceride measurements in midpregnancy might have important clinical value to predict fetal overgrowth at term in nondiabetic pregnancies.
Another interesting finding in our study was the relationship between maternal plasma glucose levels and newborn weight. In nondiabetic and diabetic pregnancies, it is still controversial whether midpregnancy maternal fasting or postprandial plasma glucose level is an independent predictor of newborn weight at term. In the current study, newborn weight at term correlated significantly with fasting plasma glucose but not with postprandial levels at 24–32 weeks' gestation in univariable linear regression analysis. However, in the multiple logistic regression model that included maternal triglyceride, prepregnant BMI, and maternal weight gain during pregnancy as covariates, the association between fasting glucose level and birth weight did not remain statistically significant (P = .06). This result is consistent with another recent study in which there was no significant effect of plasma glucose values on macrosomia in screening-positive and GTT-negative subjects.16 Although it might be a result of a lack of statistical power because of our small sample size, it seemed that maternal serum triglyceride level had a more important effect on fetal growth than maternal plasma glucose in nondiabetic women.
The positive association between maternal triglyceride level and newborn weight is confusing, because maternal serum triglyceride does not appear to cross the placenta.13 With respect to the physiologic mechanism, it is hypothesized that hydrolysis of maternal triglyceride by placental lipoprotein lipase to free fatty acids that cross the placenta is increased.13 Enhanced insulin resistance during late pregnancy would explain the association between maternal triglyceride level and fetal growth. Hypertriglyceridemia is one of the major characteristics of insulin resistance syndrome in nonpregnant adults.17 Also in pregnancy, hyperlipidemia including hypertriglyceridemia was found in women with gestational diabetes,18,19 which is associated with both fetal macrosomia and decreased insulin sensitivity during pregnancy compared with normal pregnancy.20 Conversely, some authors reported that maternal insulin sensitivity was negatively associated with infantile birth weight.21,22 We speculate that maternal hypertriglyceridemia in midpregnancy reflects increased insulin resistance, and consequently, is associated with newborn birth weight.
1. Pedersen J. Weight and length at birth of infants of diabetic mothers. Acta Endocrinol 1954;16:330–42.
2. Landon MB, Gabbe SG, Piana R, Mennuti MT, Main EK. Neonatal morbidity in pregnancy complicated by diabetes mellitus: Predictive value of maternal glycemic profiles. Am J Obstet Gynecol 1987;156:1089–95.
3. 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;161:646–53.
4. Lin CC, River J, River P, Blix BA, Moawad AH. Good diabetic control early in pregnancy and favorable fetal outcome. Obstet Gynecol 1986;67:51–6.
5. Tallarigo L, Giampietro O, Penno G, Miccoli R, Gregoli G, Navalesi R. Relation of glucose tolerance to complications of pregnancy in nondiabetic women. N Engl J Med 1986;315:989–92.
6. Leikin EL, Jenkins JH, Pomerantz GA, Klein L. Abnormal glucose screening tests in pregnancy: A risk factor for fetal macrosomia. Obstet Gynecol 1987;69:570–3.
7. Okun N, Verma A, Mitchell BF, Flowerdew G. Relative importance of maternal constitutional factors and glucose intolerance of pregnancy in the development of newborn macrosomia. J Matern Fetal Med 1997;6:285–90.
8. Freinkel N. Banting lecture 1980: Of pregnancy and progeny. Diabetes 1980;29:1023–35.
9. Herrera E, Munilla MA. Maternal lipid metabolism and its implications for fetal growth. In: Battaglia FC, ed. Placental function and fetal nutrition, Nestlé Nutrition Workshop Series, Vol. 39. Philadelphia: Lippincott-Raven, 1997:169–82.
10. Knopp RH, Magee MS, Walden CE, Bonet B, Benedetti TJ. Prediction of infant birth weight by GDM screening tests: Importance of plasma triglyceride. Diabetes Care 1992;15:1605–13.
11. Nishida H, Sakamoto S, Sakanoue M. New fetal growth curves for Japanese. Acta Paediatr Scand 1985;319(Suppl):62–7.
12. The Committee on Nutrition and Metabolism of the Japan Society of Obsterics and Gynecology. The committee report. Acta Obstet Gynaecol Jpn 1984;36:2055–8.
13. Knopp RH, Bonet B, Zhu X. Lipid metabolism in pregnancy. In: Cowett RM, ed. Principles of perinatal-neonatal metabolism. 2nd ed. New York: Springer, 1998:221–58.
14. Ordovas JM, Pocovi M, Grande F. Plasma lipids and cholesterol esterification rate during pregnancy. Obstet Gynecol 1984;63:20–5.
15. Soria A, Chicco A, Mocchiutti N, Gutman RA, Lombardo YB, Martin-Hidalgo A, et al. A sucrose-rich diet affects triglyceride metabolism differently in pregnant and nonpregnant rats and has negative effects on fetal growth. J Nutr 1996;126:2481–6.
16. Verma A, Mitchell BF, Demianczuk N, Flowerdew G, Okun NB. Relationship between plasma glucose levels in glucose-intolerant women and newborn macrosomia. J Matern Fetal Med 1997;6:187–93.
17. Mantzoros CS, Flier JS. Insulin resistance: The clinical spectrum. In: Mazzaferi E, ed. Advances in endocrinology and metabolism. Vol. 6. St. Louis: Mosby, 1995:193–232.
18. Hollingworth DR. Alterations of maternal metabolism in normal and diabetic pregnancies: Differences in insulin-dependent, noninsulin-dependent, and gestational diabetes. Am J Obstet Gynecol 1983;146:417–29.
19. Meyers-Seifer CH, Vohr BR. Lipid levels in former gestational diabetic mothers. Diabetes Care 1996;19:1351–6.
20. Catalano PM, Tyzbyr ED, Wolfe RR, Calles J, Roman NM, Amini SB, et al. Carbohydrate metabolism during pregnancy in control subjects and women with gestational diabetes. Am J Physiol 1993;264:E60–7.
21. 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.
22. Caruzo A, Paradisi G, Ferrazzani S, Lucchese A, Moretti S, Fulghesu AM. Effect of maternal carbohydrate metabolism on fetal growth. Obstet Gynecol 1998;92:8–12.