Obstetric intervention prevents adverse perinatal outcome

It has often been said that ever increasing obstetric intervention is not accompanied by significant reductions in perinatal mortality or cerebral palsy, with the result that obstetricians are often criticised for advocating electronic fetal monitoring too strongly and caesarean section too frequently. Three articles in this issue should give comfort to obstetricians.

Svetlana Glinianaia and her colleagues (pages 452–460) conducted an epidemiological survey to compare perinatal mortality in two consecutive epochs in twin and singleton pregnancies. The authors obtained information from the Northern Region Perinatal Mortality Survey in the United Kingdom, a population survey operating since 1981 involving sixteen health districts in the region. The period of the survey was arbitrarily divided into two equal halves in order to examine secular trends in perinatal mortality. The analysis is complex, for it involves comparison of perinatal mortality in twin and singleton pregnancy throughout the whole period, as well as comparison of perinatal mortality in singletons and twins separately in the two epochs. As expected, all causes of perinatal mortality were greater in twin pregnancy, but it is the secular trends which are especially important. The perinatal mortality due to fetal deformity, antepartum anoxia in twins and antepartum asphyxia or trauma in singletons declined substantially in the later epoch. There was no change in perinatal mortality due to immaturity or antepartum anoxia in singleton pregnancies; in fact, the perinatal mortality from antepartum anoxia increased in singleton pregnancies where the birthweight was greater than 2500 grammes.

Glinianaia and colleagues suggest that the decrease in perinatal mortality due to fetal deformity is the result of prenatal diagnosis and termination of pregnancy; that the decrease in perinatal mortality due to anoxia in twin pregnancies may be the result of routine surveillance in pregnancy with ultrasound and Doppler; and that the increase in perinatal mortality due to anoxia in singletons weighing more than 2500 grammes may be the result of greater emphasis on antenatal care in the community. And although the authors do not say so, a reader may infer that the decrease in perinatal mortality due to intrapartum asphyxia or trauma is the result of fetal monitoring and caesarean section in labour.

The strengths of this epidemiological survey are that it describes the experience of a population and not just a single hospital; that the collection of data was complete; and that establishment of the causes of perinatal death was rigorous. Another virtue of the study is the use of Wigglesworth's classification of perinatal mortality, this classification giving the best indication of the effects of medical treatments. The authors are rightly cautious about the interpretation of their study, however, for the comparison of the later epoch was with the historical controls of the earlier, and so the findings of the study may be affected by confounding. Furthermore, the outcomes studied are rare and the methods of fetal surveillance are intrinsically unreliable, so that many pregnant women would have undergone invasive procedures such as induction of labour and caesarean section in order to save a few infants. The true place of fetal surveillance will be realised by decision analysis in each pregnant woman, where the risks of these invasive procedures are weighed against the risks of perinatal death.

Another measure of obstetric care is the changing frequency of hypoxic-ischaemic encephalopathy in term infants, studied by Julie Smith and her colleagues (pages 461–466). The authors compared the frequency of hypoxic-ischaemic encephalopathy in three epochs in a single maternity unit and found a progressive decrease, being 7.6 per 1000 live births in 1980; 4.6 per 1000 live births in 1988; and 1.9 per 1000 live births in 1996. The same trend was seen if only moderate or severe cases were considered. Furthermore, the outcome of infants who sustained hypoxic-ischaemic encephalopathy improved, with 48% achieving normal life in 1980; 59% in 1988; and 69% in 1996, this change being statistically significant (χ2 test for trend 11.14; P < 0.001). These improvements occurred in an era of increasing delivery in a specialist maternity unit, greater use of electronic fetal heart rate monitoring and caesarean section, and better staffing of the paediatric and anaesthetic departments. Hypoxic-ischaemic encephalopathy in term infants is a direct measure of obstetric care and is little affected by confounding variables. The results of this study strongly suggest that it is the changes in the maternity service that are responsible for the declining frequency of hypoxic-ischaemic encephalopathy and the decrease in long term damage due to this condition. Hypoxic-ischaemic encephalopathy can be caused by acute severe fetal hypoxia in the second stage of labour where it is necessary to effect a vaginal delivery as quickly as possible. Is forceps delivery quicker than vacuum extraction? This is the question asked by Yetunde Okunwobi-Smith and colleagues (pages 467–471), who carried out a retrospective survey in 225 women of the interval from decision to delivery by the two instruments. The main finding of the study was that in fetal distress delivery was accomplished in six minutes more quickly with forceps than with vacuum extraction. The slowness of vacuum extraction was partly explained by its failure to deliver the infant in one in six instances, where forceps delivery or caesarean section was then performed. Six minutes may not seem long, but the delay may be significant where fetal hypoxia has already occurred in the first stage of labour, and the fetal heart rate trace suggests anoxia in the second stage.

These papers suggest that doubts about the effectiveness of obstetric treatments are misplaced, for the organisational and technical changes in obstetric services in the past two decades are strongly associated with diminishing perinatal mortality and morbidity. Serious adverse outcomes for the infant are so rare that randomised trials may be impossible, and we shall rely for evidence on epidemiological survey and observational studies. We shall never be able to allow completely for confounding factors, but if the findings of several surveys are consistent we can confidently apply their results to the provision of obstetric services.