aMRC Programme on AIDS, Entebbe, Uganda; bDepartment of International Health, Imperial College of Science and Technology, London; and cLiverpool School of Tropical Medicine, Liverpool, UK.
Sponsorship: This work was funded by the United Kingdom Medical Research Council (G9323636) and supported by TASO Uganda. Dr French is supported by a career development fellowship from the Wellcome Trust (061230/Z/00/Z).
Received: 9 July 2003; revised: 9 January 2004; accepted: 28 January 2003.
23-Valent pneumococcal polysaccharide vaccine was previously reported to be ineffective in HIV-infected Ugandan adults. Prolonged follow-up of trial participants confirmed a persistent excess of all-cause pneumonia in vaccine recipients [hazard ratio (HR) 1.6; 95% confidence interval (CI) 1.0–2.4], but surprisingly a survival advantage favouring vaccination (HR 0.84; CI 0.7–1.0). An explanation for the improvement in survival in the face of excess morbid events is lacking; a role for vaccine in HIV care in Africa remains unlikely.
We have previously reported an increase in all-cause pneumonia and an excess of invasive pneumococcal events in HIV-infected Ugandan adults who received vaccine when participating in a randomized, controlled trial of 23-valent pneumococcal polysaccharide vaccine (PPV) . Explanations for these findings were suggested and immunological mechanisms were postulated. Subsequent correspondence raised concerns about the use of PPV in other immunodeficient groups , or in maternal immunization programmes in Africa. However, a chance outcome in a clinical trial must be considered as an alternative explanation for the results.
After the planned study period ended in June 1998, we continued to monitor the trial cohort using clinical, data collection and statistical procedures defined in the original trial manual and described previously . Both the participants and the trial staff remained blind to the vaccine status. Co-trimoxazole prophylaxis (960 mg a day) has been routinely offered to members of the cohort since August 2000. Isoniazid prophylaxis was administered to a number of skin-test-positive cohort members between November 1998 and April 2000 as part of an operational assessment of this approach . This extended period of follow-up has provided an opportunity to assess the robustness of the earlier results and reconsider the possibility that these were chance findings.
At the end of December 2001, 260 vaccine and 213 placebo recipients (36% of the original cohort) were alive and under follow-up, 711 participants (54%) had died and 139 (10%) were recorded as defaulters (70 vaccine, 69 placebo). Overall, the excess rate of first-event pneumococcal-related disease persisted in vaccine recipients. However, none of this excess was attributable to the later period, when little difference between rates was measured (see Table 1). The risk of pneumococcal disease in vaccine recipients strengthened with increasing CD4 cell counts, particularly all-cause pneumonia (see Table 1). During the later period, survival was greater in vaccine recipients, such that overall survival favoured vaccination but particularly in the group with CD4 cell counts of 200–500 cells/μl at enrolment. Co-trimoxazole prophylaxis was used by 302 study participants (169 in the placebo arm and 133 in the vaccine arm). Adjustment for its use did not significantly alter the hazard estimates (data not shown). No protective effect of co-trimoxazole on all-cause pneumonia or invasive pneumococcal events was measured, with incidence rate ratios of 1.11 (0.3–3.7) and 0.66 (0.2–1.9), respectively (Poisson regression adjusted for CD4 cell count). The numbers of individuals taking isoniazid were similar in the placebo and vaccine arms, 36 and 33, respectively. Multiple adjustment of the proportional hazard model for age, sex, enrolment CD4 cell count, clinical stage and pneumococcal carriage at the time of enrolment did not significantly alter the hazard ratio estimates for disease or death (data not shown).
These data confirm the robustness of the original results and the poor efficacy of PPV in this population for preventing pneumococcal disease. Although the tendency towards increased disease episodes in the vaccine recipients has persisted, little difference in invasive pneumococcal events or all-cause pneumonia rates was noted by vaccine status in the later period. The overall rate of all-cause pneumonia fell during the extended follow-up period. This, we believe, primarily results from reduced access to radiology services and a failure to fulfil the chest radiographic criteria for all-cause pneumonia. The consistent rate of pneumococcal bacteremia in placebo recipients suggests that there has been no significant change in the natural history of pneumococcal disease over time.
The tendency towards an increasing hazard of pneumonia in vaccine recipients with higher CD4 cell counts at vaccination is directly contradictory to findings from retrospective studies in north America . Current recommendations in the Unites States  support vaccination as early as possible after HIV infection. These findings lend no support for this approach in the Ugandan population. The findings are consistent with the originally postulated explanation for an adverse effect of vaccination – pneumococcal polysaccharide vaccination induces a selective defect in the humoral response to re-challenge with these antigens. When the risk of disease is low, at higher CD4 cell counts, a selective defect in antipneumococcal immunity is likely to produce a greater difference in disease rates than when CD4 cell counts are low, and background rates are high in association with broad and established HIV-induced defects in immune response.
The measured survival advantage in vaccine recipients is surprising, given the clear lack of efficacy of PPV in preventing pneumococcal disease. A theoretical potential for pneumococcal polysaccharides to stimulate crossreactive protection against other serious polysaccharide-coated pathogens must be considered. However, rates of non-typhi Salmonella bacteremia and cryptococcal disease, both frequent in this population, have been unaffected by vaccination (see Table 1). The rates of clinic visits for all febrile illness were unaffected by vaccination status, suggesting that vaccine recipients were not accessing a greater amount of healthcare. Similarly, the rates (Table 1) and time to diagnosis of tuberculosis were similar in both groups; the median time (interquartile range) in months to diagnosis in vaccine recipients was 4.4 (1.8–11.1) and for placebo recipients was 5.1 (2.2–13.8). The possibility that increased respiratory disease in vaccine recipients may thus have enhanced surveillance and investigation for tuberculosis also seems unlikely. A direct effect on HIV pathogenesis also seems implausible; the upregulation of HIV transcription has been reported by one group . The remaining possibility is that this result represents a further chance finding.
Mortality rates in this cohort without antiretroviral therapy are catastrophically high. A 16% improvement in survival may be modest in comparison to the expected effect of antiretroviral therapy; however, this compares favourably with other prophylactic strategies recommended in sub-Saharan Africa [7,8], particularly as a single vaccination. We had originally dismissed the idea of a further clinical evaluation of PPV in HIV-infected Africans, but this may need to be re-considered.
Cotrimoxazole prophylaxis has had little effect on pneumococcal disease in this cohort. High rates of resistance among pneumococcal isolates to co-trimoxazole were recorded in the period before the commencement of chemoprophylaxis, 70% with a minimal inhibitory concentration of 1/19 μg/ml or greater trimethoprim–sulphamethoxazole, n = 167 (N. French, unpublished data). The lack of effect is therefore understandable. The pneumococcus continues to be a frequent and serious infection in HIV-infected African adults, and the evaluation of new and alternative control strategies remains a priority.
The authors would like to thank the se staff and members of the TASO Entebbe office and the Ugands Virus Research Institute for their help and participation.
1. French N, Nakiyingi J, Carpenter LM, Lugada E, Watera C, Moi K, et al
. 23-Valent pneumococcal polysaccharide vaccine in HIV-1-infected Ugandan adults: double-blind, randomised and placebo controlled trial. Lancet
2. Spoulou V, Theodoridou M, Papaevangelou VG, Mostrou GI, Ioannidis JP. 23-Valent pneumococcal vaccination and HIV. Lancet
3. Lugada ES, Watera C, Nakiyingi J, Elliott A, Brink A, Nanyunja M, et al
. Operational assessment of isoniazid prophylaxis in a community AIDS service organisation in Uganda. Int J Tubercl Lung Dis
4. Breiman RF, Keller DW, Phelan MA, Sniadack DH, Stephens DS, Rimland D, et al
. Evaluation of effectiveness of the 23-valent pneumococcal capsular polysaccharide vaccine for HIV-infected patients. Arch Intern Med
5. Kaplan JE, Masur H, Holmes KK. Guidelines for preventing opportunistic infections among HIV-infected persons – 2002. Recommendations of the U.S. Public Health Service and the Infectious Diseases Society of America. MMWR Recom Rep
6. Brichacek B, Swindells S, Janoff EN, Pirruccello S, Stevenson M. Increased plasma human immunodeficiency virus type 1 burden following antigenic challenge with pneumococcal vaccine. J Infect Dis
7. Quigley MA, Mwinga A, Hosp M, Lisse I, Fuchs D, Porter JDH, et al
. Long-term effect of preventive therapy for tuberculosis in a cohort of HIV-infected Zambian adults. AIDS
8. Grimwade K, Gilks C. Cotrimoxazole prophylaxis in adults infected with HIV in low-income countries. Curr Opin Infect Dis