The CAPITA study of protein-conjugate pneumococcal vaccine and its implications for use in adults in developed countries

Abstract

Until 1990, Hemophilus influenzae type b (HITB) was a major cause of morbidity and mortality in toddlers and young children. A vaccine consisting of purified polyribosyl ribitol phosphate (PRP), the capsular polysaccharide (CPS) of HITB, had been shown to be ineffective as an antigen in the population at risk, and this vaccine was withdrawn from the market within a few years of its introduction. By contrast, the discovery that PRP, when covalently bound to an antigenic protein, stimulated antibody production in infants and toddlers,¹ led to the development of a vaccine that has all but eradicated HITB infection and brought about a near-disappearance of this organism in the United States.

Keywords: :

• Pneumococcus
• vaccine
• conjugate vaccine
• pneumonia
• CAPITA

These same principles were rapidly applied to capsular polysaccharides of Streptococcus pneumoniae. A randomized, controlled study of 38 000 infants in the Kaiser-Permanente network showed that a protein-conjugate vaccine consisting of CPS from the 7 pneumococcal strains that most commonly infected infants and young children (PCV7) was > 95% effective in preventing invasive disease due to vaccine serotypes.²

Three important additional observations were made: (1) Because PCV7 also stimulated mucosal immunity, it greatly reduced the frequency of colonization by vaccine strains; (2) The result of decreased colonization was a remarkable protection of nonvaccinated members of society of all ages, such that, over the ensuing decade, pneumococcal disease due to these 7 common serotypes decreased by >95% in unvaccinated children and adults (so-called “herd immunity”);³ and (3) new strains emerged to partially fill the ecological niche that had been vacated by vaccine strains.⁴

This brief recapitulation clearly shows potential strengths and weaknesses of the pneumococcal conjugate vaccine (PCV), now expanded to include 13 capsular types (PCV13). Once this vaccine was shown to be so brilliantly effective in children, no one would have doubted its likely efficacy in adults. We correctly predicted⁵ a successful result from the recently concluded Community-acquired Pneumonia Immunization Trial in Adults (CAPITA), in which adults who lived in the Netherlands were randomized to receive PCV13 or a placebo. Enrollment was begun in the fall of 2008 and completed on January 31, 2010; observation was completed on August 28, 2013. The Netherlands was chosen because its well-functioning nationalized health system designates a physician for most of the inhabitants and because there had been relatively little use of pneumococcal vaccine in adults up to that time. A total of 84 496 subjects were randomized to receive PCV13 or placebo. The mean time for observation was set to exceed 3 y, and subjects were to be followed until the first bout of: (1) bacteremic pneumococcal pneumonia (BPP); (2) non-bacteremic pneumococcal pneumonia (non-BPP), largely diagnosed by a multiplex assay that used monoclonal antibodies to detect CPS from vaccine types (VT) in urine; or (3) non-pneumonia invasive pneumococcal infection (IPD). Pneumococcal infections were to be further distinguished into those due to VT and all others.

Before examining the results, it is essential to comment on the context in which CAPITA was undertaken. PCV7 was first recommended for Dutch infants and children in June, 2006 and this recommendation was changed to PCV10 in March, 2011. During the study period, only 18 cases of BPP were caused by PCV7 strains, verifying the efficacy with which PCV7 eliminated vaccine strains from the population. The four most commonly infecting serotypes were 1, 3, 7F and 19A, accounting for 67 of 90 total infections. None of these is contained in PCV7 and two are also not contained in PCV10; PCV 13 includes them all, suggesting that complete success in vaccinating the infant population will lead to the near-disappearance of these strains, as well.

The vaccine successfully prevented BPP and non-BPP, reducing the occurrence of disease in each category by about 45%. All other IPD was reduced by 75%. Regarding protection against non-BBP, the use of a sensitive and specific diagnostic test enabled a clear demonstration of vaccine efficacy. This phenomenon has consistently been observed during testing of vaccines—e.g., the effect of PCV7 on otitis media: the more sensitive and specific the test, the greater is the possibility of demonstrating vaccine efficacy.

A well-designed and well-executed study, the CAPITA trial was designed to answer the question of whether PCV13 protects adults against pneumococcal disease. Unfortunately, it leaves several important questions unanswered, many of which would have given the health care practitioner important guidance as to its use.

First, because CAPITA only compared PCV13 to placebo, it did not address the crucial question of whether PPV23, the existing pneumococcal polysaccharide vaccine is equally protective. (A study may be well-designed but may fail to examine important scientific questions in the groups selected for study.) The most recent meta-analysis of PPV23⁶ supported an earlier one⁷ in showing that good efficacy for PPV23—in fact, quite similar to that shown by PCV13 in CAPITA. In the meta-analysis, PPV23 reduced BPP, non-BPP and IPD by 74% (OR 0.26, 95% CI 0.15–0.46), 54% (OR 0.46, CI 0.25–0.84), and 80% (OR 0.2, CI 0.1–0.39), respectively. A meta-analysis by Huss et al.⁸ showed a 36% reduction for non-BPP and no significant efficacy for the other outcomes. These authors, however, disregarded many studies that other authorities consider valid, and their conclusions on the lack of efficacy of PPV23 for BPP rely heavily on only two papers, one of which used non-validated diagnostic methods to prove pneumococcal infection. The recently reported Spanish study (CAPAMIS)⁹ also showed 48% protection against non-BPP when it confined analysis to persons vaccinated within the preceding 5 y, albeit not against BPP. Absent a direct comparison of PPV23 and PCV13, even without the complication introduced by herd effect, we cannot conclude that one of these vaccines is better than the other.

Second, what will be the anticipated efficacy of PCV13 once infants are universally vaccinated? Because PCV13 eliminates colonization, its routine use in infants and toddlers will almost certainly be followed by a disappearance of these VT from the population, just as already has occurred in the United States and in the Netherlands with serotypes contained in PCV7. Thus, the routine use of PCV13 in adults will become increasingly trivial, whereas protection against other, currently less-common serotypes that are likely to be included in PPV23, may become increasingly important. If PPV23 is an effective vaccine, then recommending widespread use of PCV13 in non-immunocompromised adults does not appear to be sound public policy.

Third, should PCV13 be used to “prime” immunocompromised subjects who can then receive PPV23 as a boosting dose? This role for conjugate followed by polysaccharide vaccine is implied in the current ACIP recommendations.¹⁰ As we have reviewed earlier,⁵ there are insufficient data to support this concept, and the absence of a benefit was specifically demonstrated in transplant recipients.¹¹

Fourth, is PCV13 more likely than PPV23 to provide better protection for immune-compromised subjects? Our earlier reviews^5,12 of antibody responses to PPV23 vs PCV showed no distinct advantage for PCV. Evidence for better efficacy of the conjugate vaccine in patients with AIDS was found in HIV-infected adults in two field trials conducted by the same group of investigators. In a study in Uganda,¹³ French et al. showed that PPV23 failed to protect HIV-infected adults, most of whom had advanced infection and were not on treatment, whereas, in a similar study in Malawi, done at a time in which a national program for antiretroviral therapy was being initiated, showed that two doses of PCV7¹⁴ were protective, at least for the first year of study. Importantly, in the relatively small number of immunodeficient or immunosuppressed subjects in CAPITA, PCV13 exhibited no protective effect (22 total cases of pneumococcal disease in vaccine-recipients vs. 24 cases in placebo-recipients). It is unclear whether the results of the two studies by French et al.^13,14 in HIV infected patients support ACIP’s broad recommendation to administer PCV13 to all immunocompromised subjects over the age of 19. For example, earlier studies comparing antibody responses in organ transplant patients showed no meaningful differences between PPV23 and PCV7,^15,16 and the meager literature that addresses the question in patients with hematologic malignancies does not consistently document better responses to PCV.^5,17

Finally, should PCV13 be used routinely in healthy adults ≥ 50 y of age to protect them against pneumococcal infection? The rate of pneumococcal infection as a cause of pneumonia is substantially lower than it was a few decades ago.¹⁸ Recent studies in the US suggest that no more than about 10–12% of CAP is caused by pneumococcus,^19-21 Taking this observation together with the steady decline in disease caused by serotypes contained in PCV13, it might well be regarded as entirely superfluous to initiate a campaign to vaccinate adults 50–64 y old with PCV13.

In summary, the CAPITA trial demonstrates that PCV13 effectively protects adults against infection by pneumococcal serotypes that are represented in the vaccine. This trial was performed in a country where PCV10 utilization in infants and toddlers was in its earliest stages and where PCV13 had not yet been used at all. The trial did not support the use of PCV13 in immunocompromised adults and addressed neither the question of immunological “priming” nor the comparative effectiveness of this vaccine vs. PPV23. Taken together with the importance of the herd effect and the generally low rate with which S. pneumoniae is implicated in CAP, these facts suggest that widespread use of PCV13 in adults ages 50–64, or even in all immunocompromised adults ≥ 19 y of age will have limited protective effect as a public health measure.