Adverse effects of maternal lead levels on birth outcomes in the ALSPAC study: a prospective birth cohort study

OBJECTIVE

To study the associations of prenatal blood lead levels (B-Pb) with pregnancy outcomes in a large cohort of mother-child pairs in the UK.

DESIGN

Prospective birth cohort study.

SETTING

Avon area of Bristol, UK.

POPULATION

Pregnant women enrolled in the Avon Longitudinal Study of Parents and Children (ALSPAC).

METHODS

Whole blood samples were collected and analysed by inductively coupled plasma dynamic reaction cell mass spectrometry (n = 4285). Data collected on the infants included anthropometric variables and gestational age at delivery. Linear regression models for continuous outcomes and logistic regression models for categorical outcomes were adjusted for covariates including maternal height, smoking, parity, sex of the baby and gestational age.

MAIN OUTCOME MEASURES

Birthweight, head circumference and crown-heel length, preterm delivery and low birthweight.

RESULTS

The mean blood lead level (B-Pb) was 3.67 ± 1.47 μg/dl. B-Pb ≥ 5 μg/dl significantly increased the risk of preterm delivery (adjusted odds ratio [OR] 2.00 95% confidence interval [95% CI] 1.35-3.00) but not of having a low birthweight baby (adjusted OR 1.37, 95% CI 0.86-2.18) in multivariable binary logistic models. Increasing B-Pb was significantly associated with reductions in birth weight (β -13.23, 95% CI -23.75 to -2.70), head circumference (β -0.04, 95% CI -0.07 to -0.06) and crown-heel length (β -0.05, 95% CI -0.10 to -0.00) in multivariable linear regression models.

CONCLUSIONS

There was evidence for adverse effects of maternal B-Pb on the incidence of preterm delivery, birthweight, head circumference and crown-heel length, but not on the incidence of low birthweight, in this group of women.

Girl or boy? Prenatal lead, cadmium and mercury exposure and the secondary sex ratio in the ALSPAC study

The aim of this study was to evaluate the effect of prenatal exposure to lead, cadmium and mercury levels on the secondary sex ratio. Whole blood samples were collected from pregnant women enrolled in the Avon Longitudinal Study of Parents and Children (ALSPAC) study at a median gestational age of 11 weeks and were analyzed for lead, cadmium and mercury. Regression analysis was used to identify associations between maternal lead, cadmium and mercury levels and the secondary sex ratio with adjustment for confounders. There was no evidence for associations between maternal lead, cadmium or mercury levels and the secondary sex ratio in this sample. It appears unlikely that alterations in the secondary sex ratio are influenced by exposure to heavy metals, but further work should be done in large cohorts in other countries to confirm these findings.

Effects of early childhood lead exposure on academic performance and behaviour of school age children

AIM

To determine whether early lead exposure at levels below 10 microg/dl has an impact on educational and behavioural outcomes at school.

METHODS

Venous samples were taken from a subgroup of the Avon Longitudinal Study of Parents and Children (ALSPAC) attending a research clinic at 30 months of age (n = 582), and lead levels were measured by atomic absorption spectrometry. Developmental, behavioural and standardised educational outcomes (Standard Assessment Tests, SATs) were collected on these children at age 7-8 years. In the analysis, blood lead concentration was investigated both as a continuous covariate and as a categorical variable.

RESULTS

488 cases (84%) had complete data on confounders and outcomes. After adjustment for confounders and using a log dose-response model for lead concentration, blood lead levels showed significant associations with reading, writing and spelling grades on SATs, and antisocial behaviour. A doubling in lead concentration was associated with a 0.3 point (95% CI -0.5 to -0.1) decline in SATs grades. Treating lead levels categorically, with the reference group 0-2 microg/dl, no effects on outcomes were apparent at 2-5 microg/dl, but levels of 5-10 microg/dl were associated with a reduction in scores for reading (OR 0.51, p = 0.006) and writing (OR 0.49, p = 0.003). Lead levels >10 microg/dl were also associated with increased scores for antisocial behaviour (OR 2.9, p = 0.040) and hyperactivity (OR 2.82, p = 0.034).

CONCLUSIONS

Exposure to lead early in childhood has effects on subsequent educational attainment, even at blood levels below 10 microg/dl. These data suggest that the threshold for clinical concern should be reduced to 5 microg/dl.

The variation in stool patterns from 1 to 42 months: a population-based observational study

OBJECTIVE

To document the normal stool patterns of young children.

DESIGN

Prospective population-based longitudinal study.

SETTING

Avon Longitudinal Study of Parents and Children (ALSPAC).

SUBJECTS

12,984 children, whose parents completed questionnaires at 4 weeks, 6, 18, 30 and 42 months on their frequency of bowel movements and the consistency and colour of their stools.

RESULTS

Stool frequency declined from a mean of 3.0 times/day (3rd centile 0.6, 97th centile 5.9) at 4 weeks to 1.3 times/day (0.6, 2.7) at 42 months. Stool consistency was soft in most babies with nearly half passing liquid or curdy stool at 4 weeks. 14% of babies usually passed a hard stool at 4 weeks, rising to 30% at 42 months. Stool colour was commonly yellow at 4 weeks and had changed to brown by 6 months. Black stools were extremely unusual at all ages.

CONCLUSIONS

These data on the changes with age in the stool patterns of young children will be useful for clinicians.

Family, socioeconomic and prenatal factors associated with failure to thrive in the Avon Longitudinal Study of Parents and Children (ALSPAC)

BACKGROUND

The epidemiological profile of infants failing to thrive is unclear. The aim of this study is to investigate the prenatal and socioeconomic factors associated with these infants using standardized weight gain conditional on previous weight.

METHODS

In a large UK population cohort study, 11 718 infants born at term in 1991-1992 with no major congenital abnormalities were identified. Using a weight gain criterion conditional on initial weight from birth to 6-8 weeks, 6-8 weeks to 9 months, and birth to 9 months, the slowest gaining 5% were identified.

RESULTS

None of the prenatal factors was associated with failure to thrive in the multivariable analysis nor were traditional markers of socioeconomic deprivation such as poor parental education or low occupational status. Parental height was significantly correlated with slow infant weight gain in both separate periods and from birth to 9 months (Pearson's r = +0.20, P < 0.001). Eight times as many infants born to shorter parents (8.7%, 95% CI: 6.6, 11.3) showed slow weight gain as infants born to taller parents (1.1%, 95% CI: 0.5, 2.5). Higher parity was also related to slow infant weight gain; infants born in the fourth or subsequent pregnancy were twice as likely to fail to thrive from birth to 9 months (8.3%, 95% CI: 6.4, 10.6) as first-born infants (3.4%, 95% CI: 2.9, 10.6).

CONCLUSIONS

Future studies need to take account of parental height when calculating growth standards and look at why failure to thrive is more common, not in poorer families but in larger families.

Health and social factors for health visitor caseload weighting: reliability, accuracy and current and potential use

The reliability and validity of data collection and recording for a health visitor caseload weighting system operating in a large combined acute and community trust in Bristol was examined. Client families were recruited from all participating (service) health visitors and were interviewed by a research health visitor attached to the project. The presence or absence of 28 health needs factors, selected for their known impact on health in the longer term, was subsequently compared with the service health visitor's own caseload records and with the computer record derived from them. Substantial differences were determined in the records obtained between the service and research health visitors and between the written and computerised records. Whilst the mean total score recorded for each family by the service health visitor was 2.0, the average of the differences in the total number of health factors recorded by the service and research health visitors was 1.9. Discrepancies were mostly associated with differences in interpretation of definitions, knowledge of recent events, changing circumstances and issues of confidentiality. Health factors at particular risk of being misinterpreted and those associated with other health factors were identified in order to propose a reduced factor set with greater inherent reliability and validity. At the level of the ward, the caseload weighting score, as currently defined, is highly correlated with standard deprivation indices in widespread common use. The results of this study indicate the need for users of the caseload weighting data to decide on the primary function of this data set - area-based community profiling or identification of at-risk client families in the community. This decision will inform further efforts to identify the most useful factors, tighten definitions, streamline data collection and train health visitors in their use. Interpretation of data will be facilitated by a scientifically developed scoring system. This work will assist trusts, both locally and nationally, in rationalising their allocation of health visiting activities to areas of greatest need.

Project Pró-natal: population-based study of perinatal and infant mortality in natal, Northeast Brazil

The Pró-Natal project is a collaborative initiative that aims to improve maternal and infant health in a deprived community in Natal, Northeast Brazil. To assess the perinatal and infant mortality in this population of 40,000, we have collected over a 2-year period a consecutive series of 39 autopsy examinations on deaths under 1 year of age. During this period there were 2212 live births in the study population. The 14 perinatal deaths are described using the Wrigglesworth classification, and the 25 infant deaths, using a clinicopathological system. The contribution of normally formed stillbirths was small (14%), which probably reflects the underreporting of stillbirths in this community. The most common cause of death in the live births was complications of prematurity (43%). Specific causes (22%) of perinatal deaths were predominantly infections, including one case of congenital syphilis. Perinatal asphyxia was diagnosed in 14%, and there was one case (7%) of a chromosome abnormality. Infant deaths were predominantly due to respiratory (45%) and gastrointestinal infections (28%), with chronic malnutrition as an underlying cause in 80% of cases. Prenatal care could theoretically have prevented three of the perinatal deaths, and a further six deaths could have been avoided by improved management of labor and the immediate neonatal period. Prevention of malnutrition and improved treatment of acute infections would contribute to a reduction in infant mortality in this population. The Pró-Natal project will use these data to design preventative interventions to reduce perinatal and infant mortality in this community.

The effects of housing on the health of preterm infants

Using a prospective study based on cases and controls selected from a geographical population, we have investigated the effects of housing on the health of very preterm infants (< or = 32 weeks' gestation) during the first year of life. Information on health morbidity of the 117 preterm and 226 term babies was collected using a parent-held record, and housing data by a validated self-completion questionnaire. The most common health problems in the first year-upper (UR) and lower (LR) respiratory tract infection, otitis media (OM) and diarrhoea and vomiting (DV)-were all more frequent in the preterm group, There were no significant differences in the housing conditions to which preterm and control infants were exposed. Relative risks (RR) and 95% confidence intervals (CI) associated with each housing factor were calculated for preterm and control infants separately. Significant (P < 0.05) interaction effects were found for overcrowding and gas cooking. Overcrowding was associated with an increased incidence of LR [RR = 1.53; CI 0.96-2.42] and DV [RR = 1.57; CI 0.92-2.67] in the preterm, but with a decreased incidence of LR [RR = 0.28; CI 0.04-1.86] and DV [RR = 0.85; CI 0.30-2.38] in the term controls. The use of gas ovens was found to be associated in preterm infants with an increase in LR [RR = 1.48; CI 0.96-2.28] and DV [RR = 2.24; CI 1.28-3.93] but the controls did not show this effect for LR [RR = 0.67; CI 0.40-1.09] or DV [RR = 0.93; CI 0.56-1.56]. These associations are robust-even after allowing for confounding social factors-but causality has not been proved. This work suggests that preterm infants may be vulnerable to specific adverse housing factors, and further studies are now indicated to clarify potential mechanisms and interactive effects behind these associations.

Haemoglobin and ferritin concentrations in infants at 8 months of age

AIM

To identify the optimum age to screen for iron deficiency, the normal distribution of haemoglobin and ferritin in a representative population sample was investigated.

METHODS

Normal values for haemoglobin and ferritin were measured from heel prick capillary samples obtained from a representative cohort of 1175 infants at 8 months old who were randomly selected from children taking part in the Avon Longitudinal Study of Pregnancy and Childhood (ALSPAC).

RESULTS

Haemoglobin was normally distributed: mean (SD) 117 (11) milligrams, 95% confidence interval (CI) 116 to 118, and range 72-153 milligrams. Ferritin was log normally distributed: geometric mean 38.5 micrograms/l, 95% CI 37.0 to 39.9, range 7.1-224 micrograms/l. The 5th centile for haemoglobin was 97 milligrams and for ferritin 16.9 micrograms/l. No correlation was found between haemoglobin and ferritin. Multiple regression analysis showed ferritin concentrations to be positively related to birth weight (p < 0.0001) and the sex of the child (girls with higher concentrations) (p < 0.0001) but negatively with the child's weight at 8 months (p < 0.0001). Haemoglobin concentrations were positively related to the child's weight at 8 months (p = 0.04). Neither haemoglobin nor ferritin concentrations were related to social class as measured by maternal education level.

CONCLUSION

These data define the normal range for haemoglobin and ferritin in capillary samples in the UK population, and suggest that anaemia is common in infancy. Using current recommendations, 23% of infants would be identified as anaemic. For British infants at 8 months of age, a more representative 'cut off' for anaemia would be haemoglobin concentration < 97 milligrams and for iron deficiency ferritin < 16 micrograms/l.

Acute splenic sequestration in homozygous sickle cell disease: natural history and management

Of a cohort of 308 children with homozygous sickle cell disease diagnosed at birth, 89 experienced 132 clinically significant attacks of acute splenic sequestration (ASS) over a 10-year period. The age at first attack ranged from 3 months to 6 years. Survival curve analysis of the interval until first attack indicated a cumulative probability of 0.225 by 2 years, and 0.265 by 3 years, and 0.297 by 5 years of age. Thirteen events were fatal, 11 during the first attack, and all before transfusion could be instituted. Recurrences occurred in 49% of survivors of the first attacks, and there were diminishing intervals between subsequent events. Respiratory symptoms were associated with 52 of 132 events, but bacterial isolates on blood culture were less frequent, and ASS was not prevented by pneumococcal vaccine or penicillin prophylaxis. A high fetal hemoglobin level protected against attacks of ASS. A parental education program aimed at early diagnosis of ASS was followed by an increase in the incidence rate for ASS from a mean of 4.6 per 100 patient-years to 11.3 per 100 patient-years, probably reflecting increased awareness of the complication. During the same periods, the fatality rate fell from 29.4 per 100 events to 3.1 per 100 events. The improvement in outcome is likely to have resulted from improvement in medical management and earlier detection of ASS.

Role of splenectomy in homozygous sickle cell disease in childhood

60 Jamaican children with homozygous sickle cell (SS) disease underwent splenectomy, 14 for prophylaxis against recurrent acute splenic sequestration and 46 for treatment of sustained hypersplenism. Age at operation varied from 9 months to 16 years. Patients were followed up for 1 month to 27 years (median 6 years), with a total of 369 years of patient-observation. None of the 3 patients who died, at ages 2 1/2, 6 1/2, and 21 years, had received prophylaxis against infection. Overwhelming sepsis was possible but not confirmed in the first two deaths which occurred 11 months and 2 1/2 years after operation; the third died from chronic renal failure 11 years after splenectomy. After operation, there were no confirmed cases of pneumococcal septicaemia or meningitis, and the commonest clinical event was the acute chest syndrome.

Priapism and impotence in homozygous sickle cell disease

A questionnaire study of Jamaican patients with homozygous sickle cell (SS) disease indicated a 42% prevalence of priapism, with a median age at onset of 21 years. Two predominantly different patterns of priapism were recorded: short "stuttering" episodes lasting less than three hours, with normal consequent sexual function, and severe prolonged attacks (generally more than 24 hours) commonly followed by impotence. Stuttering episodes were frequently a prodrome to a major attack. Over one fourth of those who had suffered priapism had some degree of impotence. Hematologic analysis indicated that priapism was significantly associated with low hemoglobin F levels and high platelet counts. Patients with severe attacks of priapism had lower hemoglobin F levels and reticulocyte count, and a higher mean corpuscular volume, than patients with only stuttering episodes. Priapism and impotence contribute to the morbidity of SS disease more frequently than previously recognized.