Examining Duration of Protection: Should a Booster Dose Be Part of All Infant Pneumococcal Conjugate Vaccine Programs?

(See the Major Article by Jayasinghe et al on pages 367–74.) In this issue of Clinical Infectious Diseases, Jayasinghe et al [1] report good overall effectiveness of 7-valent and 13-valent pneumococcal conjugate vaccines (PCVs) to prevent invasive pneumococcal disease in Australia for their infant schedule of 3 primary doses (at ages 2, 4, and 6 months) without a booster. The article adds more evidence to the large and growing body of scientific work showing the success of PCVs in preventing disease. Previous work has demonstrated that PCVs not only prevent a range of syndromes from otitis media to pneumonia to meningitis in vaccinated children but that they also stop transmission of vaccine-type strains by preventing nasopharyngeal carriage, thereby preventing disease in unvaccinated children and adults [2]. Around the globe, nearly all would agree that the benefits of PCV programs have been as good or better than anticipated. With this background, the current report is notable because the investigators detected evidence of a limitation in what PCVs can do: when given on a schedule of 3 doses at 2, 4, and 6 months of age with no booster dose, the ability of PCV to prevent invasive disease among Australian children appeared to wane over time. In fact, the odds of contracting vaccine-type invasive pneumococcal disease were >5-fold higher among children who were vaccinated >24 months previously compared with those who had been vaccinated within 12 months. The findings convinced the Australian National Immunisation Technical Advisory Group to change their long-standing recommended schedule from 3 primary doses without a booster (3 + 0) to 2 primary doses with a booster (2 + 1). Before their recent schedule change, Australia had been the lone high-income country using pneumococcal conjugate vaccine in infants without a booster dose, although a 3-dose primary series without a booster is standard practice in low-income countries that use PCVs. Whether a PCV schedule without a booster dose provides durable protection has been a concern, given past experience with conjugate vaccines to prevent Haemophilus influenzae type b (Hib) disease. The United Kingdom, for example, had used a schedule without a booster dose for Hib vaccine for nearly a decade and changed their schedule to add a booster after detecting an increase in Hib cases in older children [3]. Until the current analysis, however, no study had shown limited duration of protection for PCVs in a general population of children. South African researchers evaluated duration of protection during 4 years of extended follow-up after the end of a clinical trial for a 9-valent vaccine given on a 3 + 0 schedule (a total of 6 years of follow-up since the trial began); vaccine efficacy remained unchanged in healthy children but dropped among children with human immunodeficiency virus (HIV) [4]. In the setting of the South African clinical trial, spread of pneumococci from carriers in the community who were not part of the trial may have been enough to provide natural boosting for healthy children who had received their primary series, and therefore no drop in efficacy was seen. Of note, the South African routine PCV program began with a schedule of 3 primary doses and later changed to 2 primary doses with a booster because of concerns about vaccine effectiveness and duration of protection among children with HIV [5]. In low-income countries, studies detect pneumococcal carriage in up to 90% of children aged <5 years, suggesting a high amount of pneumococcal transmission in the community [6]. In high-income settings, typically less than half of children in this age range will have pneumococci detected [7]. Natural boosting might be occurring in high-transmission settings in some countries currently using PCV in routine programs, in particular those with lower immunization coverage. In Australia, a low-transmission setting that has had high coverage with 3 doses of PCV for over a decade, very few vaccine-type pneumococci would be around to provide natural boosting, which might explain why the authors detected waning protection from PCV13 with increasing time since the primary series. Whether a substantial proportion of the children developing invasive pneumococcal disease >24 months after vaccination in the Australian study had underlying conditions that may have limited their initial immune response to vaccination is unknown because data on underlying medical conditions were not available. The authors used multiple approaches to demonstrate the overall effectiveness of PCVs in Australia, as well as the declining effectiveness with increasing time since vaccination. First, a simple analysis of their disease reports showed that the number of vaccine-serotype cases among vaccine-eligible children increased during the first few years after PCV13 introduction. Next, they used 2 different case–control approaches to measure vaccine effectiveness—a conventional method that used controls matched from children in the birth registry and the indirect cohort method, which uses children with nonvaccine serotype pneumococcal disease as controls. The conventional case–control study and the indirect cohort analysis produced similar vaccine effectiveness estimates, as was seen in a comparison of the 2 methods using US data [8]. An analysis of the potential for the indirect cohort analysis to overestimate vaccine effectiveness in the United Kingdom suggested that the method might have overestimated effectiveness by 2%–5% in their setting, an assessment that used a model to account for the increase in carriage and disease caused by nonvaccine serotypes seen after PCV introduction [9]. Although observational studies can be susceptible to bias [10], the ability of the authors to support their conclusions using different approaches strengthens their findings. These analyses highlight how observational study designs can be used to answer questions not measurable using randomized controlled clinical trials, such as those requiring large populations and multiple years of observation. The results from Australia may lead one to ask whether booster doses should be used for pneumococcal vaccines in every country. With the change in Australia’s schedule to include a booster dose, the question is now relevant only for low- and middle-income countries. On the plus side, booster doses result in higher antibody levels [11] that may lead to longer duration of protection and help to prevent carriage of vaccine-serotype pneumococci among 1-year-olds, who may be the pneumococcal “super-spreaders” in many communities. If toddlers are indeed the leading source of pneumococcal transmission, controlling carriage in this group could be the key to generating strong herd effects and community protection. Conversely, shifting to a schedule with a booster dose could jeopardize direct protection early in life, particularly in settings with fragile immunization infrastructure. Using a booster dose usually means dropping 1 of the 3 doses from the primary series; keeping 3 primary doses and adding a fourth (booster) dose is often not a consideration because of program costs or an effort to limit the total number of shots. Dropping a dose from the primary series may mean less direct protection during the months leading up to the booster dose, and most deaths from pneumococcal disease in children occur in those aged <1 year. Better infrastructure is needed in many places for delivering vaccines to children at around a year of age or later. The latest data from the United Nations Children’s Fund and the World Health Organization (WHO) show that achieving vaccination coverage goals for measles vaccines at 9 months and after the first birthday is a work-in-progress in some countries that likely have high pneumococcal disease burdens [12]. The WHO’s Strategic Advisory Group of Experts is now considering whether to give a preference to the 3 + 0 or the 2 + 1 schedule as they update their guidance for use of PCVs. In the WHO position paper from 2012, the 3 + 0 schedule was recommended, with the 2 + 1 schedule an acceptable alternative [13]. Previous reviews of the literature around the world have found that both schedules are highly effective and have not clearly pointed to one or the other schedule as producing better results [14]. The findings from Australia in this issue of CID provide a new and important piece of evidence to consider in this important schedule decision; however, more data are needed from other settings where disease transmission, disease risk, and ability to deliver vaccine doses might be quite different than that seen in Australia. Disclaimer. The findings and conclusions in this report are those of the author and do not necessarily represent the official position of the US Centers for Disease Control and Prevention. Potential conflicts of interest. The author: No reported conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.