Objective To study the characteristics of birthweight and gestational age of third trimester fetal deaths which occurred before the onset of labour, Design Review of computerised confidential perinatal mortality records. Data originated from the 1992 Trent Region Perinatal Mortality Survey.Sample One hundred and forty-nine antepartum stillbirths of at least 24 weeks of gestation confirmed by early ultrasound scan. Congenital abnormalities and multiple pregnancies were excluded.Main outcome measures Reported causes of stillbirth, weight-for-gestational age centiles based on a standard derived from normal pregnancies; pregnancy characteristics compared with the local maternity population.Results Of 149 stillbirths, 83 (56%) were preterm and 66 were at term, and the majority (126; 85%) occurred from 3 1 weeks. Most of the deaths (97; 65%) were reported as 'unexplained' even though post-mortems had been carried out in 60% of all cases. Using a gestational age-specific fetal weight standard derived from normal, term live births, 41% of all cases of stillborn infants were small-forgestational age (< 10th centile; OR 6.2; 95% CI 3.3-1 1.5); 39% of which had been classified as unexplained were small for gestational age (OR 5.6; 2.6-12.0). This excess of small stillbirths was most pronounced between 3 1 and 33 weeks, where the weights of 63% of all stillbirths and 72% of unexplained fetal deaths were < 10th centile. Overall, a higher proportion of preterm (< 37 weeks) than term stillbirths were small for gestational age: 53% vs 26% (OR 3.3; 1.6-6.5). However, at term there were also more subtle differences in weight deficit, with more fetuses with a weight between the 10th and 50th centiles than between 50th and 90th (36 vs 11; OR 3.3; 1-4-7.8). Mothers of pregnancies ending in stillbirth were similar in age, size, parity and ethnic group to mothers of live born babies, but were more likely to be smokers (37 vs 27%, OR 1.6; 1.2-2.3).Conclusions Many stillborn babies are small for gestational age. In the absence of significant differences in physiological pregnancy characteristics, this is unlikely to be a constitutional smallness, but represents a preponderance of intrauterine growth restriction. For a full appreciation of the strength of this association, appropriate weight standards and classifications need to be applied in perinatal mortality surveys. Many antepartum stillbirths which are currently designated as unexplained may be avoidable if slow fetal growth could be recognised as a warning sign.

Design-based stereology is employed to estimate total numbers of myocyte nuclei and mean myocyte volume per nucleus in ventricles of fetal and early postnatal human hearts. Organs were collected postmortem from subjects varying in age from 16 gestational wk to 40 postnatal wk. Numbers of myocyte nuclei per unit volume of ventricle were estimated using physical disectors (parallel pairs of sections). Absolute numbers were calculated by multiplying nuclear packing densities by ventricular volumes estimated from ventricular mass and tissue density. Volumes per nucleus were obtained via estimates of the combined volumes of all myocytes (or of the myocardium as a whole) and the numbers of myocyte nuclei. The findings showed that numbers of myocyte nuclei increase linearly from 16 wk towards term. They were also consistent with the notion that hyperplasia ceases abruptly at birth or soon afterwards. The net rate of production of myocyte nuclei was about 38i10(\wk (2.3 million nuclei\h). The total volume of myocytes continued to expand in the same way from 16 wk to at least 35 wk of gestation. Published studies on the incidence of binucleate myocytes during early postnatal growth of the ventricles of rats suggest that the volume of a myocyte doubles prior to nuclear division. Prenatal growth in the human heart is consistent with this mechanism. Myocardial hypertrophy after birth must occur by cellular hypertrophy without karyokinesis.

SUMMARY The brains of 30 infants who died after at least one real time ultrasound scan were examined after fixation. The ultrasound diagnosis of either periventricular haemorrhage or periventricular leucomalacia was compared with the macroscopic and histological appearances. Each hemisphere was considered separately for both periventricular haemorrhage and periventricular leucomalacia. The accuracy of ultrasound diagnosis for periventricular haemorrhage was 88%, with sensitivity of 91% and specificity of 85%. The accuracy for periventricular leucomalacia was 90%, with sensitivity of 85% and specificity of 93%. Ultrasound was shown to diagnose the entire range of periventricular leucomalacia lesions. Three hemispheres showed the appearance of prolonged flare, and this correlated with extensive spongiosis and microcalcification of the periventricular white matter, although no macroscopic lesion was seen.

The syndrome of prolonged expirational apnoea is well recognised.' In this syndrome children may rapidly develop very profound arterial hypoxaemia after relatively short periods of apnoea.2 It has been postulated that this hypoxaemia is due to the sudden development of intrapulmonary arteriovenous shunting.It is a commonly held belief that there is no anatomical basis for the shunting, but it is observed at ages by which the foramen ovale and the ductus arteriosus have closed. In live children exhibiting shunting of this type echocardiography has been used to demonstrate the absence of intracardiac shunts and that the ductus arteriosus is closed.3 Subsequent necropsy confirmed the closure of the ductus but the shunting remained unexplained.The belief that there is no anatomical basis for the shunting appears to be based on the work of Hislop and Reid who described the development and anatomy of the pulmonary vasculature in a group of children with ages ranging between 3 days and 11 years but found no evidence of shunts.4 Their method outlined the pulmonary vessels by injection of a concentrated solution of gelatin (25%) into the pulmonary artery of previously frozen lungs. The vessels were assessed on radiographs of the lung with limited histology. Vessels down to 15 iim in diameter were filled with the mixture but no gelatin was seen in the pulmonary capillaries or in the pulmonary veins.We used a similar technique to Hislop and Reid with a less concentrated solution and lower injection pressure in fresh lungs and found that in some cases the injection mixture flowed freely from the pulmonary vein. Histologically gelatin can be seen in the pulmonary veins, occasional capillaries, and airspaces.Material and methods Since 1985 we have been injecting the pulmonary arteries with a solution of gelatin to study the development of the pulmonary vasculature. In some cases we noted that the gelatin flowed freely from the pulmonary vein, so we began to measure the pressure required. The fresh lungs of 49 infants up to 44 weeks' old, whose deaths had been referred to the coroner, were used in this study. All the infants had died to some degree unexpectedly both in hospital and at home. The Emery system of classifying infant death was used.' In this class A are infants with a known fatal illness who are expected to die; class B are infants with a curable but potentially life threatening illness that may be treated, class C are infants with antemortem or postmortem evidence of minor illness who die unexpectedly, and class D are infants with little or no postmortem evidence of disease to explain death. It must be emphasised that the classification is based on findings at necropsy and on examination of subsequent histology. Thus significant but unexpected findings are classified as for the disease, not the state of apparent wellbeing of the infant before death.The thoracic contents were removed intact, the right lung isolated, and the right pulmonary artery cannulated. A 50 ml syringe was connected by a three way tap to the c...

SUMMARY Samples of fresh and fixed tissues from infants with exomphalos treated by thiomersal application were analysed for mercury content. The results showed that thiomersal can induce blood and organ levels of organic mercury which are well in excess of the minimum toxic level in adults and fetuses. The analysis of fresh and fixed tissues must be carefully controlled against normal tissues in order to interpret mercury levels accurately.