Adult meningitis in a setting of high HIV and TB prevalence: findings from 4961 suspected cases

South Africa has the highest burden of human immunodeficiency virus (HIV) disease in the world, with an estimated 5.5 million cases accounting for 17% of the global total [1]. This is associated with one of the world's highest tuberculosis (TB) incidence rates [1]. The presentation and causes of adult meningitis have changed substantially as a result of these ongoing epidemics [2, 3]. Advanced HIV-infection can make interpretation of both clinical and laboratory findings difficult, hence knowledge of aetiology and laboratory findings in patients presenting with meningitis are important in guiding management. We performed a retrospective study to determine these findings in a setting of high HIV and tuberculosis prevalence in Cape Town.

The study was approved by the Research Ethics Committee of the University of Cape Town, and performed at GF Jooste Hospital, a public sector adult referral hospital serving a population of 1.3 million. Forty-three percent of patients admitted to the medical wards have proven or clinically suspected HIV [4], and TB incidence in the referral area exceeds 1000/100,000 person-years [5].

This is the largest reported series of adult meningitis in South Africa. CM and TBM are the commonest causes of adult meningitis in our setting (63% and 27% of microbiological diagnoses respectively). This is in marked contrast to the pre-HIV era [6, 7], and in keeping with recent data from central and southern Africa showing that cryptococcosis is now the leading cause of adult meningitis, accounting for 27-45% of all cases [2, 3, 8–11]. As in these previously reported series, CM in our study was almost exclusively in the context of advanced HIV disease. Of note, during the study period approximately one third of initial episodes of CM occurred in patients already on ART [12], so that both initial and relapse episodes of CM were a mix of patients on and not on ART. This may explain why the proportion of cases diagnosed on India Ink and culture were lower, and those diagnosed by CrAg higher, than prior series, and also why CSF indices did not differ between initial and relapse episodes.

The proportion of TBM in our study was higher than in most reported cohorts from central and southern Africa, which range from 1-17% [2, 3, 8, 10], however similar to figures from two small South African studies. TBM accounted for 25.4% of all adult meningitis in Soweto in 1996 [13], and increased from 16% in 1994 to 31% in 1998 in Pretoria [14]. The proportion of TBM in our study is almost certainly an under-estimate. TBM is under-diagnosed by routine laboratory investigation, with reported rates of microbiological confirmation ranging from 25 to 70% [15]. In a recent Vietnamese study, microbiological confirmation of TBM was obtained in 45% of HIV-positive patients, versus 33% of HIV-negative patients [16]. Assuming similar confirmation rates in our setting, it is likely that a substantial number of the 539 cases with no microbiological diagnosis, and lymphocytic CSF, represented TBM; an assumption supported by the very high community TB rates in our setting [5].

Further reasons for possible under-diagnosis of TBM were that not all CSF samples were sent for TB studies, volumes of CSF examined were low (typically 1-2 mls), and time for smear microscopy was limited. There is good evidence that diagnostic yields can be improved by collection of large CSF volumes for culture, and careful microscopic examination of centrifuged samples [17, 18]. Smear positivity rates of 58-69% have been reported from Vietnam, where up to 10 mls of CSF was routinely obtained [17, 18], and efforts are needed to improve the diagnosis of TBM in our setting.

Numbers of bacterial meningitis were relatively low. It is possible that prior antibiotic exposure hindered diagnosis in some cases, however numbers of sterile pyogenic CSFs were low. Numbers of bacterial meningitis appear to have stayed more or less constant despite the evolving HIV epidemic [14], but with a marked switch in aetiology. Meningococcus, which previously predominated, causing ~30% of adult bacterial meningitis [6, 7], accounted for only 3% in our series, with Pneumococcus accounting for over 90% of bacterial meningitis.

The study highlights the fact that HIV co-infection, as well as changing the causative organisms, can lead to atypical laboratory presentations of meningitis. Substantial numbers of patients with CM, TB and bacterial meningitis had normal or nearly normal CSF findings. Of particular note were the findings in bacterial meningitis, almost all of which was HIV-associated. Four (6%) of 68 patients had acellular CSF, 30 (44%) of 68 had a lymphocyte predominance, and the majority had very elevated CSF protein. This adds to the diagnostic difficulties posed by the often overlapping clinical features of meningitis a population with a high HIV prevalence, further compounded by the presence of dual pathology in a small proportion of cases (table).

A limitation is that, as a laboratory based study, detailed clinical information was lacking. CM is unlikely to be missed, however, as discussed, TBM is likely to under-diagnosed by laboratory investigation, and this caveat must be taken into account when interpreting these data.

The HIV and TB epidemics in South Africa have led to a marked change in the aetiology of adult meningitis. Cryptococcus has emerged as the dominant causative agent, the proportion of meningitis due to TB is increasing, and the spectrum of bacterial meningitis has been markedly altered. High rates of HIV in a population may make interpretation of CSF findings difficult, and it is important to take these data into account when investigating and managing patients with suspected meningitis. Mortality of both cryptococcal and TB meningitis remains high in this setting [11, 18], and these results underscore the importance of continuing efforts to improve the early diagnosis and treatment of cryptococcal and TB meningitis.