No evidence that vaccination with a polysaccharide pneumococcal vaccine protects drug users against all-cause pneumonia

Abstract

Individuals infected with HIV are at high risk of pneumonia and invasive disease caused by Streptococcus pneumoniae[1]. To date, the effect of vaccination with a polysaccharide pneumococcal vaccine on this group is unclear. Breiman et al.[2] suggested a protective effect of vaccination on the incidence of invasive disease, whereas a recent study in Uganda [3] showed an increased risk of invasive disease and no protection against all-cause pneumonia among vaccinees. Since November 1994, pneumovax (Pasteur Merieux MSD, Brussels, Belgium) has been offered to drug users in the Amsterdam Cohort Studies (ACS) who have a history of pneumonia, an asthmatic constitution, or HIV-positive serostatus, because these conditions (besides intravenous drug use) increase the risk of pneumonia [4]. In the ACS, immunological, virological, and clinical data are collected at standardized cohort visits every 4 months. Physicians decide whether an anamnestic event is a pneumonia based on symptoms, diagnostics, and treatment. We studied the effect of pneumovax on the incidence of all-cause pneumonia in a group of 48 drug users who seroconverted to HIV during ACS follow-up. Subjects were part of a clinical follow-up project in which immunological, virological and clinical data, in addition to ACS data, were collected in hospitals in Amsterdam. In the clinical follow-up, data were collected retrospectively and prospectively until 1 January 1998. To be included in our study and be eligible for vaccination, individuals had to seroconvert to HIV after ACS entry and continue follow-up after 1 November 1994. The study period ran from ACS entry until death or 1 January 1998. Our study population consisted of 33 men and 15 women, whose mean age at ACS entry was 29.4 years (SD 6); 92% were of west European origin. Of the 48 subjects, 12 refused vaccination and two had medical contraindications; for two subjects, the reason for not receiving a vaccination was unknown. The remainder of 32 subjects received pneumovax at a mean age of 36.2 years (SD 6). Of these, 31 individuals were vaccinated after seroconversion to HIV; one individual was vaccinated before that date. Five subjects died during follow-up. After July 1996, when highly active antiretroviral therapy became the standard of care, seven participants received this therapy: at least one protease inhibitor or non-nucleoside reverse transcriptase inhibitor, combined with nucleoside reverse transcriptase inhibitors. To compare the baseline characteristics of vaccinees with non-vaccinees, we considered vaccinees eligible for vaccination at the date of vaccination. Those who did not receive vaccination were considered to be eligible for vaccination on 1 November 1994 if they had seroconverted before that date (n = 36), or at their first HIV-positive visit if they seroconverted after 1 November 1994 (n = 12). At the date they were eligible for vaccination, vaccinees (n = 32) and non-vaccinees (n = 16) were comparable in age, years since seroconversion, age at seroconversion, CD4 cell count, episode(s) of pneumonia before seroconversion, the use of antiretroviral therapy (according to the intention-to-treat principle), and the number of drug injections since the last ACS visit. Therefore we assumed the two groups to be at the same risk of pneumonia. Poisson regression was used to estimate the incidence of pneumonia and to calculate the relative risks (RR) with 95% confidence intervals (CI). Evaluated as potential risk factors were vaccination, sex, years since seroconversion (HIV negative; ≤ 2; > 2 and ≥ 5; > 5), age at seroconversion (≤ 29; > 29 and ≤ 35; > 35), HIV serostatus with CD4 cell counts (HIV-negative; HIV-positive, CD4 count > 200 cells/mm3; HIV-positive, CD4 count ≤ 200 cells/mm3), and prophylaxis against Pneumocystis carinii pneumonia (according to the intention-to-treat principle). All risk factors were treated as time-dependent variables except for sex and age at seroconversion. During the study period, 65 episodes of pneumonia were documented (incidence 0.15 per person-year; 95% CI 0.12; 0.20), three with a sputum culture positive for S. pneumoniae and five positive for another identifiable pathogen. In 57 cases, the causative pathogen was unknown. In univariate analysis, vaccinees had a non-significantly higher risk of pneumonia than the non-vaccinees (RR 1.36; 95% CI 0.76; 2.46). Compared with HIV-negative individuals, the risk of pneumonia was significantly higher among HIV-positive individuals (RR 3.00; 95% CI 1.34; 6.71) and among HIV-positive individuals with a CD4 cell count of 200 cells/mm3 or less (RR 3.69; 95% CI 1.40; 9.69) (Table 1). In comparison with HIV-negative individuals, those who had seroconverted more than 2 years previously had a significantly increased risk of pneumonia (RR 3.82; 95% CI 1.46; 10.00). In multivariate analysis, HIV-positive serostatus (RR 3.00; 95% CI 1.32; 6.79) and HIV-positive serostatus with a CD4 cell count of 200 cells/mm3 or less (RR 3.68; 95% CI 1.36; 9.95) remained significant risk factors. However, the RR for pneumonia among vaccinees decreased to 1.01 (95% CI 0.53; 1.91). To avoid recall-biased data and misclassification during ACS visits we repeated the analysis, including only pneumonia documented in hospital charts (n = 21), and found an unadjusted RR of 1.99 (95% CI 0.77; 5.12) for vaccinees (data not shown). HIV-positive serostatus with a CD4 cell count of 200 cells/mm3 or less was significantly associated with pneumonia in univariate analysis (RR 3.87; 95% CI 1.09; 13.73). In multivariate analysis (RR for vaccinees 1.47; 95% CI 0.48; 4.55), no significant risk factors were found, possibly because of the small numbers. To correct for dependency among the visits of each subject, the model was re-analysed using generalised estimation equations, and the results were comparable. In the multivariate models, no statistically significant interaction terms were found.Table 1: Uni- and multivariate associations between risk factors and all-cause pneumonia in a group of 48 HIV seroconverters from the Amsterdam Cohort Studies. This study confirms that HIV-positive serostatus and a deteriorating immune system are associated with a significantly higher risk of all-cause pneumonia [4]. Our results suggest that vaccination with pneumovax has no protective effect against all-cause pneumonia and should not be given to HIV-positive drug users for that reason. However, the power in this observational study was limited, and additional studies need to be conducted in order to evaluate pneumovax further, and to provide evidence-based recommendations on the use of pneumococcal vaccine in HIV-positive patients in the era of highly active antiretroviral therapy. Catharina E. A. Lindenburga Miranda W. Langendama Birgit H. B. v. Benthema Frank Miedemabc Roel A. Coutinhoac