Does malnutrition in utero determine diabetes and coronary heart disease in adulthood? Results from the Leningrad siege study, a cross sectional studyBMJ 1997; 315 doi: http://dx.doi.org/10.1136/bmj.315.7119.1342 (Published 22 November 1997) Cite this as: BMJ 1997;315:1342
- S A Stanner, project coordinator ()1,
- K Bulmer, research techniciana,
- C Andrès, research techniciana,
- O E Lantseva, endocrinologistb,
- V Borodina, biologistb,
- V V Poteen, professor of medicinea,
- J S Yudkin, professor of medicineb
- a Department of Medicine, University College London Medical School, Whittington Hospital, London N19 3UA
- b Ott Institute of Obstetrics and Gynaecology, Russian Academy of Medical Science, St Petersburg, Russia
- Correspondence to: Ms Stanner
- Accepted 26 August 1997
Objective: To investigate the relation between decreased maternal food intake and risk factors for coronary heart disease in adult
Design: Cross sectional study.
Subjects: 169 subjects exposed to malnutrition in utero (intrauterine group) during the siege of Leningrad (now St Petersburg) in 1941-4; 192 subjects born in Leningrad just before rationing began, before the siege (infant group); and 188 subjects born concurrently with the first two groups but outside the area of the siege (unexposed group).
Setting: Ott Institute of Obstetrics and Gynaecology, St Petersburg.
Main outcome measures: Development of risk factors for coronary heart disease and diabetes mellitus—obesity, blood pressure, glucose tolerance, insulin concentrations, lipids, albumin excretion rate, and clotting factors.
Results: There was no difference between the subjects exposed to starvation in utero and those starved during infant life in: (a) glucose tolerance (mean fasting glucose: intrauterine group 5.2(95% confidence interval 5.1 to 5.3), infant group 5.3 (5.1 to 5.5), P=0.94; mean 2 hour glucose: intrauterine group 6.1 (5.8 to 6.4), infant group 6.0 (5.7 to 6.3), P=0.99);(b) insulin concentration; (c) blood pressure; (d)lipid concentration; or (e) coagulation factors. Concentrations of von Willebrand factor were raised in the intrauterine group (156.5 (79.1 to 309.5)) compared with the infant group (127.6 (63.9 to 254.8); P<0.001), and female subjects in the intrauterine group had a stronger interaction between obesity and both systolic (P=0.01) and diastolic (P=0.04) blood pressure than in the infant group. Short adult stature was associated with raised concentrations of glucose and insulin 2 hours after a glucose load—independently of siege exposure. Subjects in the unexposed group had non-systematic differences in subscapular to triceps skinfold ratio, diastolic blood pressure, and clotting factors compared with the exposed groups.
Conclusions:Intrauterine malnutrition was not associated with glucose intolerance, dyslipidaemia, hypertension, or cardiovascular disease in adulthood. Subjects exposed to malnutrition showed evidence of endothelial dysfunction and a stronger influence of obesity on blood pressure.
Relations between intrauterine growth and adult disease such as diabetes and cardiovascular disease have been linked to poor nutrition during pregnancy
In this study, however, intrauterine exposure to malnutrition was not associated with glucose intolerance
Intrauterine malnutrition did not affect insulin concentration, blood pressure, or concentration of lipids or coagulation factors
Concentration of von Willebrand factor, a marker of endothelial damage, was raised in the subjects exposed to intrauterine malnutrition
Obesity and blood pressure were more strongly related in subjects exposed to intrauterine malnutrition than in subjects either unexposed to malnutrition or exposed to malnutrition only as infants
Several reports have shown that low birth weight is associated with diabetes and hypertension in adult life.1 2 3 4 5 6 78 910 Other studies have reported that thin babies develop insulin resistance in adulthood.1112 These observations have led to thehypothesis that growth retardation affecting the development and vascularisation of particular organs at different stages of fetal development will predispose the individual to impaired organ function, with consequent disease in later life.13
Many of these reports have assumed that early growth retardation is synonymous with fetal malnutrition and that this in turn is related to maternal supply of nutrients during pregnancy. A “thrifty phenotype” hypothesis has been proposed to suggest that many of the diseases of Western civilisation, which occur in epidemic proportions when populations move from malnutrition to a Western lifestyle, may be the result of programming of the metabolism and function of a tissue or organ as a result of diminished supply of certain nutrients during critical stages of development.14 15 Animal studies have shown that intrauterine protein deficiency is associated with impaired pancreatic ß cell function16 17 18 and increased blood pressure19 in later life. Nevertheless the relation between maternal nutrient intake and either the birth weight of offspring20 or subsequent disease21 in humans remains poorly documented.
The siege of Leningrad (the German blockade of the city now known as St Petersburg) between 8 September 1941 and 27 January 1944 prevented supplies from reaching the city for 872 days. Of a population of 2.4 million, between 750 000 and one million people died, mostly from starvation. Most of these deaths occurred during the “hunger winter” of November 1941 to February 1942, when the siege was in full force and the bread ration of 250 g for workers and 125 g for others was all that was available.22 The average daily ration for most of the citizens of Leningrad during this time therefore provided around 300 calories and contained virtually no protein. Although the situation improved when Lake Ladoga froze sufficiently to allow supplies to be transported across, it was April/May 1942 before food supplies increased substantially. Average male and female birth weights fell by 18% and 16% respectively.23
We investigated the relation between decreased maternal food intake and risk factors for coronary heart disease in adult life, with specific reference to malnutrition in utero during the siege of Leningrad.
Subjects and methods
Subjects and study design
We investigated two groups of subjects exposed to the siege of Leningrad. These two groups were identified from the register of the Society of Children of the Siege, which maintains a complete and updated record of all people living in, or born in, the city of Leningrad during the siege. Subjects exposed to the siege in utero (intrauterine group) comprised adults born in the city of Leningrad between 1 November 1941 and 30 June 1942, the first date being 54 days after the start of the siege. Subjects exposed to the siege as infants (infant group)comprised adults also born in the city of Leningrad but between 1 January and 30 June 1941, and consequently they were at least 10 weeks old at the beginning of the siege. Historical records document the plentiful supply of food in Leningrad until rationing was imposed on 18 July 1941 in preparation for the impending siege.
We identified 1229 subjects (548 male) born between 1 January 1941 and 30 June 1942. Four of these subjects had died and 209 were not contactable or had changed address, leaving 1016 (464 male) available for invitation to the study. Of these, 443 (44%) subjects attended for screening (125 male (27% of those contacted), 318 female (58%)), of whom 10 were excluded as known diabetics and 72 were excluded either because their birth date fell between 1 July 1941 and 31 October 1941 (n=51) or because, owing to inaccurate entry on the register, the date fell outside the limits of the study group (n=21). This left a population of 169 (37 male)in the intrauterine group and 192 (62 male) in the infant group. Of the subjects in these two groups, 115/167 (69%) and 129/192 (67%) respectively remained in Leningrad until the siege ended, with most of the evacuations occurring after July 1942. Thus all the subjects in the intrauterine group would have been additionally exposed to siege in infancy.
In addition to the two groups exposed to the siege, we studied a third group. This comprised 188 adults (50 male) who were born in the province of Leningrad but outside the city (and thus the siege limits) during the same period as subjects in the other two groups (1 January 1941 to 30 June 1942). The subjects in this group (unexposed group) were invited from two sources—the radial kerotomy clinic of the local hospital, where patients had been referred for surgery for refractive eye problems, and six local workplaces. Subjects with known diabetes, glaucoma, or hypertensive or diabetic retinopathy were excluded (n=7); 102 subjects (24 male) from the kerotomy clinic attended, and 86 subjects (26 male) from the local workplaces attended. The two subgroups were not found to be significantly different in any of the criteria studied (data not shown)and were therefore combined.
Subjects were invited to attend the department of endocrinology at the Ott Institute, St Petersburg, the morning after an overnight fast. Subjects were weighed in light clothing and without shoes on a beam balance (Seca, Birmingham) and height recorded on a stadiometer (Pribordetal Plant, Zuevo, Russia). The waist to hip ratio and the subscapular to triceps skinfold ratio were calculated.24 Blood pressure was measured in triplicate with a random zero sphygmomanometer (Hawksley Gelman, Lancing, Sussex) that had been calibrated against a similar machine in the department of medicine at the University College London Medical School, Whittington Hospital, London. A resting 12 lead electrocardiogram was recorded. Personal and family medical histories were taken to ascertain previous health and experience of the siege, and a standardised questionnaire was administered for ischaemic heart disease.25 Smoking and alcohol histories were also obtained.
Blood sampling and biochemical analyses
A venous blood sample was taken after an overnight fast for several measures (see below), and each patient was asked to void his or her bladder. Each patient was then given 75 g of anhydrous glucose (Fortical, Cow and Gate, Trowbridge, Wiltshire)dissolved in 300 ml water to drink over 5 minutes; 30 minutes and 120 minutes later, blood samples were taken for measurement of glucose and insulin concentrations. The patient passed urine again after the glucose tolerance test, and the time was recorded. Each subject was then given a urine collection bottle to make a timed overnight urine collection. Plasma glucose concentration was measured in samples taken after fasting and after 30 and 120 minutes with a Beckman analyser (Beckman Instruments UK, High Wycombe, Buckinghamshire). Fifty four per cent of the samples were also assayed in the department of medicine at the University College London Medical School, with good agreement between the two assays (mean difference 0.12 (SE 0.03) mmol/l, P=0.001).
Total and high density lipoprotein cholesterol were measured using the enzymatic colorimetric method (Sigma Diagnostics, Poole, Dorset), the latter after precipitation of triglyceride-rich lipoproteins with heparin and manganese. Triglyceride concentrations were also assayed with a Sigma kit (Sigma Diagnostics), and low density lipoprotein cholesterol concentration was calculated with the Friedewald formula.26 Plasma insulin concentrations were measured with a commercial enzyme linked immunosorbent assay (ELISA) (Dako Diagnostics, Ely, Cambridgeshire), which is specific for insulin without cross reaction with either intact or des 31, 32 proinsulin.27 Intact and des 31, 32 proinsulin, fibrinogen, factor VII and urinary albumin were assayed as previously described.28 Fasting C peptide concentration was measured with a radioimmunoassay kit for human C peptide (Biodata, Rome, Italy); plasminogen activator inhibitor was measured both as activity and as antigen (Biopool, Bio-Stat, Stockport, Cheshire); and plasma concentration of von Willebrand factor was measured with an in house enzyme linked immunosorbent assay by using antibodies from Dako (Copenhagen, Denmark)and validated by using controls from the National Institute for Biological Standards and Controls. All assays were performed in a single run, except for plasminogen activator inhibitor, factor VII, and fibrinogen, for which all statistical analyses were adjusted for analytical batch.
Classification of subjects
Because of the absence of satisfactory Russian criteria for classifying social class, subjects were classified according to manual or non-manual occupation and the Office of Population Censuses and Surveys' classification of occupations.29 Glucose intolerance was classified as impaired glucose tolerance or diabetes mellitus, according to the World Health Organisation's criteria.30 The electrocardiographic results were coded in Britain according to Minnesota code criteria,25 and subjects were classified in three groups. Group 1 comprised subjects with confirmed myocardial infarction according to codes 1.1, 1.2, or 7.1 for electrocardiographic changes. Group 2 comprised subjects with definite or possible ischaemia (all subjects with the changes that defined group 1, and subjects fitting one or more of the codes 1.3, 4.1-4.4, 5.1-5.3). Group 3 comprised all subjects with the abnormalities that defined groups 1 and 2 or with a history of angina (on a World Health Organisation questionnaire). Microalbuminuria was defined as an albumin excretion rate of 20–200 μg/min on either a two hour sample or an overnight sample. Hypertension was defined as a systolic blood pressure >140 mm Hg, diastolic blood pressure >90 mm Hg, or treatment for hypertension. The percentage of proinsulin-like molecules was calculated as:100x(proinsulin+des 31, 32 proinsulin)/(insulin+proinsulin+des 31, 32 proinsulin).
We compared the distribution of variables in the intrauterine group and the infant group, as the source of the subjects in these groups was identical and most of the subjects had remained in Leningrad throughout the siege. We used analysis of variance, with logarithmic transformation for skewed variables and adjustment for sex. Categoric variables were compared by using the χ2 test in a similar fashion. We used correlation and linear regression analysis to study associations between continuous variables, again with logarithmic transformation of skewed data. Analysis of covariance was used to test the homogeneity of regression slope of blood pressure and body mass index (weight(kg)/height(m)2))for these two groups. All analyses were repeated across all three study groups, with the same adjustment for sex.A P value of <0.05 was taken as significant, although, where multiple comparisons have been performed, it would be appropriate to use more rigorous criteria for significance.
From the variance of these measures in an age matched population in London31 we calculated that 200 subjects in each group would permit the detection of a 0.22 mmol/l difference in fasting plasma glucose and 6.1 mm Hg difference in systolic blood pressure between groups at the 1% level with a power of 90%.
Intrauterine and infant groups
The subjects in the intrauterine group and the infant group were similar for parents' nationality, percentage of subjects in employment, social class, and years of education. In the intrauterine group 21% (26/122) of subjects recalled the loss of a first degree relative during the siege, compared with 24% (36/151) of those exposed during infancy (P=0.52); the percentages of subjects recalling mothers and siblings lost through starvation or illness were 4% (5/122) and 5% (7/151) respectively (P=0.83). Table 1 shows the characteristics of the subjects studied. The intrauterine and infant groups did not differ in height or in degree of central obesity (determined by waist to hip ratio), regardless of whether men and women were analysed together or separately. Blood pressure and the prevalence of coronary heart disease were similar for the two groups, although a significantly higher proportion of the infant group reported symptoms of angina. Smoking rates and alcohol intake were similar for all groups (data not shown).
There was no difference in concentrations of fasting and two hour plasma glucose during an oral glucose tolerance test (table 2) or any excess of known diabetes or glucose intolerance associated with the intrauterine or infant group. These conclusions were unaffected by using assay results from London rather than from St Petersburg. Levels of insulin-like molecules, lipid concentrations, and daytime albumin excretion rate did not differ, but the intrauterine group had a significantly lower overnight albumin excretion rate (table 3.
Concentrations of fibrinogen, factor VII, and plasminogen activator inhibitor (activity and antigen) were similar for the intrauterine and infant groups (table 4), but subjects in the intrauterine group had a significantly higher concentration of von Willebrand factor. This remained significant after further adjustment for obesity, smoking, and current coronary heart disease (P=0.009).
The subjects in the unexposed group were similar for all lifestyle and demographic factors studied. They had marginally lower diastolic blood pressure than the exposed subjects (P<0.05). They also had significantly higher concentrations of factor VII but lower concentrations of plasminogen activator inhibitor antigen and activity.
Leon et al have suggested that adult blood pressure is more markedly affected by obesity in individuals with intrauterine growth retardation than in those without growth retardation.9 We therefore explored the relation between exposure to the siege and adult obesity. In all subjects combined, blood pressure was positively related to obesity (systolic blood pressure partial r=0.29, P<0.001, with sex, age, and exposure controlled for). However, the relation in women between body mass index and (a) systolic blood pressure (1) and (b)diastolic blood pressure was significantly stronger in the intrauterine group than in the infant group (P=0.01 and P=0.04 respectively). This suggests that siege exposure and adult obesity may act synergistically to increase susceptibility to hypertension.
We explored the relation between adult height and glucose intolerance, hypertension, and microalbuminuria. Glucose intolerant subjects were shorter than those with normal glucose tolerance, with a mean difference of 21 mm (P=0.02) after adjustment for age, sex, and exposure to siege. This difference did not persist after further adjustment for adult social class and education (adjusted mean difference 13 mm, analysis of variance P=0.935, with age, sex, and exposure to siege controlled for). The partial correlation coefficients of height were r=-0.10, P=0.019 with two hour plasma glucose, and r=-0.13, P=0.002 with insulin two hours after glucose challenge, after adjustment for the same covariates. The height of subjects with hypertension or microalbuminuria did not differ significantly from the height of subjects without.
We studied 549 subjects born in and around Leningrad at the time of the siege, one third of whom are likely to have been exposed to severe intrauterine starvation. This population has allowed us to explore the relation between maternal malnutrition and cardiovascular risk factors, which otherwise have been predominantly studied in vitro or in animal models. We did not find evidence for the hypothesis suggesting that intrauterine exposure to maternal malnutrition was linked with glucose intolerance, dyslipidaemia, hypertension, or cardiovascular disease in adulthood. Although these observations suggest that increased levels of cardiovascular risk factors are not consequent on maternal nutritional intake, endothelial dysfunction and a stronger relation between obesity and blood pressure were more common in the subjects who were exposed to the siege in utero.
Effect of losses to follow up
We studied 44% of eligible subjects (those in the exposed groups who were invited for screening). The nature of studies relating to fetal origin inevitably means that follow up is less complete than for the usual sort of epidemiological study. Low ascertainment rates (as low as 5.8% in some studies32) have been criticised,33 34 but Martyn et al have argued that as the analyses are based on internal comparisons, selection bias would arise only if the associations of early life events and coronary heart disease were different between those studied and those lost to follow up.35 In our study the only characteristic that distinguished the intrauterine group from the infant group was a date of birth before July or after November 1941. This thus determined whether they were exposed to the siege conditions only as an infant or additionally as a fetus, the two siege groups being selected from the same register with similar attendance rates. If programming of adult metabolism is consequent on exposure to malnutrition during certain stages of organ development in utero, substantial differences would be expected between these two groups.
Lack of reliable data on birth weight
We could not obtain reliable data on birth weight. Twenty two per cent (120) of the subjects provided a recalled birth weight, the intrauterine group having a significantly lower mean birth weight (2.7 (SD 0.9) kg v infancy group 3.4 (SD 1.1)kg vunexposed group, 3.6 (SD 1.0) kg; analysis of variance, with sex controlled for, P<0.001). This is consistent with the birth weights recorded during this period of the siege.23 This study used exposure to siege, rather than birth weight, as a surrogate marker of maternal nutrient intake. Historical records show that maternal diet in pregnancy was likely to have been good for the infant group, who were aged at least 10 weeks when the siege began, with food supply being plentiful until rationing began. We found no evidence that subjects born during the siege had parents in occupations with preferential access to food.
Other possible explanations for our negative findings include the possibility that, as observed by Godfrey et al in human studies,20and as recognised in sheep,36 the effect of protein deprivation on fetal development is more marked in offspring of women with a high carbohydrate intake in pregnancy, a situation likely to prevail in malnourished populations of developing countries. Another interpretation is that malnutrition may be necessary for more prolonged periods or for more than one generation before the relations between growth retardation and adult disease are seen.35 Studies have shown powerful intergenerational effects on birth weight—largely transmitted through the maternal line37—which seem to apply even when women migrate from communities with poor nutrition to an environment with an improved food supply.
Effect of malnutrition on adult height
Short stature has been related both to incidence of coronary heart disease38 and to glucose intolerance.39 We confirmed the relation between short stature and glucose intolerance, this association being independent of study group. Adult height is determined both by birth weight and by childhood environment. We found no difference in height between the study groups, and the relation between height and glucose intolerance seems to be explained partly by occupational grade and education, which we have used as indicators of later social deprivation. Studies of men born during or after the hunger winter in the Netherlands show that height was unaffected by exposure to famine conditions in utero, indicating the potential for “catch up” growth if later food supply is adequate.40 We did not find in this population the described associations between microalbuminuria31 or blood pressure41 and adult height.
We also studied subjects born outside siege conditions, who differed in several variables from the two siege groups. However, these differences may relate to different selection criteria and not to different maternal nutritional intakes.
We are grateful to Professor D J P Barker for help in seeking financial support for this project and discussions about the design of the study, to Drs John and Simon Griffin for helping with assays in St Petersburg, and Stephanie Goubet and Dr David Leon for advice on analysis of our data. We are indebted to the staff of the Ott Institute who helped with subject recruitment and screening and are particularly grateful to Elena Yablochkina for all her help in translating questionnaires, organising trips, and coordinating this project in St Petersburg.
Funding: This study was supported by grants from the British Diabetic Association and the International Association for the Promotion of Cooperation with Scientists from the Former Soviet Union.
Conflict of interest: None.