Acute bacterial meningitis during and after pregnancy
Prof S Lucas, Department of Histopathology, St Thomas’ Hospital, London, SE1 7EH, UK. Email firstname.lastname@example.org
Please cite this paper as: Lucas S. Acute bacterial meningitis during and after pregnancy. BJOG 2012;119:1555–1557.
Relatively few obstetricians in high-income countries will experience the tragedy of a maternal death, and fewer still will see one result from meningitis. When such cases arise, the obstetricians will be partners in care alongside physicians and intensivists. It is more than half a century since a major review on meningitis in pregnancy has been published, and here BJOG presents a significant report and review of the condition.
Acute bacterial meningitis (ABM; that is, acute pyogenic, non-tuberculous-viral-fungal infection) still has a high mortality rate of >20% in adults, with rapid evolution and relatively non-specific presenting signs and symptoms. Antibiotic resistance is not an issue with the usual causative organisms. The problems lie in delayed diagnosis and the biology of the disease. The observational study on fatal meningitis (all causes) among pregnant women in the Netherlands published in this issue is significant for several reasons.1 Firstly, it is the first study of meningitis in pregnancy published in 50 years that is not a case report, or case series of between one and six cases,2 but instead represents all the known cases for a country over a period of 23 years (1983–2007). Fifteen of 489 maternal deaths (direct and indirect causes) were ABM or other meningoencephalitis. In fact, all of the meningitis cases happened in the last 15 years, when 344 mothers died, making meningitis the cause of 4.4% of maternal deaths. From admittedly small numbers, the case fatality rate could be as high as 38%. Secondly, the rate of pregnancy-associated ABM in the Netherlands appears to be higher than that in the only other European country (the UK) that publishes similar audits of all known maternal deaths. The incidence of meningitis in the Netherlands for 1993–2007 was 0.5/100 000 live births; the overall maternal mortality rate (MMR) was 11.8/100 000 live births. Excluding the three non-ABM cases, the Netherlands ABM rate was actually 0.4. In the UK, with a similar overall MMR, the ABM fatality rate in 2006–2008 was 0.13, and the highest recorded rate in any triennium was only 0.24.1,3 Why the difference between countries with similar socio-economic conditions? Thirdly, the organisms and chronologies are documented. Ten (67%) deaths were caused by Streptococcus pneumoniae (pneumococcus), and there were single cases of Streptococcus milleri, Haemophilus influenzae, Mycobacterium tuberculosis, Cryptococcus neoformans and herpes simplex meningoencephalitis. All the streptococcal and the herpes simplex virus (HSV) infections were diagnosed before delivery (15–37 weeks of gestation; median 34 weeks of gestation); the cryptococcal, tuberculous and H. influenzae patients presented 10–27 days postpartum. Only one woman had an HIV infection and was receiving antiretroviral therapy, and she developed pneumococcal infection. The pattern and chronology of the infections raise questions over the susceptibility to infection in pregnancy. Fourthly, the clinical features, diagnostic processes, treatments and possible risk factors for fatal meningitis are documented. Nearly all (87%) presented with headache, most had altered mental status and fever, but only five had documented neck stiffness. The microbiological diagnoses were based on cultures of cerebrospinal fluid (CSF) and blood; only one patient was autopsied. All but one patient received antibiotics before death: the cryptococcal patient appears to have presented with postpartum psychosis, and was diagnosed post-mortem. Regarding risk factors for infection, no patient had spinal anaesthesia or cranial operative interventions. But nine of 15 patients had clinical and/or radiological evidence of otitis infection prior to meningitis. In fact, excluding the patients with HSV, tuberculosis and cryptococcosis, where ear infection would not be relevant pathogenetically, nine of 12 (75%) women had preceding ear infection. This raises questions over the management of otitis in pregnancy.
Obviously, such a retrospective study has limitations, which are acknowledged. All pregnancy-associated meningitis deaths may not be reported to the Dutch Maternal Mortality Committee (MMC), particularly if they are postpartum. The absence of reported cases from 1983 to 1993 suggests that under-reporting must happen alongside, presumably, improved diagnostic methods over time. Some clinical data are always missing in national studies such as this, perhaps including more HIV-positive women. There are no denominator data on women, pregnant or not, with meningitis who did not die. This is a standard problem with all medical confidential enquiries that are based on analysing deaths. These facts are also common to the Confidential Enquiries into Maternal Deaths in the UK. Nonetheless, we can ask why pneumococcus infection appears to be so important in pregnancy-associated meningitis, why ear infection appears to be an important predisposing factor and whether this explains the differences in data from Netherlands and the UK?
Infection in pregnancy
Bacterial sepsis in pregnancy has at least five distinct clinicopathological and pathogenetic subtypes,3 and only some cases represent the WHO definition of ‘puerperal sepsis’ (see the Appendix). Unsafe termination of pregnancy and infections introduced iatrogenically (for example via epidural anaesthesia or craniotomy for cerebral swelling) are well documented, and the latter are known to be associated with ABM. Ascending genital tract infection in the second trimester, as a cause of maternal systemic sepsis, does not present as meningitis, and is usually caused by gram-negative bacilli or group B, but not group A, streptococci. Pregnant women presenting with overwhelming septic shock before labour and rupture of membranes mostly have community-acquired group A Streptococcus pyogenes (GAS) infection, but do sometimes have pneumococcal infection. Similarly, most women with postpartum genital tract infection and systemic sepsis (‘classical puerperal sepsis’) have GAS infection, usually acquired in the community, not in healthcare centres. The latter two subtypes do not involve meningitis as a presenting feature.4
Pregnancy-associated ABM is thus a separate case from these scenarios, and when fatal (in the UK classification) is classed as an indirect death: that is, a disease not directly related to pregnancy but presumed to have been worsened by it. This is a contentious statement as it begs the question of whether pregnant women are more susceptible to infection than non-pregnant women. Since Hippocrates there has been a general assumption that they are more susceptible to infection, but real evidence is hard to find. Perhaps many of these infections are actually coincidental, but are counted assiduously because of the temporal association with pregnancy? In fact, the debate centres on whether there is an innate pregnancy-associated immunodeficiency. No one disputes, for example, the increased susceptibility to ascending genital tract infection after the membranes have ruptured.
There has been much speculation on, and experimental study of, the putative reversible T-cell immune response—the Th1/Th2 shift—in pregnancy, which could reduce antibacterial and general immunity in pregnancy.5 Some infections, such as listeriosis, are indeed more common in pregnancy, and leprosy rebounds in clinical severity after delivery. Pregnant mice are certainly more susceptible to leishmania infection,6 but for human tuberculosis, case–control studies provide contradictory evidence of susceptibility during and after pregnancy.7,8 And for the numerically more important GAS infections, there is no good evidence for a pregnancy-specific susceptibility. Although it is an attractive proposition that pregnancy requires a reduced immune response in order to preserve the fetus, there is no good evidence that either T-cell or B-cell responses are impaired during pregnancy, nor that there is a general susceptibility to infections apart from pneumococcal, group B streptococcus (GBS) and Listeria.9,10
Supporting in part the Schaap study, a recent population-based study of invasive streptococcal infection in pregnancy in the USA found that pneumococcal infection was three times more common postpartum compared with non-pregnant women, whereas during pregnancy the incidence rates were similar.11 Conversely, for GAS and GBS women were >20 times more likely to acquire infection postpartum compared with non-pregnant women, but during pregnancy there was a doubling of incidence for GBS infections only. The marked postpartum increase for GAS and GBS infections presumably relate to the recently disrupted genital tract. The difference from the Dutch scenario is that there, all the pneumococcal infections presented before delivery, not afterwards.
Why pneumococcal infection in pregnancy?
In adults (and children) the pneumococcus is the most common cause of ABM. In the USA it accounted for 58% of all infections in 1998–2007.12 However, in that large study, pregnancy was not identified as a factor except for the uncommon Listeria infection.
Susceptibility to pneumococcal infection is widely known to be increased by many risk factors, including HIV disease, sickle-cell disease and H1N1 lung infection, probably through different physical and immunological factors. These recent meningitis and pregnancy studies show that pregnancy and the postpartum status can reasonably be added to that list, although the relative increase in incidence is not so great. Schaap et al. do not discuss whether universal pneumococcal vaccination in pregnant women is worthwhile, and it would probably require much more evidence to convince public health policy makers to go down that route. However, it appears that pneumococcal infections should indeed be considered as indirect in pregnancy associations, not coincidental.
Otitis infection and meningitis
The Schaap study does not mention another recent publication reporting on six pregnant women with ABM, diagnosed since 2005, at a Dutch hospital.2 All had pneumococcal infection: two died and four had otitis. Pneumonococcus is the most common pathogen isolated from the ear in otitis.13 This reinforces one of the main messages from the study: that otitis needs to be treated seriously in pregnancy, and that patients should be followed up to ensure the eradication of infection.
Why the Netherlands should be so different from UK is not clear. The most recent confidential enquiry found only four cases of meningitis: two were pneumococcal and one was a GAS infection following craniotomy.3 There was no mention of otitis. Of note in the Dutch study is that no meningitis patients had had any spinal anaesthesia. Did the UK case ascertainment process miss meningitis cases? Do UK doctors treat otitis in pregnancy more aggressively? This study is a reminder for the organisers of the next enquiry to think of meningitis, both pre- and postpartum. The UK The Health Protection Agency (HPA) presents many data on the epidemiology of the streptococcal infections (more on GAS than pneumococcus), but until now has failed to include pregnancy states as data categories.
The main and important conclusion from the Dutch study is that meningitis in pregnancy and postpartum has been historically underestimated in terms of frequency and mortality. The clinical symptoms are frequently vague and not those of the classical meningitis triad, meaning that healthcare workers need to consider the possibility of meningitis more proactively and diagnose it faster.
Disclosure of interests
There are no conflicts of interest to disclose.
Contribution to authorship
SL is the sole author.
Details of ethics approval
Puerperal sepsis is defined by the WHO as:
‘Infection of the genital tract occurring at any time between rupture of membranes or labour, and the 42nd day postpartum, in which two or more of the following are present
- • Pelvic pain
- • Fever
- • Abnormal vaginal discharge
- • Abnormal smell of discharge
- • Delay in reduction of size of uterus’