Fetal growth and cardiovascular risk factors in Jamaican schoolchildrenBMJ 1996; 312 doi: https://doi.org/10.1136/bmj.312.7024.156 (Published 20 January 1996) Cite this as: BMJ 1996;312:156
- Terrence E Forrester, directora,
- Rainford J Wilks, lecturera,
- Franklyn I Bennett, senior lecturera,
- Donald Simeon, lecturera,
- Clive Osmond, statisticianb,
- Mureen Allen, research fellowa,
- Andrene P Chung, research fellowa,
- Paul Scott, research fellowa
- a Tropical Metabolism Research Unit, University of the West Indies, Mona, Kingston 7, Jamaica
- b Medical Research Council Environmental Epidemiology Unit, University of Southampton, Southampton General Hospital, Southampton SO9 4XY
- Correspondence to: Dr Forrester.
- Accepted 2 November 1995
Objective: To determine relation between schoolchildren's blood pressure, glycated haemoglobin level, and cholesterol concentration and their anthropometry, socioeconomic status, and birth measurements.
Design: Retrospective cohort study.
Setting: 27 schools closest to University Hospital of the West Indies, Kingston, Jamaica.
Subjects: 2337 children aged 6-16 years who were born at university hospital were recruited, and their birth records were recovered: 1610 had suitable records, 659 had records including birth length, and 610 of these were prepubertal.
Main outcome measures: Blood pressure, glycated haemoglobin level, serum cholesterol concentration, anthropometry at birth, current anthropometry, and socioeconomic status.
Results: Multiple regression analysis showed that children's systolic blood pressure was inversely related to their birth weight (P<0.0001) and directly related to their current weight. Glycated haemoglobin level was higher in children with thicker triceps skinfolds (P<0.001) and who had been shorter at birth (P=0.003). Serum cholesterol concentration was inversely related to current height (P=0.001) and to length at birth (P=0.09) and was directly related to triceps skinfold thickness and higher socioeconomic status (P<0.001).
Conclusions: Blood pressure in childhood was inversely related to birth weight and directly to current weight. Glycaemic control and serum cholesterol were related to short length at birth, height deficit in childhood, and childhood obesity.
Hypertension and diabetes mellitus contribute substantially to cardiovascular mortality in the Caribbean. Fetal growth retardation has recently been added to the list of established risk factors for these diseases.1 2 3 4 5 6 7 8 Thus, blood pressure in adults was inversely related to birth weight, blood glucose concentration after an oral glucose challenge was inversely related to birth weight and to weight at 1 year, and serum cholesterol concentration was inversely related to abdominal circumference at birth.4 5 6 7 8
The underlying mechanisms are largely unknown. Continued exposure to risk factors associated with poor socioeconomic circumstances might explain associations between weight and proportions at birth and disease in adulthood.9 Alternatively, poor maternal nutrition might curtail fetal growth and thus affect the control of blood pressure and concentrations of blood glucose and serum lipids in later life.10 In Caribbean communities the prevalence of hypertension and diabetes is high, and both social deprivation and poor maternal nutrition are common. Describing how social deprivation and “fetal programming” affect the prevalence of hypertension and diabetes is important for public health.
Retardation of fetal growth has been linked with blood pressure in adults and children.4 6 11 12 13 However, there are few data for glycated haemoglobin and serum cholesterol concentrations in childhood.14 In this study we determined how blood pressure, glucose tolerance (glycated haemoglobin level), and serum cholesterol concentration in Jamaican schoolchildren were related to socioeconomic status, body size and composition, anthropometry at birth, and maternal blood pressure and anthropometry.
Schoolchildren born in the University Hospital of the West Indies, Kingston Jamaica, were eligible for entry into the study, and we took a sample from the 27 schools nearest to the hospital. Attendance at school for children aged 6 to 16 is mandatory in Jamaica, and actual attendance is 83.5% in Kingston. Five of the 27 schools were primary or preparatory, and 22 were secondary or high schools. Two of the five primary schools were private, but all the secondary schools were state financed.
Parents identified those children in school who were born at the hospital, and we invited children to participate in the study after we had obtained their oral consent and parental written consent. The study was approved by the medical ethics committee of the University of the West Indies.
MEASUREMENTS OF CHILDREN
The children were studied in school.
Blood pressure—Each child was seated for 10 minutes before blood pressure was measured by one of two observers with an oscillometric sphygmomanometer (Dinamap 1846SX15). Five measurements were taken a minute apart, and the fourth and fifth measurements were averaged to give the blood pressure. Middle upper arm circumference dictated cuff size.
Anthropometry—Two trained nutritionists measured each child's weight, height, middle upper arm circumference, triceps and subscapular skinfold thicknesses, and waist and hip circumferences.16
Blood tests—Venous blood was drawn after measuring blood pressure and anthropometry. Haemoglobin concentration was measured with a Coulter haemoglobinometer. Total serum cholesterol concentration and glycated haemoglobin level represented lipid status and usual blood glucose control respectively. Cholesterol concentration was measured enzymatically,17 and glycated haemoglobin was measured by affinity chromatography.18 The coefficient of variation for the cholesterol assays was 9.8% at a concentration of 4.5 mmol/l, while for the glycated haemoglobin assay it was 10.6% at 6%.
Socioeconomic status—Two observers conducted structured interviews to determine socioeconomic status while the children were seated awaiting measurement of blood pressure. Equal ratings from data on family possessions, utilities in the house, and school materials were used to compute a socioeconomic index that was directly related to wealth. As a guide to the accuracy of the responses, a random sample of 10% of the 1610 children enrolled in the study was chosen and one parent (head of household) was asked to answer the same questionnaire. Responses of parents and children were compared with scores for socioeconomic status. Each parent's score was subtracted from their child's score to obtain the difference. The mean of the differences (-1.4) provided a measure of bias, and the standard deviation (2.3) gave a measure of precision. The mean difference was not significantly different from zero.
Observers were trained and the techniques used were standardised before the study began. Two observers administered the questionnaire, and two others measured blood pressure and anthropometry. Estimates of variability within and between observers were obtained by testing 20 subjects before the study began and then again at points marking 25%, 50%, and 75% completion of the study. Bias and precision were calculated for test-retest differences. On each occasion collection of data was not resumed until measurements within and between observers had a correlation co-efficient of 0.9 or greater. For blood pressure, the Dinamap sphygmomanometer was calibrated weekly against a mercury manometer and tested for leaks every day.
INFORMATION ON FETAL DEVELOPMENT
After children were enrolled, maternal hospital records were located to obtain data on mothers' weight, height, and blood pressure throughout pregnancy and on the children's birth and placental weights, crownheel length, crown-rump length, and head circumference. While birth and placental weights were routinely recorded, lengths were not recorded for many of the older children.
Data analysis was confined to the 1610 children who were born after 28 weeks' gestation and had suitable records. Of these, 659 had complete records including measurements of head circumference and crown-heel length.
Descriptive data for the population are based on the total sample (1610). Multiple regression techniques were used to determine associations between children's current blood pressure, glycated haemoglobin level, cholesterol concentration, and anthropometry and their birth weight and placental weight, placental to birth weight ratio, and maternal anthropometry in the first trimester.
Testing for associations between current blood pressure, glycated haemoglobin level, and cholesterol concentration and anthropometry at birth (including body length and head circumference) was restricted to the 659 children with complete records. Because tracking of blood pressure and serum lipids is lost at puberty,19 20 21 which can obscure associations with birth measurements, we selected the 610 children with complete records who were prepubertal (aged </= 10 years) for these analyses.
In these analyses birth weight, placental weight, and crown-heel length were divided into groups and tested for trend in association with blood pressure, glycated haemoglobin level, and cholesterol concentration. The sample of 1610 children provided 80% power at the 95% level of significance to show a difference in blood pressure of 2 mm Hg between children in the highest and lowest quarters of the birth weight distribution (SD for blood pressure=10 mm Hg).
The 27 schools screened yielded 2983 eligible subjects, of whom 2337 were willing to take part in the study. The remaining 646 (302 boys, 344 girls) declined because of parental or personal objection. No further data were available on these 646 children. Of the 2337 children willing to take part, 1610 had sufficient records for inclusion in the study population and were born after 28 weeks' gestation (1409 were born at term, between 38 and 42 weeks' gestation).
Table 1 shows the group comparisons for the 610 children (309 girls, 301 boys) aged </=10 years and the 1000 children (570 girls, 430 boys) aged >/=11. Table 2 shows the birth measurements for boys and girls: boys were heavier than girls and had greater head circumferences, but placental weights were not different.
For the whole sample of 1610 children, systolic pressure was inversely related to birth weight (b=-2.6 mm Hg/kg (SE=0.5, P<0.0001) in a simultaneous regression analysis with age, sex, and current weight.
Prepubertal children—Table 3 shows that blood pressure was also inversely related to birth weight in the 610 prepubertal children. Placental weight was inversely related to systolic pressure after adjustment for age, sex, and current weight (test for trend, P=0.003). This relation became non-significant in a simultaneous regression with birth weight. There was no significant relation between the ratio of placental to birth weights and blood pressure. The associations between birth weight and blood pressure were independent of gestational age and of maternal blood pressure during pregnancy (test for trend allowing for maternal systolic blood pressure, P<0.001).
For the whole sample of 1610 children, glycated haemoglobin was directly related only to triceps skinfold thickness (P<0.001) in a simultaneous regression analysis with age, sex, and current weight.
Prepubertal children—Glycated haemoglobin level was non-significantly related to weight, ponderal index, and head circumference at birth in a simultaneous regression with age, sex, and current weight. However, table 4 shows that crown-heel length was inversely related to glycated haemoglobin level independently of gestation (test for trend, P=0.003). The ratio of placental to birth weights was not related to glycated haemoglobin level.
For the whole sample of 1610 children, cholesterol concentration was positively related to socioeconomic status (P<0.001), triceps skinfold thickness (P<0.001), and first trimester maternal body mass index (P=0.002) and inversely related to current height (P=0.001) in a simultaneous regression analysis with age and sex.
Prepubertal children—Cholesterol concentration was not associated with birth weight in a simultaneous regression analysis with age, sex, and current weight. Cholesterol was, however, inversely associated with crown-heel length (table 5), but this association was not significant in a simultaneous analysis with placental weight.
Our findings indicate that childhood blood pressure, glycated haemoglobin, and serum cholesterol concentration were related to weight and proportions at birth as well as to body size and proportions in childhood. The associations with length at birth were established in the 610 children with complete records who were aged 10 years or less. This analysis of 38% of the total sample might have introduced bias, but these children were similar to the total sample in their gestational age, anthropometry at birth, sex distribution, and maternal characteristics. They had lower blood pressure because they were younger and smaller.
Our finding that systolic blood pressure was inversely related to birth weight is consistent with previous studies.4 5 6 10 11 12 13 In adolescence tracking of blood pressure is disturbed and associations between blood pressure and birth weight are obscured.19 20 21 As expected, systolic blood pressure was related to childhood weight and maternal blood pressure and the relation between birth weight and blood pressure was independent of maternal blood pressure.22 23 24 25
A high level of glycated haemoglobin was related to shortness at birth as well as fatness in childhood. Thus, factors that influenced birth length were later associated with blood glucose control. The relation between glycated haemoglobin and skinfold thickness is not surprising given the known relation between obesity and insulin resistance and glucose intolerance.6 26 27 28 29
Serum total cholesterol concentration in Jamaican children was related to shortness at birth as well as shortness in childhood. The relation between weight and serum cholesterol concentration has been noted before in children and adults.27 28 29 30 31 32 The association between body length at birth and cholesterol might reflect abnormal intrauterine growth in which retarded trunk and visceral growth was associated with alterations in lipid metabolism. Abdominal circumference at birth, reflecting visceral growth, has been related to serum cholesterol concentration in adults.8 29
Antecedents of chronic cardiovascular disease track from childhood through to adulthood.30 31 32 33 34 35 36 37 38 39 Tracking of blood pressure is evident by 1 year of age and becomes stronger with increasing age.33 34 35 36 Serum cholesterol concentration also tracks.37 38 39 This has been related to body composition and physical activity, but it might be related to metabolic programming accompanying restraint of fetal growth, which presents as shortness at birth.30 31 32 33
Environmental forces acting throughout childhood and adult life are confounding variables, and familial microenvironments might influence birth weight.9 In this study socioeconomic status was assumed to reflect social deprivation, an independent predictor of death from cardiovascular disease.9 However, socioeconomic status had no independent association with blood pressure or glycated haemoglobin level. In addition cholesterol concentration was higher in the richer rather than the poorer children. Thus, fetal programming is perhaps more important than social deprivation in determining cardiovascular risk factors in children.
Raised blood pressure in Jamaican children was related to lower birth weight independently of maternal blood pressure. Children who had been shorter at birth had raised glycated haemoglobin levels and serum cholesterol concentration. Shortness at birth probably identifies children whose growth was retarded in late gestation. Thus, retardation of intrauterine growth might be associated with the programming of glucose and lipid metabolism.
We thank all the children who participated in this study, as well as to the staff of the 27 schools who helped. We thank Miss Jean Andrade and the staff of the records department, University Hospital of the West Indies, for tracing the antenatal and delivery records. We also thank Marvin Reid, Audrey Morris, Yvonne Grandison, and Margaret Stewart for field work and Professor D J P Barker for his comments during analysis of this study.
Funding The Carnegie Corporation and the Wellcome Trust funded this study.
Confict of interest None.