The death of a woman who is pregnant is doubly tragic, resulting in loss of two lives or, at a terrible minimum, the risks of prematurity or a motherless upbringing. In the UK, regular audits of maternal and neonatal mortality over 60 years have aimed to minimise the risk to mothers and babies. The most recent was published in late 2014 from the Maternal, Newborn and Infant Clinical Outcome Review Programme produced by MBRRACE-UK (Mothers and Babies: Reducing Risks through Audits and Confidential Enquiries across the UK). These data highlight the need for changing the structures of care or pregnant women with pre-existing neurological conditions as well as those with neurological complications of pregnancy.

The confidential enquiries into maternal mortality began in 1952, initially just for England and Wales, with later versions including Scotland and latterly the Republic of Ireland. The most recent report covers the 3 years from 2009. The process involves classification (table 1) and expert assessors’ scrutiny of maternal deaths. Any deficiencies in care identified are rated on whether or not they affected outcome.

Ascertainment rates and data collection rates are now exceptionally high, the yield being twice that using death certification alone, with complete records obtained for a remarkable 99% of cases. This near-complete ascertainment may explain some of the trends in increasing deaths documented in the late 1980s and 1990s.

The early audits found the UK maternal mortality rate was around 90 per 100 000 pregnancies. The prevention and treatment of pre-eclampsia, perinatal haemorrhage and sepsis appears to have driven a 90% fall in maternal mortality in six decades. Even since the early 1980s, improved obstetric care was bringing about significant reductions in direct mortality from around 6 to 3 per 100 000. However, in contrast to these improvements in obstetric complication rates, there has been a significant and steady rise in deaths in pregnancy from other causes. The recent audits have analysed these by specialty to determine how these could be improved.

The most recent audit confirms these previous trends and this may leave obstetricians feeling relatively upbeat; despite increasing numbers of births, increasing levels of obesity, increasing maternal age and increasing numbers of pregnancies in non-UK born women, the death rate among pregnant women has continued to fall. However, the results still give cause for concern since those deaths not directly related to pregnancy continue at the same rate. Increasing ascertainment may explain an initial jump in indirect mortality in the 1990s, but there has been no progress in reducing this.

Neurological disorders account for about 20–30% of indirect maternal deaths in pregnancy in the UK since the 1980s (table 2). Wills and Kelso reviewed the stroke management of 26 patients dying following intracranial haemorrhages from 2009 to 2012.1 There were equal numbers of subarachnoid haemorrhages and intracerebral bleeds. Fifty-six per cent of the stroke cases were felt to have had optimal care, with potentially contributory deficiencies in only 19%. Most cases presented no warning signs or symptoms, but in a few patients, a misdiagnosis of migraine was followed by fatal rebleeding of subarachnoid haemorrhages. The authors recommend a high index of suspicion in pregnant patients with headaches who have neurological deficits or neurological signs (including neck stiffness)—though this probably applies to non-pregnant patients too. The main message is that stroke care should not be changed by any stages or outcomes of pregnancy, and that there is no specific contraindication to interventions for subarachnoid haemorrhage or cerebral infarction.

Epilepsy remains a prominent cause of maternal mortality. Between 2009 and 2012, 14 pregnant women with epilepsy died up to 42 days after delivery: two were ascribed to drowning, while the rest were attributed to sudden unexpected death in epilepsy. The assessors identified 10 cases (71%) where changes in care may have influenced outcome and a further two where care could have been improved without necessarily implying contribution to outcome.

Importantly, there were a further 12 deaths in women with epilepsy occurring between 6 and 52 weeks postpartum. Again, most were felt to have been preventable or potentially affected by improvements in epilepsy management.

Access to specialist services was the main factor lacking in the majority of preventable deaths as this stopped provision of advice about continuing treatment, managing breakthrough seizures and anticipating/managing sleep deprivation. The single rooms were considered as a factor in two in-hospital deaths.