Microalbuminuria in Adults after Prenatal Exposure to the Dutch Famine

Materials and Methods

All participants were part of a cohort of 2414 liveborn term singletons, who were born at the Wilhelmina Gasthuis in Amsterdam between November 1, 1943, and February 28, 1947. The selection procedure for this clinical study has been described in detail elsewhere (25). The local Medical Ethics Committee approved the study, and all participants gave informed consent. The study was carried out in accordance with the declaration of Helsinki. A urine sample was collected from 724 of the 741 people who visited the clinic. Their mean age at clinic visit was 50 yr (range 48 to 53 yr). Their mean birth weight (3352 g) did not differ significantly from the 1690 who were not included (mean 3343 g).

We defined the famine period as the time that the official adult food rations were under 1000 calories per day: November 26, 1944, to May 12, 1945. Although supplementation of the official food ration by church organizations, the black market, and central kitchens was possible and indeed total food intake may have been up to twice as high as the official rations indicate, the official rations do adequately reflect the variation over time of total food availability throughout the famine period (28).

We considered fetuses to have been exposed to famine when the average daily rations for adults during any 13-wk period of gestation were <1000 calories. Therefore, infants who were born between January 7, 1945, and December 8, 1945, were exposed to famine in utero. Three 16-wk periods were used to differentiate between people who were exposed in late gestation (born between January 7 and April 28, 1945), in mid gestation (born between April 29 and August 18, 1945), and in early gestation (born between August 19 and December 8, 1945). Children who were younger than 1 yr were relatively protected during the famine, because their official daily rations never fell below 1000 calories, and the specific nutritional components were always above the standards used by the Oxford Nutritional Survey (29). As in all previous publications from this study (25–27,30–33), the nonexposed group comprises those who were born before the famine and those who were conceived after the famine. The medical birth records provided information about the mother, the course of the pregnancy, and the size of the infant at birth (birth weight, length, ponderal index, and head circumference).

An overnight urine sample was collected from 10 p.m. the evening before the clinic visit. Plasma and urinary creatinine were measured. Urinary MA was assessed using immunonephelometry. An albumin/creatinine ratio (ACR) was calculated by dividing urinary microalbumin (mg/L) by urinary creatinine (mmol/L). MA was defined as an ACR ≥2.5 g/mol, according to the hospital laboratory’s guidelines. Creatinine clearance was estimated using the Cockcroft-Gault formula (34): The validity of the Cockcroft-Gault formula in the general population has been confirmed in a number of studies (34–36).

During the clinic visit, a standard oral glucose tolerance test (25) was performed, an electrocardiogram (ECG) was made, weight and height were measured, BP was measured using an automated auscultatory device (Profilomat; Disentronic Medical Systems AG, Burgdorf, Switzerland) (37), and serum cholesterol was measured. A cardiologist, blinded to clinical data, examined the ECG and marked them either (borderline) normal or abnormal. Diabetes or impaired glucose tolerance (IGT) was defined as either self-reported diabetes (type 1 and type 2) or a 2-h blood glucose >7.8 mmol/L (25).

Health, medication use, lifestyle, and socioeconomic data had been recorded during a home interview. People who were receiving hypertension treatment were not excluded from analyses. Socioeconomic status (SES) was determined from the participant’s or the partner’s occupation, whichever was highest, according to the socioeconomic index (ISEI-92) scale ranging from 16 for the lowest to 87 for the highest status.

Body mass index (BMI) and 2-h glucose concentrations were log transformed because they had skewed distributions; therefore, results are reported as geometric means and SD. We were not able to transform the skewed ACR distribution, so ACR results are reported using the dichotomized MA definition only.

Results

A total of 288 (40%) of the 724 participants studied had been exposed to famine in utero (Table 1). Weight at the last antenatal clinic visit was lower in mothers of participants who were exposed to famine during late and mid gestation than in mothers of nonexposed individuals. Infants who were exposed to famine during mid or late gestation were lighter and shorter and had smaller placentas than infants who were not exposed to famine in utero. Adult BMI was higher among those who were exposed early in gestation (P = 0.04).

In total, 52 (7.2%) people had MA (Table 2). People with MA had a higher BMI, had a lower SES, more often were smokers, were older, had higher diastolic and systolic BP, and were more likely to have IGT or diabetes and ECG abnormalities than those without MA. The prevalence of MA did not differ between men (7.3%) and women (7.1%; P = 0.89). People with MA tended to have been smaller at birth than people without MA (3269 and 3358 g; P = 0.2, corrected for gender).

The prevalence of MA was significantly higher in those who were exposed to famine in mid gestation (12%) than in people who were not exposed prenatally (7%; P = 0.05). The prevalence of MA in those who were exposed in early (9%; P = 0.3) or late gestation (7%; P = 0.8) was not significantly increased.

The effect of exposure to famine in mid gestation was independent of maternal weight (adjusted OR 2.1; 95% CI 1.0 to 4.4), birth weight (adjusted OR 1.9; 95% CI 0.9 to 4.0), and other measures of size at birth. None of the parameters of size at birth were themselves associated with MA.

After adjusting for variables that are known to influence MA, including BMI, gender, age, smoking, IGT or diabetes, systolic BP, the presence of ECG abnormalities and serum cholesterol, and adult SES in a multivariate logistic regression model, the effect of exposure to famine in mid gestation on adult MA remained significant (OR 3.2; P < 0.01). The most marked increase in the point estimate for the OR was observed when adjusting for BMI. Again, there was no significant association with exposure in early (P = 0.3) or late (P = 0.6) gestation after adjusting for these risk factors in a multivariate model (Table 1). The association between mid gestational exposure to famine and MA was independent of the use of antihypertensive medication, diagnosed hypertension, and pulse pressure. In a gender-stratified analysis, we found no significant modification of the effect of famine exposure on MA prevalence.

We were able to estimate creatinine clearance in 667 of the 724 participants: Body weight measurements were missing in 48, and plasma creatinine was missing in 9 participants. The overall mean creatinine clearance was 115 ml/min in women and 124 ml/min in men. Lower SES, higher fasting glucose, higher systolic BP, and higher adult weight were associated with higher creatinine clearance.

The mean creatinine clearance of individuals who were born before the famine was significantly lower than the mean clearance in the group that was conceived and born after the famine (age- and gender-adjusted P < 0.001). We therefore used the conceived after group as the control group for analyses of creatinine clearance, comparing creatinine clearance per exposure group adjusting for confounders using linear regression.

Those who were conceived after the famine had the highest mean clearance. In comparison, those who were born before the famine had a gender- and age-adjusted decrease of 16.4 ml/min (95% CI 7.4 to 25.4; P < 0.01), whereas people who were exposed to famine in mid gestation had a decrease of 11.9 ml/min (95% CI 4.0 to 19.8; P < 0.01). Adjusting for confounders such as adult weight, systolic BP, fasting glucose, adult SES, as well as for measures of size at birth or maternal parameters, did not attenuate this association. Neither measures of size at birth nor maternal parameters were associated with mean creatinine clearance.

Discussion

We found that people who were exposed to famine in mid gestation had higher rates of MA at age 50 compared with people who had not been exposed to famine in utero. This is the first direct evidence suggesting that maternal starvation during gestation is linked to impaired renal function in the offspring. This effect was not mediated by reduced maternal weight or small size of the infant at birth. Adjusting for variables that are known to influence MA did not alter the effect.

We did not find a significant association between small birth size and MA, although people with MA tended to have been smaller at birth than people without MA. Some studies have found an association of either birth weight (6,7) or ponderal index with MA (8), whereas other studies were able to demonstrate only a trend to that effect (38,39). This inconsistency may be because small size at birth is a crude marker of fetal undernutrition (40). In female rats that were exposed to a low-protein diet for 1 wk in mid pregnancy, the nephron number of the offspring was profoundly reduced, whereas the effect on birth weight was negligible (16). This illustrates that organ-specific effects of a short period of undernutrition during gestation may be independent of effects on total body size.

Previously, we found people who were exposed to famine in late gestation to have reduced glucose tolerance (25), whereas exposure to famine in early gestation was linked to higher rates of obesity (30), a more atherogenic lipid profile (26), and cardiovascular disease (27). We now have evidence that mid gestational exposure to famine is linked to reduced renal function. This distinct relation between prenatal exposure to famine at different stages of gestation provides additional evidence that long-term effects of maternal undernutrition depend on its timing during gestation.

Our findings are consistent with the hypothesis that mid gestational undernutrition leads to reduced glomerular endowment at birth and a consequent increased risk for MA as a result of single nephron hyperfiltration (41,42). We are limited, however, by the small number of individuals in each of the exposure groups in our ability to make firm statements about the precise timing of the effects per trimester of exposure. Our conclusions are based on noninvasive outcome measures only. Renal biopsy, allowing glomerular counting and sizing, is not available in this cohort of healthy volunteers because of ethical considerations. We are limited by the small sample size in this study in drawing conclusions for the repercussions of our findings for ESRD rates. However, the literature suggests that adverse intrauterine conditions do contribute to higher ESRD prevalence (5,9).

We conclude that exposure to famine in mid gestation is linked to a 3.2-fold increase in occurrence of MA in adulthood and a 10% decrease in creatinine clearance, neither of which can be explained by cardiovascular confounders. We propose that mid gestational exposure to famine may prevent formation of sufficient glomeruli and thus increase the risk for MA and possibly affect renal function in adulthood. This supports the concept that intrauterine conditions during distinct, organ-specific periods of sensitivity may permanently determine health outcome in later life.