HIV-1 infection is probably the single most important health problem facing adults in sub-Saharan African nations . Effective interventions to reduce morbidity and mortality are urgently needed as the HIV-1 epidemic evolves into the AIDS epidemic. Because Streptococcus pneumoniae is a leading cause of febrile illness in HIV-1-infected adults in Africa , with an incidence of invasive disease of 4% per year in Nairobi  (compared with 1% per year in the United States) [4,5], pneumococcal disease is a prime target for intervention. A strategy of prevention based on a single-dose vaccine is particularly well-suited to developing nations where resources are limited and the disease is so common.
Despite its safety, low cost, and potential clinical value, the efficacy of pneumococcal polysaccharide vaccine in HIV-1-infected patients is unknown. Retrospective analyses have suggested a protective effect of vaccine in other patients at high risk [6,7], results that may extend to those with HIV-1 infection . Although routinely recommended for use in HIV-1-infected patients in the United States and the United Kingdom [9,10], broad distribution of the vaccine cannot be justified as yet in Africa, where annual health-care budgets are of the order of US$2–10 per person. A clear demonstration of efficacy (and cost-effectiveness) is required before vaccination can be justified for public use in HIV-1-infected adults in these countries.
We performed a feasibility study for the use of pneumococcal polysaccharide vaccine in HIV-1-infected adults in Entebbe, Uganda as an open-label pilot trial. The aims of this study in a high-risk population were to assess the recruitment and compliance of HIV-1-infected adults in the Entebbe district to be vaccinated and to complete follow-up and to determine the immunogenicity of the vaccine to elicit capsule-specific antibodies, which serve as the primary determinant of defense against the organism .
We describe a successful vaccine recruitment program, vaccine responses in 77 HIV-1-infected and 10 HIV-seronegative adults, outcome and mortality in vaccine recipients, as well as two cases of pneumococcal disease after vaccination. The mechanisms identified for these failures (impaired humoral responses and disease with a non-vaccine serotype) may help to guide the development of effective vaccine strategies for this common and serious invasive infection.
Materials and methods
HIV-1-infected adults were recruited from the Entebbe office of a local community-based HIV-1 care center (the AIDS Support Organization of Uganda; TASO). This clinic has approximately 1000 registered active HIV-1-infected clients. HIV-1-seronegative adults were age-matched from TASO staff. Vaccination was undertaken in sequential volunteers following an announcement at a center meeting in late October 1994. Pregnant women or those suffering an acute febrile illness were excluded. Participants gave individual written informed consent following discussions in the local language (Lugandan) of protocols approved by the AIDS research committee of the Ugandan Ministry of Health, in accord with UN/WHO standards. Recruits were considered under follow-up if they had attended the study clinic in the 3 months preceding the beginning of August 1997. Clinical characteristics of study participants are summarized in Table 1.
Vaccine and sampling
Pneumococcal vaccine (Pnu-immune, generously supplied by Wyeth-Lederle-Cyanamid, Pearl River, New York, USA), which contained 25 μg of each of 23 capsular polysaccharides, was given as an intradeltoid injection. Blood sampling was performed immediately prior to and 1 month after vaccination. Serum samples were separated within 6 h and stored at −70°C until testing.
Capsule-specific immunoglobulin G measurement
Levels of pneumococcal capsule-specific immunoglobulin G (IgG) were measured by enzyme-linked immunosorbant assay (ELISA) as previously described . The capture antigens were capsular polysaccharides (American Type Culture Collection, Rockville, Maryland, USA; 5 mg/l) for serotypes common in East Africa (1, 6B, 14, and 19F). Sera were pre-adsorbed with cell wall polysaccharide (50 mg/l) to remove noncapsular immunoglobulin and six dilutions were tested in triplicate. Horseradish peroxidase-labeled affinity-purified goat anti-human IgG served as the detector antibody. Standard (CDC/FDA 89SF) and control sera were included on each plate. The coefficient of variation was < 12% for all serotypes.
Serum killing activity
A fixed inoculum of log phase S. pneumoniae (type 1 from Patient 1, type 13 from Patient 2), following passage through mice , was incubated in microwells with serial 1 : 2 dilutions of heat-inactivated human serum (56°C for 30 min), 0.2 mg complement from a single screened lot (Baby Rabbit Complement, Cedarlane Laboratories Ltd, Accurate Scientific, Hornby, Ontario, Canada), and 4 × 105 phagocytic cells (HL-60, a myelomonocytic cell line, differentiated to a granulocyte line using 120 mM dimethylformamide), yielding a granulocyte : bacteria (effector : target) ratio of 400 : 1. After rotation at 37°C for 60 min, samples were plated and incubated on blood agar plates. Results are reported as endpoint titers giving 50% kill compared with the number of organisms after incubation with fetal calf serum, complement, and effector cells. Sera obtained before and 1 month after vaccination were tested from Patients 1 and 2, three HIV-seronegative control subjects, and six randomly selected HIV-1-infected patients (CD4+ T cell counts 26–1218 × 106 cells/l). Acute and convalescent sera were also available from Patient 1 at the time of subsequent presentation with pneumococcal bacteremia and 1 month later.
Differences in levels of capsule-specific IgG in serum were compared between clinical groups by analysis of variance (for three groups) and unpaired two-tailed t-test (two group comparison), and within groups (preand post-vaccination) by paired two-tailed t-test using the Stat-view 4.01 statistical package (Abacus Concepts, Berkeley, California, USA). Increases (as fold rises) were determined by dividing each subject's post-vaccination levels by their pre-vaccination levels. A rise of greater than twofold was considered significant.
Compliance with follow-up and natural history of HIV-1 disease
We followed 95% of vaccinees (83 of 87) for up to 2.8 years (Table 1). Deaths occurred in 29 of 77 (38%) HIV-1-infected vaccine recipients and none of the 10 HIV-seronegative control subjects. Survival was strongly associated with CD4+ T cell count at vaccination (Table 1). The relative risk of death with a CD4+ T cell count < 300 × 106 cells/l was 3.8 [95% confidence interval (CI), 1.9–7.9] compared with the rate in patients with higher counts. Of the surviving 44 HIV-1-infected participants, 40 were regularly attending the study site; four were visited at home and reported good health as their reason for non-attendance at the clinic. Therefore, prospective vaccine trials can be successfully performed in semi-urban areas in Africa. The primary limitation to effective surveillance is the high mortality rate, rather than lack of compliance, in this population.
Response to pneumococcal vaccine: reactogenicity
Side-effects of vaccination were infrequent and minor. Discomfort at the vaccine site lasting less than 24 h was reported by 12 HIV-1-infected vaccine recipients (16%) and two HIV-seronegative control subjects (20%). A third control subject reported a self-limited flu-like illness for 48 h post-vaccination. Three HIV-1-infected vaccine recipients died within 6 weeks of vaccination from vaccine unrelated causes and did not contribute post-vaccination serum samples.
Baseline levels of capsule-specific IgG were similar among HIV-1-infected patients and control subjects for all capsule types tested (Table 2). Pneumococcal vaccine elicited significant increases in levels of capsule-specific IgG measured 1 month after vaccination for three of the four serotypes tested (serotypes 1, the most common pathogenic type in East Africa, 6B, and 14) in both groups. Only HIV-seronegative control subjects showed significant rises in antibody to serotype 19F. Moreover, levels of capsule-specific IgG post-vaccination were lower in the HIV-1-infected patients compared with those in control subjects for all serotypes. Among HIV-1-infected patients, CD4+ T cell counts did not significantly affect either baseline or post-vaccination levels of capsule-specific antibody (Fig. 1).
Both HIV-1-infected patients and HIV-seronegative control subjects generated appreciable mean increases in each serotype-specific antibody (Table 2, Fig. 1). Type 14 was the most reliably immunogenic serotype in both clinical groups. Differences in the mean increases between groups were significant only for type 6B, but the proportion of HIV-1-infected vaccinees with increases greater than twofold was consistently lower compared with that in control subjects.
A 41-year-old Ugandan woman received pneumococcal vaccine in November 1994. At that time she complained of chronic cough and suffered recurrent documented fevers that had lasted more than a month (WHO clinical stage 3) . Her CD4+ T cell count was 216 × 106 cells/l. Thirteen months after vaccination she visited the clinic with a history of fever and worsening cough in the preceding 3 days associated with production of purulent sputum. Examination revealed newly present rales at the right lung base, a temperature of 39°C, pulse 120 beats/min and respiratory rate 32/min. Sputum and blood cultures grew S. pneumoniae. She made an uneventful recovery with a single parenteral 5 million unit dose of benzylpenicillin and subsequent oral amoxicillin 500 mg three times daily for 5 days. The organism was penicillin-sensitive (based on oxacillin disc susceptibility) and capsular serotype 1 (based on the Quellung reaction, graciously performed by Dr Robert Austrian), a vaccine serotype that causes up to half of cases of invasive pneumococcal disease in sub-Saharan Africa [3,15]. Sera were obtained acutely and after 4 weeks of convalescence.
A 34-year-old Ugandan woman received vaccine in November 1994. She gave a history consistent with episodes of sinusitis occurring at least twice in the previous 6 months (WHO clinical stage 2). Her CD4+ T cell count was 494 × 106 cells/l. Fourteen months after vaccination she presented with a 6-day history of fever, left pleuritic chest pain and new cough productive of purulent sputum. On examination, she had bronchial breathing at the left lung base, fever of 37.8°C, pulse 100 beats/min, and respiratory rate 18/min. A chest radiograph confirmed the presence of left lower lobe pneumonia. Sputum culture grew only S. pneumoniae. She received a 5-day course of oral amoxicillin 500 mg three times daily. When seen 3 weeks later, she had improved and returned to work. The organism was penicillin sensitive and capsular serotype 13, a non-vaccine serotype. Peri-event sera were not available.
Serum killing activity
The ability of sera from three randomly selected control subjects and six HIV-1-infected patients to support complement-mediated killing by phagocytes of serotype 1 S. pneumoniae increased appreciably 4 weeks after vaccination compared with pre-vaccination values (Fig. 2; top two panels); the titers varied within each group. In contrast, sera from the two HIV-1-infected vaccine failures, Patients 1 and 2, showed no such ability to mediate killing of this organism in either pre- or post-vaccination sera (Fig. 2; left bottom panel). In addition, even after natural bacteremic infection with S. pneumoniae serotype 1, sera from Patient 1 showed no detectable activity to kill the organism (Fig. 2; right bottom panel). As expected, no change was noted in serum killing activity against serotype 13 (a nonvaccine serotype) in either of the two HIV-1-infected patients or the three HIV-seronegative control subjects (not shown).
We have shown that patients at appreciable risk of invasive pneumococcal disease, HIV-1-infected adults in East Africa, can be efficiently recruited and reliably followed for a prospective vaccine trial. These patients generate a significant antibody response to three of four capsule types contained within the 23-valent polysaccharide vaccine, including serotype 1, which comprises up to half of serious pneumococcal infections in this area . These responses appear independent of CD4+ T cell counts, as described in other populations, but are lower than those produced by the HIV-seronegative control subjects [13,16].
Antibodies directed against the polysaccharide capsule are the principal mechanisms of defense against invasive pneumococcal infection . The importance of these antibodies in protection is supported by the observations that HIV-1-infected Patient 1 failed to generate functionally active capsule-specific antibodies to serotype 1 after immunization; she subsequently developed pneumococcal bacteremia with that serotype and failed to produce functionally active antibodies to the organism after infection. That other HIV-1-infected patients did respond to the vaccine, as assessed by ELISA and the killing assay, and that quantitative antibody responses to capsular polysaccharides in the vaccine correlate with functional activity [13,17] suggest that vaccination may be helpful in protecting against serious invasive pneumococcal disease (e.g., bacteremia, meningitis) at all stages of HIV-1 infection. Nevertheless, as with Patient 1, there may be a subgroup of HIV-1-infected individuals who are unable to generate a response to capsular polysaccharide antigens, whether presented in vaccine or in the organism itself during natural infection. Identifying the specific mechanisms of unresponsiveness and the proportion of patients who show such impaired responses is important for several reasons.
First, the number of individuals with such inadequate responses in a cohort of patients and their contribution to the excess rates of pneumococcal disease seen in HIV-1-infected adults will be a powerful predictor of whether pneumococcal polysaccharide vaccination will be effective as a preventive public-health strategy. If, as with Patient 1, there is a small subgroup of HIV-1-infected adults at extremely high risk who account for most cases of disease and they do not respond to vaccination, it is unlikely that we will see an impact of vaccination at the population level. It is more likely, however, that the cases we have reported represent the extreme end of the distribution of impaired humoral immune response to polysaccharides. Among HIV-1-infected women in Nairobi, the frequency and magnitude of responses to vaccine among those with prior pneumococcal bacteremia was similar to those in women without previous pneumococcal disease . These data suggest that identifying a clear subgroup of non-responders to polysaccharide antigens at increased risk of pneumococcal disease may be difficult as the risks may be multifactorial . A prospective trial is now in progress in Uganda to establish whether vaccination will demonstrate a protective effect at the population level and whether assessment of vaccine-specific immunological responses correlate directly with clinical outcome.
Second, any protective effect of the vaccine will only relate to the serotypes of S. pneumoniae contained within the vaccine. Serotype 13, identified in Patient 2, is a non-vaccine type but is thought to be uncommon in East Africa. The distribution of serotypes included in the current vaccine were determined predominantly from data from developed countries but appear to represent over 80% of significant isolates found in HIV-1-infected adults in Nairobi, Kenya, and South Africa [3,19]. The public-health impact of vaccination will be limited if a high proportion of disease is caused by nonvaccine serotypes irrespective of any serotype-specific efficacy. This is a particular problem when fewer serotypes can be included in newer protein–polysaccharide vaccines. Consideration of local results on the prevalent serotypes causing disease will be important in planning vaccination programs.
Although pneumococcal vaccine efficacy may be limited by the immunogenicity of antigens in the population at risk and by the appropriateness of the vaccine serotypes selected, rigorous evaluation of vaccine efficacy in East Africa does not appear to be limited by patient recruitment or compliance. In Entebbe, access to a population of HIV-1-infected adults is good. The TASO clinic in Entebbe serves a large number of HIV-1-infected individuals as the only center primarily caring for their needs in the district. Interest in pneumococcal vaccination was strong and recruitment rapid (October/November 1994) because (i) vaccination (irrespective of type) is perceived positively, a consequence of the work of UNEPI (United Nations Extended Program of Immunization) and other organizations promoting childhood vaccination; and (ii) pneumococcal vaccination was considered by the clients as a relevant intervention for HIV-1-infected adults, when previously no other prophylactic strategies existed or were affordable. The willingness and enthusiasm of individuals to take part were also reflected in excellent follow-up and low default rate, both of which contributed to a successful trial. There are probably several reasons for these achievements. The use of an acceptable product may fill a perceived need in a population. Access to an established centralized clinic, TASO, is appealing because of its holistic approach to care and its provision of counseling, social, and subsidized medical services, which are unavailable elsewhere in the district. The geographically contained region around Entebbe and the surrounding settlements, which are situated on a peninsula into Lake Victoria, provided a circumscribed population. Finally, migration was low because, for many residents, Entebbe is their family home and where they will be buried. Consequently, they will remain in the area for sickness and terminal care. Indeed, the principal losses to follow-up were as a result of low rates of survival, which were in turn a consequence of the large proportion of individuals with relatively advanced diseases as measured by low CD4+ T cell counts. Recruitment for the vaccine efficacy trial will need to adjust sample sizes to achieve adequate observation time to account for the poor survival of HIV-1-infected patients attending this center. Nevertheless, high rates of pneumococcal disease should help to establish whether the vaccine has a substantial impact on rates of pneumococcal disease.
Our estimated rates of S. pneumoniae infections, 13 per 1000 person-years of observation (95% CI, 2–49) in vaccinated adults, may under-represent the true rates of disease in this population. Events may have been missed in the first 9 months when diagnostic microbiology facilities were under development. As a result, we cannot determine vaccine efficacy. However, disease rates will probably be high enough to support a randomized placebo-controlled trial. If rates of pneumococcal disease are closer to those found in neighboring Nairobi, Kenya (25–42 per 1000 person-years) , a 66% efficacy of the vaccine would be identifiable with sample sizes of approximately 500 subjects in each arm.
In summary, prophylaxis for opportunistic infections has had a dramatic impact on HIV-1 care and survival in the developed world. However, many of these regimens would be inappropriate in the developing world, where Pneumocystis carinii, cytomegalovirus, and Mycobacterium avium complex disease are infrequently reported . In venues such as sub-Saharan Africa, cost-effective prevention of pneumococcal disease, along with tuberculosis, non-typhi salmonellosis, and cryptococcosis, is a priority. The currently available polysaccharide vaccine offers the most realistic hope to date for inexpensive, accessible, and effective prevention of a common invasive secondary infection in patients with HIV-1 disease in Africa. This study establishes the feasibility of successfully performing an efficacy trial for prevention of invasive S. pneumoniae infections with pneumococcal vaccine in this high-risk population as well as for identifying potential mechanisms for vaccine failure (e.g., poor immunogenicity or inappropriate selection of serotypes).
We thank Dr Robert Austrian (University of Pennsylvania, USA) for performing serotyping of pneumococcal isolates, Ann Emery for manuscript preparation, and especially the dedicated staff and involved clients at TASO Entebbe for their support and care of the study and study team.
1. Mulder DW, Nunn AJ, Kamali A, Nakiyingi J, Wagner H-U, Kengeya-Kayondo JF: Two-year HIV-1-associated mortality in a Ugandan rural population. Lancet 1994, 343:1021–1023.
2. Gilks CF, Brindle RJ, Otieno LS, et al.: Life-threatening bacteraemia in HIV-1 seropositive adults admitted to hospital in Nairobi, Kenya. Lancet 1990, 336:545–549.
3. Gilks CF, Ojoo SA, Ojoo JC, et al.: Invasive pneumococcal disease in a cohort of predominantly HIV-1 infected female sex-workers in Nairobi, Kenya. Lancet 1996, 347:718–723.
4. Redd SC, Rutherford GW, Sande MA, et al.: The role of human immunodeficiency virus in pneumococcal bacteremia in San Francisco residents. J Infect Dis 1990, 162:1012–1017.
5. Schuchat A, Broome CV, Hightower A, Costa SJ, Parkin W: Use of surveillance for invasive pneumococcal disease to estimate the size of the immunosuppressed HIV-infected population. JAMA 1991, 265:3275–3279.
6. Shapiro ED, Berg AT, Austrian R, et al.: The protective efficacy of polyvalent pneumococcal polysaccharide vaccine. N Engl J Med 1991, 325:1453–1460.
7. Fine MJ, Smith MA, Carson CA, et al.: Efficacy of pneumococcal vaccine in adults. A meta analysis of randomized controlled trials. Arch Intern Med 1994, 154:2666–2677.
8. Ward JW, Hanson DL, Jones J, Kaplan JOM: Pneumococcal vaccination and the incidence of pneumoniae among HIV-1-infected persons. 34th Annual Meeting of the Infectious Disease Society of America, San Francisco, California, 1997 [abstract].
9. Centers for Disease Control: Recommendations of the immunization practices advisory committee on pneumococcal polysaccharide vaccine. MMWR 1989, 38:64–76.
10. HMSO: Immunisation Against Infectious Diseases. London: HMSO; 1992.
11. Bruyn GAW, Zegers BJM, van Furth R: Mechanisms of host defense against infection with Streptococcus pneumoniae. Clin Inf Dis 1992, 14:251–262.
12. Carson PJ, Schut RL, Simpson ML, O'Brien J, Janoff EN: Antibody class and subclass responses to pneumococcal polysaccharides following immunization of HIV-infected patients. J Infect Dis 1995, 172:340–345.
13. Janoff EN, Fasching C, Ojoo JC, O'Brien J, Gilks CF: Responsiveness of human immunodeficiency virus type 1-infected Kenyan women with and without prior pneumococcal disease to pneumococcal vaccine. J Infect Dis 1997, 175:975–978.
14. WHO International Collaborating Group for the Study of the WHO Staging System: Proposed World Health Organization staging system for HIV infection and disease: preliminary testing by an international collaborative cross-sectional study. AIDS 1993, 7:711–718.
15. Austrian R, Gold J: Pneumococcal bacteremia with especial reference to bacteremic pneumococcal pneumonia. Ann Intern Med 1964, 60:759–776.
16. Vandenbruaene M, Colebunders R, Mascart-Lemone F, et al.: Equal IgG antibody response to pneumococcal vaccination in all stages of human immunodeficiency virus disease. J Infect Dis 1995, 172:551–553.
17. Rubins JB, Puri AKG, Loch J, et al.: Magnitude, duration, quality and function of pneumococcal vaccine responses in elderly adults. J Infect Dis 1998, (in press).
18. Janoff EN, Rubins JB: Invasive pneumococcal disease in the immunocompromised host. Microbial Drug Resist 1997, 3:215–232.
19. Crewe-Brown HH, Karstaedt AS, Saunders GL, et al.: Streptococcus pneumoniae blood culture isolates from patients with and without human immunodeficiency virus infection: alterations in penicillin susceptibilities and in serogroups or serotypes. Clin Inf Dis 1997, 25:1165–1172.
© 1998 Lippincott Williams & Wilkins, Inc.