Long-Term Impact of Neonatal Breast-Feeding on Body Weight and Glucose Tolerance in Children of Diabetic Mothers

During the study period (1980–1989), 368 ODM were delivered at the Clinic of Obstetrics and Gynecology, Berlin-Kaulsdorf, Berlin (former East Germany). Each mother was offered the opportunity to have her child participate in pediatric follow-up with regular physical reexaminations, which included an oral glucose tolerance test (OGTT). A total of 112 infants were brought for follow-up; complete data on nutrition during the early neonatal period (day 1–7 of life) was available for all of the infants.

Maternal demographic data included age, duration of diabetes (type 1), and maternal prepregnancy BMI. Socioeconomic status was categorized by maternal (or paternal) occupation: unemployed (mother, n = 2; father, n = 0), manual worker (mother, n = 20; father, n = 79), nonmanual worker without high school diploma (mother, n = 71; father, n = 22), and nonmanual worker with high school diploma (mother, n = 19; father, n = 11). Geographic origin was defined as urban (n = 76) or rural (n = 36). Gestational age and birth weight were taken from birth records.

Gestational diabetes (GDM) was diagnosed between the 26th and 28th weeks of gestation using a 50-g OGTT, as described in detail previously (10,11,20). After diagnosis of GDM, glucose homeostasis was monitored weekly by 24-h day-night glucose profiles at the clinic. During monitoring, blood samples were measured every 2 h using the glucosidase-peroxidase method. Women maintaining mean 24-h profiles <5.5 mmol/l were treated with diet therapy (n = 14). When a woman’s mean profile was ≥5.5 mmol/l, insulin therapy was initiated (n = 15). After insulin therapy was started in women with GDM, glycemic monitoring was instituted and insulin therapy was titrated in these women, as in women with type 1 diabetes, to achieve euglycemia (20).

After delivery, all mothers stayed with their newborns at the maternity ward of the clinic for at least 1 week. All were encouraged to breast-feed their newborns. Depending on the status of maternal lactation, when necessary (as judged by the pediatrician at the ward) a supplementary feeding with banked donor breast milk was offered to the neonate. Donor breast milk was provided by the local center for donation of breast milk at the Clinic of Pediatrics Lindenhof, Berlin, East Germany. Women with diabetes were excluded from donating breast milk; therefore, all banked breast milk was nondiabetic milk. Thereby, all newborns in the study received differing amounts of breast milk from their biological diabetic mothers and banked donor breast milk from unrelated nondiabetic women. Because of acute nonavailability of banked breast milk, in 23 infants, short-time supplementation was necessary with standard term formula milk (Milasan; Milchwerke Stendal, Stendal, East Germany). The mean nutrient contents per 100 ml of the formula milk were 3.8 g fat, 1.9 g protein, and 6.2 g carbohydrate, including 4.9 g lactose; energy content per 100 ml was 282 kJ.

During each feeding on days 1–7 postpartum, milk intake was determined using a test weighing protocol (4,21): the infant was weighed before and after every nursing by trained staff. The weight difference (in g) and the type of milk ingested during the meal was noted as breast milk from the biological (diabetic) mother (diabetic breast milk [DBM]), banked donor breast milk from unrelated nondiabetic mothers (nondiabetic banked donor breast milk [BBM]), or standard term formula milk. Mean volume of milk ingested per day (g/day) was calculated for each type of milk separately by summing the volumes ingested during the period (days 1–7 postpartum) and dividing by 7. Because formula milk was given to only 23 infants, a number too small for allowing valid statistical evaluation, further analysis was restricted to the impact of DBM and BBM feeding.

During follow-up examinations, at a median age of 2 years, body weight and height were recorded. From these data, relative body weight was calculated for all children in relation to age- and sex-specific standard population measures (rounded to year) as follows: relative weight (individual weight divided by median standard weight for age and sex) divided by relative height (individual height divided by median standard height for age and sex) × 100 (7,8,10). The resulting relative body weight percentage allowed the combination of data obtained in male and female children (6,7,8,10). Standards for median weight and height according to age and sex were obtained from Flügel et al. (22) and determined from cross-sectional studies conducted between 1979 and 1982 in East Germany (38,000 boys, 35,000 girls). Relative body weight >110% was defined as overweight (10,23,24).

At 2 years of age, OGTTs were performed in the morning after a 12-h overnight fast. During the test, children remained prone, resting in bed. Capillary blood samples were collected in the fasting state and then 120 min after ingesting an oral glucose load of 1.75 g glucose/kg body wt in a 40% glucose solution. Blood glucose concentration was determined using the glucosidase-peroxidase method. Impaired glucose tolerance (IGT) was defined according to the National Diabetes Data group criteria recommended for children: fasting glucose in capillary whole blood <6.7 mmol/l and 120-min glucose concentration >6.7 mmol/l (25).

In all cases, informed consent for data evaluation was given. All procedures were in accordance with the local ethical standards and with the Helsinki Declaration of 1975, as revised in 1983.

Data are expressed as means ± SEM and n. To analyze relationships between the volume of different types of milk consumed neonatally and later outcome (relative body weight and glucose tolerance), the mean volume of DBM and BBM consumed during the first 7 days of life were divided into tertiles. Differences between tertiles were analyzed using one-way analysis of variance (followed by post hoc Student’s t test) and χ² statistics. For analysis of relationships between continuous variables, univariate linear regression analysis was performed. To estimate the risk for becoming overweight and developing IGT in relation to the volume of different types of milk consumed neonatally, the odds ratio (OR) with 95% CI were calculated using logistic regression analysis. Stepwise regression analysis was used to estimate the relative impact of different independent variables on outcome variables. For all evaluations, P <0.05 was considered significant. All statistical tests were accomplished using the SPSS for Windows version 9.0 statistical software package (SPSS, Munich, Germany).

Of the 112 children studied, 83 were born to women with type 1 diabetes and 29 were born to women with GDM. The mean age at time of follow-up was 2.1 ± 0.1 years. In women with type 1 diabetes, the mean duration of diabetes before pregnancy was 9.6 ± 0.6 years. Macrosomia (birth weight >4,500 g) occurred in 6% (5 of 83) of infants of mothers with type 1 diabetes and in none of the newborns of women with GDM. Study population characteristics are shown in Table 1.

As shown in Table 2, mean relative body weight during childhood differed significantly by tertiles of DBM volume ingested neonatally. Infants in the highest tertile of intake of breast milk from their biological (diabetic) mother had a higher relative body weight at 2 years of age, as compared with those in the first tertile (P = 0.001) (Fig. 1).

This resulted in a significantly higher prevalence of overweight children in the third tertile of distribution of DBM, as compared with those in the first tertile (P = 0.03). There was no significant relationship between the tertile of neonatal DBM intake and fasting or 2-h blood glucose concentration, or the prevalence of IGT in childhood.

Conversely, the volume of BBM ingested during neonatal life was inversely related to relative body weight at follow-up (Table 3): Infants in the third tertile of distribution of BBM had a lower body weight, as compared with those in the first tertile (P = 0.004). The rate of overweight children and mean blood glucose concentration at 0 and 120 min during OGTT in childhood did not differ significantly between tertiles of BBM distribution. Infants in the second and third tertile of neonatal BBM intake distribution had, however, a lower prevalence of IGT, as compared with those in the first tertile (P = 0.03, P = 0.04).

Univariate linear regression analysis showed that the volume of DBM ingested during the first week of life was positively correlated with relative body weight during infancy (P < 0.001) (Table 4). Conversely, volume of BBM ingested neonatally was inversely correlated with relative body weight in childhood (P = 0.001). By logistic regression analysis, risk for overweight at 2 years of age was significantly higher in children in the third versus first tertiles of DBM intake distribution (OR 1.91, 95% CI 1.10–3.30). Increased risk for overweight in infants who received more DBM during neonatal life was still observed after adjusting for birth weight, gestational age, sex, age, type of maternal diabetes, and maternal BMI (OR 2.59, 95% CI 1.32–5.04).

By univariate linear regression analysis, volume of DBM ingested during the first week of life was positively correlated with blood glucose concentration at 120 min during OGTT (P = 0.03) (Table 4). Conversely, volume of BBM ingested neonatally was inversely correlated with blood glucose concentration at 120 min during OGTT (P = 0.02). Relative body weight during childhood was positively correlated with 120-min blood glucose concentration during OGTT (P = 0.01). By logistic regression analysis, risk for IGT at 2 years of age was significantly decreased in children in the third tertile of BBM intake distribution, as compared with the first tertile (OR 0.27, 95% CI 0.10–0.77). Again, adjustment for the confounders, plus additional adjustment for relative body weight at follow-up, did not change this result (OR 0.18, 95% CI 0.05–0.68).

Finally, to analyze the independent influence of neonatal nutrition on body weight and glucose tolerance in childhood, stepwise regression models were developed using relative body weight and blood glucose concentration at 120 min during OGTT in childhood as dependent variables. Birth weight, gestational age, sex, age, type of maternal diabetes, maternal BMI, volume of DBM, and volume of BBM were entered as independent variables into a stepwise regression analysis with relative body weight during childhood as the dependent variable. As shown in Table 5, the volume of DBM ingested neonatally was the only significant predictor of relative body weight during childhood, whereas all other variables did not improve the fit. In a second stepwise regression model, blood glucose concentration at 120 min during OGTT was used as a dependent variable. Birth weight, gestational age, sex, age, type of maternal diabetes, maternal BMI, volume of DBM, volume of BBM, and relative body weight during childhood served as independent variables. Table 5 shows that, by stepwise regression analysis, relative body weight at follow-up was the only significant predictor of 120 min blood glucose concentration at 2 years of age.

In summary, the results of our study indicate that in children of mothers with diabetes during pregnancy, the early neonatal ingestion of breast milk provided by their biological (diabetic) mother may result in an increased relative body weight and increased prevalence of obesity at 2 years of age, independent of birth weight, gestational age, sex, age, type of maternal diabetes, or maternal BMI. In contrast, neonatal intake of banked breast milk from healthy nondiabetic women has a beneficial effect on later body weight and glucose tolerance in childhood. Although the volume of diabetic breast milk ingested neonatally significantly predicted relative body weight at 2 years of age, body weight itself was the only significant predictor of glucose tolerance at follow-up, as revealed by stepwise regression analysis.

To our knowledge, only one study has investigated the impact of the amount of certain types of milk ingested neonatally on later outcome, in general. Singhal et al. (26) observed a beneficial effect of the amount of banked donor breast milk ingested during neonatal life, as compared with formula milk, on blood pressure later in the childhood of preterm newborns. In a number of studies, decreased body weight, diminished risk for obesity, and decreased risk for development of type 2 diabetes were observed in breast-fed infants (1,2,3,4,5,6). Moreover, by analyzing potential influences of a wide range of social and economic factors, as well as those of nutrition during later infancy, lasting effects from the type of neonatal nutrition were described in a large study sample (5). However, all these studies were performed in healthy (i.e., nondiabetic) women and their infants. We could confirm here a protective effect of feeding breast milk from healthy mothers on later risk to become overweight and to develop IGT by a negative correlation between the volume of BBM ingested neonatally and body weight and glucose tolerance during childhood. Unfortunately, no data were available on nondiabetic control subjects followed in the same way to verify that, in this population, breast-feeding was associated with reduced risk for obesity.

Until now, only one study reflected a possible long-term influence of breast-feeding in diabetic women for the development of glucose tolerance in offspring. In the Pima Indian population, a beneficial effect of breast milk was suggested; however, no distinct data analysis was provided (27). In our study, we found an unexpected strong positive correlation between the neonatal intake of diabetic breast milk and later body weight. By adjusting for a variety of maternal and infant confounders, such as type of maternal diabetes, birth weight, gestational age, and maternal BMI, we were unable to detect any other main factor that could contribute to this relationship. Moreover, our data indicate that long-term effects of the type of early nutrition affect offspring of GDM mothers as much as offspring of type 1 diabetic mothers. For determination of milk intake, we used a weighing procedure, as developed by other investigators (4,21). Values we obtained are in good agreement with those reported in another study on milk intake of offspring of diabetic mothers (28), further supporting the validity of our data.

The question arises regarding what biological mechanisms, or what component of breast milk, could be responsible for the relationship between neonatal intake of diabetic breast milk and increased body weight in childhood. As already pointed out, some studies show that breast milk from diabetic women has an increased energy content (15). Furthermore, because glucose concentrations in breast milk were shown to follow the concentration of blood glucose proportionally (17,29), increased milk glucose concentrations can be expected to occur in type 1 diabetic women and also in mothers with GDM, who usually do not show complete normalization of glucose metabolism immediately after pregnancy (30,31). Indeed, breast milk of diabetic women was reported to be characterized by increased glucose concentrations, particularly when colostrum was analyzed (16,18,19). Furthermore, breast milk from diabetic mothers was shown to contain several-fold increased concentrations of insulin (17), which is able to cross the intestinal mucose-blood barrier during neonatal life (32). Therefore, it is possible that early in neonatal life, increased levels of intact insulin are absorbed from the breast milk of diabetic mothers. Drinking a glucose- and insulin-laden milk so early in life might, therefore, promote accelerated weight gain. However, pathophysiological mechanisms remain to be established. However, a number of studies suggest that in utero exposure to increased concentrations of glucose and insulin leads to an increased risk of developing obesity and insulin resistance later in life (8,10,11). Most recently, Singhal et al. (26) speculated that trophic factors and hormones occurring in breast milk might be responsible for the long-term effects of breast-feeding found in several studies (1,2,3,4,5,6).

To our knowledge, this study is the first one describing lasting effects of the volumes of different types of milk ingested during the first week of life on later development in humans. In animals, effects of colostrum feeding, even during the first few hours of postnatal life, were reported on a variety of parameters (33,34). More specifically, a number of animal studies support the hypothesis that early neonatal nutrition “programs” the later development of body weight and glucose metabolism, e.g., artificial reduction of the number of suckling pups in lactating rats (“small litters”) leads to increased energy content and glucose concentrations in the dam’ s milk (35). During later life, rats raised in small litters display persisting obesity and impaired glucose tolerance (36,37). Moreover, early postnatal overnutrition was shown to enhance a preexisting genetic disposition, thereby leading to the development of infantile obesity (38). Recently, it was shown that in backcrossed New Zealand Obese mice, a “genetic” animal model for type 2 diabetes, the occurrence of obesity and diabetic metabolic alterations is partly explained by neonatal ingestion of milk from their biological (diabetic) mothers and can be prevented, at least in part, by cross-fostering the offspring on nondiabetic dams (39).

Currently, diabetic women are encouraged and choose to breast-feed their newborns as often as healthy women do (40). Our findings, assuming that they can be extrapolated to other populations, might offer a dilemma to women with diabetes and their babies. Therefore, before any practical conclusions should be drawn, additional studies on the long-term consequences of the type of neonatal nutrition in offspring of diabetic mothers are needed. Meanwhile, considering the variety of advantages resulting from breast-feeding in general, in our opinion, breast-feeding should remain the preferred type of infant feeding, even in this particular population. Furthermore, in most parts of the world, banked breast milk will not be readily available and, therefore, would not offer an alternative neonatal nutrition in offspring of diabetic mothers. Consequently, future studies including data evaluation on effects of formula milk will be necessary to judge what kind of neonatal nutrition might offer a safe and practical alternative for offspring of diabetic mothers. Finally, a study of banked breast milk versus commercial milk among offspring of both diabetic and nondiabetic women is suggested to determine whether it is consumption of milk from bottle per se that leads to obesity or if banked breast milk is truly less fattening.

In conclusion, because diabetes during pregnancy is increasingly realized to be a major health problem, affecting a large and continuously increasing number of pregnant women, our data underscore the urgent need for additional studies of this important at-risk population to assess long-term consequences from the type of neonatal nutrition. Thereby, new approaches for genuine prevention of lasting adverse effects on the health outcome in offspring of diabetic mothers may be achieved.

This study was supported by the German Diabetes Association (DDG) and by the German Research Foundation (DFG, PL 241–1/3).

We thank Siri L. Kjos, MD, for helpful comments and linguistic revision of the manuscript.