Fatal meningitis during pregnancy in the Netherlands: a nationwide confidential enquiry

Authors


Dr TP Schaap, Prinses Irenestraat 2, 3981 BR, Bunnik, the Netherlands. Email tpschaap@gmail.com

Abstract

Please cite this paper as: Schaap T, Schutte J, Zwart J, Schuitemaker N, van Roosmalen J. Fatal meningitis during pregnancy in the Netherlands: a nationwide confidential enquiry. BJOG 2012;119:1558–1563.

Objective  To determine the incidence of maternal deaths attributable to meningitis in the Netherlands, and to assess clinical features and risk factors.

Design  Confidential enquiry into the causes of maternal deaths.

Setting  Nationwide in the Netherlands.

Population  A total of 4 784 408 live births.

Methods  Analysis of all maternal deaths due to meningitis in pregnancy and puerperium from 1983 up to and including 2007 reported to the Maternal Mortality Committee of the Dutch Society of Obstetrics and Gynaecology.

Main outcome measures  Incidence, clinical features and risk factors.

Results  Fifteen maternal deaths occurred due to meningitis, representing 4.4% of all maternal deaths. Twelve women (80%) presented with meningitis during pregnancy, 8 (66%) of them in the third trimester. Presenting symptoms were altered mental status (11; 73%), fever (9; 60%), nuchal rigidity (5; 33%) and headache (13; 87%). Nine women (60%) had otolaryngological infection at presentation or in the previous days or weeks. Twelve women (80%) underwent radiological examination, of which 5 (33%) showed distinct abnormalities. Cerebrospinal fluid (CSF) examination showed infected CSF in 8 (53%) women. In ten women (67%) Streptococcus pneumoniae was isolated. Substandard care was identified in 4 (27%) women.

Conclusion  Pregnant or puerperal women presenting with classical symptoms of meningitis, particularly those with a history of otolaryngological infection or headache, should undergo thorough investigation and radiological and CSF examinations. Early diagnosis and immediate antibiotic treatment are imperative because of rapid deterioration in pregnant women. In case of doubt, the threshold for antibiotic treatment should be low and close monitoring is warranted.

Introduction

Meningitis is a life-threatening disease that can develop rapidly and, if treatment is withheld, it leads to an extremely high case fatality rate. In a series of 696 men and nonpregnant women, the duration of symptoms until admission was <24 hours in 48% of cases; the case fatality rate was 21%.1 Presenting symptoms of meningitis can be diverse, the most common symptom being headache—present in almost 90% of patients.1 A classical triad exists of fever, nuchal rigidity and altered mental status. All three symptoms, however, are present in only 44–46% of patients with meningitis.1,2 The prevalence and mortality rate of meningitis in pregnant women are not known, though recent research in a low-resource country with a high HIV-prevalence has shown a high case fatality rate.3 Furthermore, it is unknown whether clinical features and risk factors in pregnant women differ from those in the general population. In the Netherlands, a national ongoing Confidential Enquiry into Maternal Deaths is being carried out by the Dutch Maternal Mortality Committee (MMC). In this study we aimed to ascertain incidence as well as to investigate the clinical characteristics and possible risk factors for maternal mortality due to meningitis in the Netherlands. We present the first comprehensive analysis of 15 cases of fatal maternal meningitis during pregnancy or puerperium. All cases were due to acute bacterial meningitis, except one with herpes meningo-encephalitis.

Methods

This study was designed as a retrospective cross-sectional study. Since 1983, all maternal deaths in the Netherlands have been reported to the MMC of the Dutch Society of Obstetrics and Gynaecology. Maternal death is defined by WHO as all deaths in pregnant women or up to and including 42 days post delivery, irrespective of the duration and site of the pregnancy, from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes.4

All cases between 1983 and 2005 were included in two successive nationwide Confidential Enquiries into Maternal Deaths in the Netherlands.5,6 During both study periods, a total of 489 (1983–92: 180; 1993–2005: 309) cases of maternal mortality were included in the Confidential Enquiries. We selected and analysed all cases of maternal meningitis. Details of the women and the course of events that preceded their deaths were extracted for all included deaths. We also retrieved cases of meningitis reported in 2006 and 2007 from the ongoing Enquiry and included them in our study. The maternal mortality ratio (MMR) was defined as the number of direct and indirect maternal deaths per 100 000 live births up to 42 days after the end of pregnancy.

Substandard care was assessed by all members of the MMC. Each panel member was requested to perform individual assessment of the patient notes using the standardised MMC audit form. Final decision of substandard care, which is used in this report, was reached by consensus. Substandard care was identified at the level of the woman, the care provider or the healthcare system.

Results

We retrieved 15 maternal deaths due to meningitis. The presenting symptoms of individual women and the causative organisms are shown in Table 1. From 1983 up to and including 1992 there were a total of 1 860 807 live births in the Netherlands, among which there were 180 maternal deaths reported to the MMC, of which none were the result of maternal meningitis. The overall MMR was 9.7 per 100 000 live births. From 1993 up to and including 2007 there were 2 923 601 live births in the Netherlands, with a total of 344 maternal deaths, 15 of which were the result of meningitis (4.4%).6 The overall MMR was 11.8 per 100 000 live births. The MMR due to meningitis was 0.5 per 100 000 live births. Twelve women (80%) presented with meningitis during pregnancy, 4 (27%) in the second trimester and 8 (66%) in the third trimester. Three women developed meningitis postpartum, with a maximum interval of 27 days. Mean age was 32.3 years (range 23–43 years), three women were from non-western countries, one of whom was HIV-positive and on antiretroviral therapy. None of the women had spinal anaesthesia or cranial surgery before developing meningitis.

Table 1. Presenting symptoms and causative organismThumbnail image of

The most frequent symptom was severe headache, present in 13 (87%) women. Other frequent symptoms were: altered mental status (defined by Glasgow Coma Scale score <15) in 11 (73%) women; fever, defined as a core temperature >38.0°C, in 9 (60%) women and nuchal rigidity, expressed as neck stiffness or a positive Brudzinski-sign, present in five (33%) women. The classical triad was present in 4 (27%) women with meningitis.

Of the 15 women with meningitis, nine (60%) showed symptoms of otolaryngological infection at presentation, one of whom was treated with antibiotics 3 weeks before admission. Five of these women presented with otitis, two complained of sinusitis. In one woman, blood cultures showed an infection due to Streptococcus milleri, a bacterium frequently found in patients with otitis. In one woman, magnetic resonance imaging (MRI) showed opacity of the os ethmoideum, subsequently an otolaryngological aetiology seemed evident.

At first presentation, 11 women (73%) underwent radiological examination using computed tomography (CT; nine women) or MRI (two women). Of these, six (40% of all women) showed distinct abnormalities leading to the diagnosis of meningitis and subsequent therapeutic action. The scans of three women (20%) showed no abnormalities; in two women (13%) CT or MRI revealed possible abnormalities leading to further testing by means of cerebrospinal fluid (CSF) examination, which eventually led to the diagnosis of meningitis.

On admission, lumbar puncture was performed in ten women (67%), eight of which (80%) showed infected CSF with the presence of white or red blood cells, microorganisms, or increased protein levels. Infections were proven with positive cultures in all women. The other two women had normal CSF and died as a result of herpes encephalitis and tuberculous meningitis.

Of the 12 women who developed meningitis in the antepartum period, seven women (58%) delivered by emergency caesarean section, resulting in live neonates. Apgar scores ranged from 3 to 9 (mean 5.9) and from 5 to 10 (mean 7.6), after 1 and 5 minutes, respectively. All umbilical cord blood pH examinations were above 7.10, all neonates were admitted to the neonatal intensive-care unit for observation and were discharged alive. Four women died before delivery; one delivered a stillbirth. All three women with postpartum meningitis had delivered live births vaginally.

Of the 15 women, 11 (73%) received broad-spectrum antibiotics following meningitis protocols. Substandard care was identified in four (27%) women. In two women the dosage or spectrum of coverage of the antibiotics initiated was insufficient for meningitis. In one woman puerperal psychosis was diagnosed so no antibiotics were given and in one case, delay on the part of the woman to seek medical attention possibly contributed to the fatal outcome. A selection of cases, representative of diagnostic difficulties and fatal outcome, is shown in Table 2.

Table 2. Selected cases of maternal deaths due to meningitis 1993–2007
Gestational age at diagnosis (weeks)Initial complaintPhysical and neurological examinationCSFRadiologyCourse of eventsDiagnosis
  1. AMS, altered mental status; CS, caesarean section; CTG, cardiotocogram; EEG, electroencephalogram; ICU, intensive-care unit; pp, postpartum.

202 weeks before admission received antibiotics for a sinusitis
Severe occipital headache and vomiting for 2 days
Temp: 39°C
Nuchal rigidity: +
AMS: +
TurbidCT-scan: no abnormalitiesDiff. diagnosis: Pneumococcal meningo-encephalitis, started with Ampicillin and Penicillin G
Progressive neurological deterioration, a repeat CT showed multiple infectious cerebral foci
EEG confirmed cerebral death: mechanical ventilation was stopped
Meningitis (S. pneumoniae)
29Severe headache and ear pain for 1 day
Within 3 hours decreased consciousness
Temp: 39°C
Nuchal rigidity: ?
AMS: +
TurbidCT-scan: pneumo-encephalitis and ethmoiditisDiff diagnosis: Pneumococcal encephalitis, started with antibiotics
Became comatose
Emergency CS: boy, Apgar score 5/7
Postoperative haemodynamically unstable, resuscitation failed, died in operating room
Meningo-encephalitis (S. pneumoniae)
34Severe headache for 2 days
One day of aphasia and motor dysfunction
Temp: 40°C
Nuchal rigidity: –
AMS: +
NormalCT-scan: possible abnormalities
(temporal swelling matching herpes encephalitis)
Diff diagnosis: Herpes encephalitis, started with Acyclovir, Amoxicillin, Claforan
Seizures successfully treated with Rivotril
Emergency CS: boy, Apgar score 9/10
Second CT scan: large intracerebral haemorrhage with elevated intracranial pressure
Mechanical ventilation stopped, woman died
Massive brain haemorrhage due to herpes encephalitis
35For 3 weeks ear pain because of otitis media
Paracentesis, after which developed decreased consciousness the same day
Temp: 38.7°C
Nuchal rigidity: +
AMS: +
TurbidCT-scan: no abnormalitiesDiff diagnosis: meningitis, started with Dexamethasone and Penicillin G
Abnormal CTG: emergency CS: boy Apgar score 6/8
Intubated and lumbar puncture performed; transferred to ICU
At ICU: EEG showed no activity after which treatment was stopped
Meningitis (S. pneumoniae)
10 days ppHeadache since delivery, slight response to NaproxenTemp: ?
Nuchal rigidity: –
AMS: +
Not performedCT-scan: post delivery, showed no abnormalitiesAt discharge received Naproxen for headache
10 days postpartum admission to psychiatric institute for puerperal psychosis
Died in her sleep while admitted
Meningitis (Cryptococcus sp.)

A postmortem was performed in only one (7%) woman. In the remaining women, cause of death was established based on physical examination, CSF or blood cultures and radiological examinations. The most frequent causative agent in our study population was Streptococcus pneumoniae, which was present in ten women (67%).

Discussion

This article describes 15 maternal deaths due to meningitis, which is an important cause of maternal mortality with a possible delay in diagnosis because of diversity of symptoms and reluctance for invasive testing during pregnancy.

In a series of 352 episodes of community-acquired meningitis in the general population by Weisfelt et al.,7 30% of patients died within 2 weeks. To our knowledge there are no publications on prevalence of meningitis in pregnant or puerperal women, therefore it is not known whether meningitis is more frequent or has a more fulminant course in this subgroup. A recent study found that although S. pneumoniae infections appeared to be about threefold more frequent in pregnant than in nonpregnant women, the duration and severity were similar in both populations.8

It is interesting that there were no reported deaths from meningitis in the first enquiry, whereas there was an increase in the second period. Whether this is a true increase in maternal meningitis, a consequence of earlier under-reporting, or coincidence is unclear. Under-reporting appeared to be 26% in the first period when compared with official statistics by Statistics Netherlands.5 From 1993 onwards, following the new WHO Guidelines, pregnancy status has been registered on death certificates in the Netherlands.4 In the second enquiry, using capture–recapture analysis by cross-matching data with Statistics Netherlands, under-reporting of maternal mortality was calculated at 33%.6 Apart from these changes, methods of registering and analysing maternal deaths have been comparable throughout the enquiries. Reviewing statistics from the UK Confidential Enquiries into Maternal Deaths, there was a concurrent tendency towards an increase in cases of meningitis: MMR rose from 0.09 per 100 000 live births in 1985–87 to 0.24 in 2003–05.9–15 However, this MMR dropped in a recent UK report to 0.13. Interestingly, no cases followed ear infections.16 Our MMR, in comparison, is almost four times as high. We hypothesise that hesitation on the part of Dutch physicians to administer antibiotics for ear infections potentially contributed to a higher incidence of acute bacterial meningitis.

In the antepartum cases, the interval between onset of symptoms and death was short, with a maximum of 48 hours. The most common early symptoms are headache and fever, symptoms that are usually caused by more prevalent, innocuous forms of infectious disease such as influenza. This analysis confirms that, in concordance with the general population, the classical triad is often only partly present at first presentation in pregnant women. Subsequently, diagnosis must be established with other symptoms like headache or ear pain and findings on radiological and CSF examination. In practice there is often reluctance to perform CT scans on pregnant women because of the fear of radiation effects on the fetus. It is important to acknowledge that the risk of delayed diagnosis and treatment of meningitis poses a much greater risk to the mother and child than the possible side effects of a CT scan. As stated by the American College of Obstetricians and Gynecologists: ‘Concern about possible effects of high-dose ionising radiation exposure should not prevent medically indicated diagnostic x-ray procedures from being performed on a pregnant woman’.17 Still, as demonstrated by this study, negative findings on CT or MRI scan do not exclude meningitis.

Examination of CSF is regarded as the reference standard in diagnosing meningitis. In our study, two women had a normal CSF examination, but ultimately died from infectious meningitis. It seems that a negative CSF examination at first presentation cannot fully exclude meningitis; in one woman a second lumbar puncture was positive for herpes virus.

In almost two-thirds of the women with meningitis, otolaryngological infections were the site of origin of the infection. This finding is supported by Van de Beek et al.1 in a retrospective cohort study in the general population. Apparently, in pregnant women with signs of the classical triad or with headache plus an otolaryngological infection at presentation or in their recent history, the differential diagnosis should include meningitis. This is supported by the analysis of substandard care, showing that critical to identifying women with meningitis is a high level of suspicion. Also, adequate follow up of otolaryngological infections, especially when the woman does not recover, could help to lower the incidence and adverse outcome of meningitis.

A Medline search, combining synonyms for both meningitis and pregnancy, revealed 148 articles, with a total of 202 maternal deaths due to meningitis. Most frequent causes were tuberculous meningitis (34 women; 17%) and herpes virus meningitis (26 women; 13%). Most articles were case reports, either from low-income countries or from the previous century. After excluding those articles, the most common mechanism of infection reported was neuraxial analgesia during labour, in 32% of women (23 of 71). Iatrogenic meningitis is a rare but clinically important entity that is estimated to occur in one in every 5000 people subjected to dural puncture.18,19 In this subgroup, meningitis was mostly caused by oropharyngeal or dermal commensal bacteria. In our study, only three women developed meningitis postpartum, none of whom received epidural or spinal analgesia during delivery.

Our study has some limitations. First, some important clinical data were missing because of the retrospective nature of the study. Second, we had no comparable data on women with meningitis during pregnancy who did not die. In a nationwide study into severe maternal morbidity and mortality in which all 98 maternity units in the Netherlands participated during a 2-year period from August 2004 until August 2006 (LEMMoN), there were eight women with meningitis.20 Five of these women survived: two developed meningitis after mechanical ventilation for pneumonia; two presented de novo with fever and headache; one woman had a ventricular peritoneal drain inserted for hydrocephalus, resulting in meningitis. Using these small numbers, the case fatality rate would be as high as 38% (three of eight). This emphasises the need for early diagnosis and treatment. Third, it is possible that maternal mortality due to meningitis during the puerperium is under-reported to the MMC because these women are primarily treated by nonobstetric specialists who are not acquainted with this Committee. However, most cases develop during pregnancy and will therefore have been included.

Conclusion

This study shows that meningitis is emerging as an important cause of maternal death in the Netherlands. There were no cases of maternal death attributed to meningitis reported between 1983 and 1992; however, it accounted for 4.4% of MMR in the subsequent 15 years. The current rate is almost four times the rate in the UK. The reason for this difference is unknown. Meningitis may present with vague symptoms and could be missed easily. Therefore it is important to exercise a high index of suspicion to avoid delay in the diagnosis and treatment with the resultant increase in fatalities.

A full and thorough investigation, including radiological or CSF examinations should be considered in pregnant or puerperal women presenting with classical symptoms of meningitis, particularly those with a recent or current otolaryngological infection or severe headache. The primary goal should be to exclude meningitis. A CT or MRI scan should not be omitted because of presumed radiation effects to the unborn child. Early diagnosis and immediate antibiotic treatment are imperative because of rapid deterioration in pregnant women. In case of doubt, the threshold for antibiotic treatment should be low and close monitoring is warranted.

Disclosure of interests

The authors have no potential conflicts of interest to declare.

Contribution to authorship

TPS analysed the collected data and extracted relevant data; performed the literature search and selected material for the literature review; reviewed the literature and compared it with the data obtained from the Confidential Enquiry into Maternal Deaths; interpreted the analysis and, together with JvR, wrote the manuscript. JMS and NWES, members of the Dutch MMC, and JJZ reviewed and corrected the manuscript. JvR, the Chairman of the Dutch MMC, was involved in setting up the study; collected data for the confidential enquiry; reviewed the collected cases and reviewed and corrected the manuscript.

Details of ethics approval

The Confidential Enquiry into Maternal Deaths involves assessment of anonymised reported data. In the Netherlands, ethical approval is not required for a confidential enquiry.

Funding

No funding was received for this publication.

Acknowledgements

The authors thank all members of the Dutch MMC and everyone who reported cases of maternal mortality.

Ancillary