Bacterial meningitis continues to be a potentially life-threatening disease worldwide, with an estimated annual incidence of 4-6 cases per 100,000 population.1,2 It is widely known that spontaneous bacterial meningitis (SBM) in adult patients, with the notable exception of meningococcal meningitis, tends to occur in those patients who are immunosuppressed or are aged.3 Bacterial meningitis has rarely been described complicating HIV-1 infection, usually in the form of case reports and a few retrospective case series.4,5 We have performed a prospective cohort study in adult patients with SBM and HIV-1 infection.
The aim of our work was to assess the burden that SBM represents in the setting of HIV-1 infection. The study was performed taking as a basis the wide epidemiological environment represented by the Spanish VACH (VIH-Aplicación de Control Hospitalario) Cohort Study Group.
PATIENTS AND METHODS
The Spanish VACH Cohort and Data Collection
The Spanish VACH Cohort and its way of collecting data have been described elsewhere.6
To be included in the present study, patients belonging to the VACH Cohort had to be diagnosed as having SBM during the study period (1996-2006). The diagnosis of SBM was based on the presence of consistent clinical findings together with: (1) isolation and identification of a pathogenic bacterium from cerebrospinal fluid (CSF) culture, excluding Mycobacterium tuberculosis, and (2) isolation and identification of a pathogenic bacterium from blood culture together with CSF cytobiochemical findings consistent with purulent meningitis. Bacterial meningitis of unknown etiology was diagnosed on the basis of CSF cytobiochemical findings consistent with purulent meningitis together with negative CSF and blood cultures. The diagnosis of purulent meningitis was made on the basis of CSF pleocytosis of more than 5 polymorphonuclear leukocytes per cubic millimeter, hypoglycorrachia defined as a blood to CSF glucose ratio <0.50, and hyperproteinorrachia of more than 0.45 g/L. Post-neurosurgical bacterial meningitides were not included in this study.
The population included in the study was divided into 3 groups:
SBM in HIV-1-Infected Patients
Patients with documented serologic evidence of HIV-1 infection who had been diagnosed as having SBM after the previously stated criteria.
SBM Control Group
Patients belonging to the same age groups of HIV-1-infected patients without serologic evidence of HIV-1 infection who had been diagnosed as having SBM during the study period in the hospitals belonging to the VACH Cohort.
HIV-1 Infection Control Group
Patients with documented serologic evidence of HIV-1 infection, belonging to the VACH Cohort, who had not been diagnosed as having SBM during the study period.
Plasma HIV-1 RNA was measured using quantitative reverse transcriptase-polymerase chain reaction (Amplicor 1.5; Roche Molecular Systems, Branchburg, NJ). The lower limit of detection of these assays is 50 copies per milliliter. For values below the limit of detection, a value of 49 copies per milliliter was imputed. Peripheral blood CD4+ T-cell counts were measured using standard flow cytometry techniques. CSF Gram-stained smear and cultures were performed with the use of standard procedures. The identification of pathogenic isolates was performed with the use of standard techniques. Antibiotic susceptibility testing was performed by means of disk diffusion in agar, according to the guidelines established by the National Committee for Clinical Laboratory Standards.7
Clinical Criteria and Definitions
Predisposing conditions, complications, treatment, and outcome have been defined elsewhere.8
Continuous variables were assessed with Student t test for normally distributed data, or the Mann-Whitney U test otherwise, and the Fisher exact test for categorical data. Incidences, relative rates (RRs), and their 95% confidence interval (CI) were calculated by means of Poisson regression models. The analysis was performed using SAS version 9.1.3 software (SAS Institute, Inc, Cary, NC), and the level of significance was established at the 0.05 level (2 sided).
From 1997 to 2006, 1371 episodes of SBM in adult patients were diagnosed in the participating hospitals for an incidence rate (95% CI) of 3.2 (3.0 to 3.4) cases per 100,000 patient-years. The VACH Cohort hospitals had a population of 13,187 HIV-1-infected patients under active follow-up. Among them, 32 patients had SBM, for an overall annual incidence of 62.0 (43.8 to 87.7) cases per 100,000 patient-years. Therefore, the relative risk (95% CI) of having SBM was 19.4 (13.6 to 27.5) times greater for an HIV-1-infected patient than for a non-HIV-1-infected individual (P < 0.001). HIV-1-infected patients with SBM represented 2.33% (95% CI: 1.61% to 3.28%) of the total adult patients diagnosed as having SBM during the study period in the VACH Cohort hospitals.
Among HIV-1-infected patients with SBM, there were 22 men (68.7%) and 10 women (31.3%), with a mean age of 29.2 ± 8.8 years (range: 15-51 years). Two hundred sixty-seven non-HIV-1-infected patients of the same age groups, diagnosed during the same study period, were used as a comparator group. The characteristics of patients belonging to this group were not different from the whole group of 1371 patients with SBM. Demographic characteristics of both groups of patients are shown in Table 1.
Predictors for Acquiring SBM in HIV-1-Infected Patients
The following factors were associated with the occurrence of SBM: last CD4 count greater than 200/mm3, with a protective risk [RR (95% CI): 0.11 (0.05 to 0.23), P < 0.0001], and viral load at entry [RR per 1-log10 increase: 1.35 (0.98 to 1.86), P = 0.0621], although statistical significance was not reached for the latter. In the multivariate analysis, the CD4 count was still the strongest predictor (P < 0.0001) and no other variable was significant when the CD4 count was included in the model. When adjusted by last CD4 count [RR (95% CI): 0.11 (0.05 to 0.23), P = 0.0026], the vaccine was unprotective [RR (95% CI): 0.69 (0.27 to 1.81) P = 0.4556].
Overall, more than a third of HIV-1-infected patients (37.5%) had conditions, other than HIV-1 infection itself, predisposing them to acquiring SBM. This figure was 24.7% for uninfected persons (Table 1). The most common for both groups of patients was alcohol abuse, followed by hepatic cirrhosis and cancer (Table 1). HIV-1-infected patients with SBM were severely immunosuppressed with a median CD4 cell count of 128/mm3 when the meningitis was diagnosed. In fact, almost 80% of the patients had a CD4 cell count <200/mm3. None of the patients who had SBM was on prophylaxis with co-trimoxazole, whereas 396 (3.0%) who had not developed it were on prophylaxis (P = 0.3190). Only half of the patients were on antiretroviral therapy, and more than 40% had developed an AIDS-defining condition (Table 1). Interestingly, 8 HIV-1-infected patients (25.0%) had had prior AIDS-defining conditions that affected the central nervous system (CNS): cerebral toxoplasmosis (4), disseminated tuberculosis (2), multifocal progressive leukoencephalopathy (1), and cryptococcal meningitis (1). However, in 6 patients (18.7%), SBM was the infection that prompted the diagnosis of HIV-1 infection. A third of patients in both groups had taken antibiotics, other than co-trimoxazole, before the diagnosis of SBM, whereas pneumococcal vaccination was uncommon in both groups of patients (Table 1).
The presence of a primary extrameningeal focus of infection was significantly more frequent in HIV-1-infected patients than in uninfected ones [odds ratio (OR) = 3.21; 95% CI: 1.36 to 7.54, P = 0.005] (Table 2), and this was especially true for the presence of pneumonia. If meningococcal meningitis was excluded, there were no differences between both groups with respect to the presence of extrameningeal infection (OR = 1.03; 95% CI: 0.40 to 2.60, P = 0.08). However, pneumonia continued to be statistically more frequent in HIV-1-infected patients (OR = 3.96; 95% CI: 1.17 to 13.29, P = 0.02). There were no significant differences between HIV-1-infected and uninfected patients with respect to clinical signs, except for a significantly higher prevalence of focal neurologic signs (OR = 3.46; 95% CI: 1.38 to 8.57, P = 0.005) and seizures (OR = 2.91; 95% CI: 0.85 to 9.37, P = 0.059) upon admission among HIV-1-infected patients (Table 2). On the other hand, the presence of nuchal rigidity was significantly less frequent among HIV-1-infected patients (OR = 0.41; 95% CI: 0.16 to 1.08, P = 0.0053). The level of consciousness on admission to the hospital was not significantly different between both groups of patients, although there was a trend to a worse level of consciousness on admission for HIV-1-infected patients. There were no differences between HIV-1-infected and non-HIV-1-infected patients with respect to CSF cytobiochemical parameters except for a significantly lower CSF to blood glucose ratio for those with HIV-1 infection (Table 2).
The most frequent etiologic agent of SBM in HIV-1-infected patients was Streptococcus pneumoniae, whereas in non-HIV-1-infected patients, it was Neisseria meningitidis. Listeria monocytogenes ranked third in both groups of patients (Table 2). There were no statistically significant differences between both groups with respect to positivity of CSF Gram-stained smear or CSF culture. However, HIV-1-infected patients had a significantly higher prevalence of bacteremia than non-HIV-1-infected patients (OR = 2.69; 95% CI: 1.18 to 61.18, P = 0.016) (Table 2).
Neurologic complications were significantly more frequent in HIV-1-infected patents than in uninfected ones (OR = 3.03; 95% CI: 1.17 to 7.68, P = 0.02), whereas extra-neurologic complications were not significantly different between both groups (Table 3). Adequacy of empiric antibiotic therapy was significantly lesser for HIV-1-infected patients (OR = 3.67; CI: 0.90 to 13.94, P = 0.0502). The overall case fatality ratio was significantly higher for HIV-1-infected patients than for those noninfected (OR = 4.33; 95% CI: 1.55 to 11.90, P = 0.004). However, when meningococcal meningitis was excluded from the analysis, although the mortality rate for HIV-1-infected patients doubled that of uninfected patients (24.2% vs. 12.4%), the difference became nonsignificant (OR = 2.25; 95% CI: 0.73 to 6.80, P = 0.14). Death was attributable to SBM in 6 of 8 HIV-1-infected patients and in 13 of 19 non-HIV-1-infected patients (P = 0.85). Mortality among HIV-1-infected patients was unrelated to the prior diagnosis of AIDS, current highly active antiretroviral therapy, prior CNS infection, predisposing conditions, extrameningeal infection, preadmission antibiotic therapy, focal neurologic signs or seizures on admission, CSF parameters, bacteremia, and adequacy of empiric antibiotic therapy. Case fatality ratio was associated with the appearance of neurologic complications (OR = 8.33; 95% CI: 1.07 to 77.69, P = 0.02), with extra-neurologic complications (OR = 14.40; 95% CI: 0.95 to 454.80, P = 0.01), and with the presence of pneumonia as primary extrameningeal infection (OR = 11.00; 95% CI: 1.11 to 138.95, P = 0.02). Patients with a CD4 cell count ≥200/mm3 had a case fatality rate of 14.3% compared with 28.0% for those with a CD4 cell count <200/mm3 (P = 0.65).
Neurologic sequelae on discharge from the hospital were present in 29% of HIV-1-infected patients and in 7% of those not infected by HIV-1 (OR = 3.50; 95% CI: 1.19 to 10.07, P = 0.01) (Table 3). The most common were neurocognitive deficits (2), hypoacusia (2), vegetative state (1), ataxia (1), and blindness (1). Neurologic sequelae were only related to the presence of neurologic complications (OR = 22.67; 95% CI: 1.39 to 801.74, P = 0.003). When meningococcal meningitis was excluded from the analysis, neurologic sequelae were still significantly more frequent in HIV-1-infected than in uninfected patients (31.8% vs. 13.1%, P = 0.0456).
Our study shows that SBM in HIV-1-infected patients is much more frequent than in the general population of the same age groups, which it usually presents in severely immunosuppressed patients who often are not taking antiretroviral therapy and who have not been previously vaccinated. Furthermore, S. pneumoniae is usually the causative agent with pneumonia or ear, nose, and throat infections as primary foci of infection. Evolving features include a higher number of neurologic complications and a poorer outcome than non-HIV-1-infected patients with higher case fatality rates and higher number of neurologic sequelae.
Like many other bacterial diseases, the incidence of SBM is higher in HIV-1-infected patients than in uninfected persons. This may be explained by the disturbances of humoral immunity present in patients with HIV-1 infection.9 A significant impairment of B-cell activation occurs, predominantly affecting the synthesis of IgM, both T-cell independent and T-cell dependent differentiation of B cells is impaired, resulting in a decreased specific humoral response to bacterial antigens, and polyclonal B-cell activation occurs that results in polyclonal hypergammaglobulinemia.9,10 These disturbances tend to be more pronounced in the advanced stages of HIV-1 infection. Therefore, not surprisingly, almost 80% of the episodes of SBM appeared in patients with less than 200 CD4 cells per cubic millimeter.
The most important difference between immunocompetent and HIV-1-infected patients is in the etiology of SBM, with S. pneumoniae as the most common etiologic agent for HIV-1-infected patients, whereas N. meningitidis has ranked the first in our epidemiologic environment until recently. However, the etiologic spectrum of SBM is highly dependent, even in HIV-1-infected patients, on the geographic location as the studies in sub-Saharan Africa demonstrate.11,12
When comparing our series with the other 2 reported, similar findings with respect to the etiologic spectrum of SBM in HIV-1-infected patients were obtained.4,5 From a clinical point of view, the presentation of SBM in HIV-1-infected patients is undistinguishable from that of uninfected persons, although among the former, there is a greater frequency of pneumonia as a primary extrameningeal infection, focal neurologic signs on admission, a more altered level of consciousness on admission, and a lower CSF to blood glucose ratio. Most of these findings may be explained by a higher prevalence of pneumococcal meningitis, but even when meningococcal meningitis was excluded from the analyses, there were still differences in these parameters between HIV-1-infected and non-HIV-1-infected patients. An additional factor could be a more labile neurologic function in the case of HIV-1-infected patients that may be due to HIV-1 infection (through neurologic damage caused by HIV-1 itself) or due to prior CNS infections suffered by these patients.
There is some discrepancy on the incidence rate of invasive pneumococcal disease in HIV-1-infected patients across the highly active antiretroviral therapy era, a decline found in surveillance studies,13,14 whereas some cohort studies suggest a stable incidence.15,16 Even in the best possible scenario, the residual burden of invasive pneumococcal disease is still 35 times higher for HIV-1-infected than for uninfected patients.17 This higher incidence of pneumococcal disease is also expressed by a high incidence of pneumonia, frequently bacteremic, which in the present study was significantly more frequent in HIV-1-infected patients presenting with SBM and was associated with a poor outcome. Data regarding the protective efficacy of pneumococcal vaccines in HIV-1-infected patients are also somewhat controversial. Case-control studies among HIV-1-infected people in developed countries who have access to antiretrovirals have generally shown protection from invasive disease among those who have received the 23-valent pneumococcal vaccine.17,18 However, the only randomized trial of this vaccine among HIV-1-infected people without access to antiretrovirals (like 50% of our HIV-1-infected population with SBM) showed an increase in pneumonia among vaccinees, and a 6-year follow-up of that study showed no further increase in pneumonia but a paradoxical significant 16% reduction in mortality in the vaccinated group.19 Our data do not let us to draw any conclusion about its protective effect in the setting of SBM.
A prior study on SBM in HIV-1-infected patients in Spain has suggested that its prognosis is similar to that of uninfected patients.5 However, in developing countries, its prognosis is ominous with mortality rates ranging from 55% for meningitis from a mixed etiology group4,12 to 65% for pneumococcal meningitis.11 Among our patients, complications, death rate, and sequelae were significantly more frequent for HIV-1-infected patients, and even when meningococcal meningitis was excluded from the analyses, the case fatality rate doubled that of non-HIV-1-infected patients.
In summary, SBM in HIV-1-infected patients has a higher incidence than among the general population, affects unvaccinated and patients who are off antiretroviral therapy, is caused by S. pneumoniae in half of the cases, and carries a bad prognosis both in terms of lethality and appearance of postmeningitic sequelae.
The authors acknowledge the assistance of Richard Pike in reviewing English language.
1. Schlech WF, Ward JI, Band JD, et al. Bacterial meningitis
in the United States, 1978 through 1981: the national bacterial meningitis
surveillance study. JAMA
2. van de Beek D, de Gans J, Tunkel AR, et al. Community-acquired bacterial meningitis
in adults. N Engl J Med
3. van de Beek D, de Gans J, Spanjaard L, et al. Clinical features and prognostic factors in adults with bacterial meningitis
. N Engl J Med
4. de Almeida S, Savalla G, Mendez B, et al. Acute bacterial meningitis
in HIV patients in southern Brazil. Arq Neuropsiquiatr
5. Almirante B, Saballs M, Ribera E, et al. Favorable prognosis
of bacterial meningitis
in patients infected with the human immunodeficiency virus. Clin Infect Dis
6. Domingo P, Suarez-Lozano I, Teira R, et al. Dyslipidemia and cardiovascular disease risk factor management in HIV-1-infected subjects treated with HAART in the Spanish VACH Cohort. Open AIDS J
7. National Committee for Clinical Laboratory Standards. Performance Standards for Antimicrobial Susceptibility Testing. Twelve Informational Supplement M102-S12
. Wayne, PA: NCCLS; 2002.
8. Domingo P, Barquet N, Alvarez M, et al. Group B streptococcal meningitis in adults. Report of twelve cases and review. Clin Infect Dis
9. Fauci AS, Pantaleo G, Stanley S, et al. Immunopathogenic mechanisms of HIV infection. Ann Intern Med
10. Lane HC, Depper JM, Greene WC, et al. Qualitative analysis of immune function in patients with the acquired immunodeficiency syndrome. Evidence for a selective defect in soluble antigen recognition. N Engl J Med
11. Gordon SB, Chaponda M, Walsh AL, et al. Pneumococcal disease
in HIV-infected Malawian adults: acute mortality and long-term survival. AIDS
12. Scarborough M, Gordon SB, Whitty CJ, et al. Corticosteroids for bacterial meningitis
in adults in sub-Saharan Africa. N Engl J Med
13. Grau I, Pallarés R, Tubau F, et al. Epidemiologic changes in bacteremic pneumococcal disease
in patients with human immunodeficiency virus in the era of highly active antiretroviral therapy. Arch Intern Med
14. Heffernan RT, Barrett NL, Gallagher KM, et al. Declining incidence of invasive Streptococcus pneumoniae
infections among persons with AIDS in an era of highly active antiretroviral therapy, 1995-2000. J Infect Dis
15. Jordano Q, Falcó V, Almirante B, et al. Invasive pneumococcal disease
in patients infected with HIV: still a threat in the era of highly active antiretroviral therapy. Clin Infect Dis
16. Barry PM, Zetola N, Keruly JC, et al. Invasive pneumococcal disease
in a cohort of HIV-infected adults: incidence and risk factors, 1990-2003. AIDS
17. Klugman KP, Madhi SA, Feldman C. HIV and pneumococcal disease
. Curr Opin Infect Dis
18. Peñaranda M, Falcó V, Payeras A, et al. Effectiveness of polysaccharide pneumococcal vaccine
in HIV-infected patients: a case-control study. Clin Infect Dis
19. Watera C, Nakiyingi J, Miiro G, et al. 23-valent pneumococcal polysaccharide vaccine in HIV infected Ugandan adults: 6-year follow-up of a clinical trial cohort. AIDS