Effect of PCV Dosing Schedules on Pneumococcal Disease
Our review identified a substantial body of research evaluating whether PCV use in young children leads to indirect effects in other age groups, although there are more data supporting some schedules than others. Most of the data were from studies evaluating 2 or 3 primary dose schedules with a booster dose (2+1, 3+1 or 3+PPV23), and among these, studies evaluating a 3+1 dosing schedule were most common. While studies have evaluated pneumonia, VT-NP carriage and VT-IPD, the demonstration of indirect effects was most consistent across studies and for all schedules for VT-IPD.
Because the first countries to introduce PCV used a 3+1 schedule, most of the available literature on indirect effects is for that schedule. The weight of evidence suggests that the use of a 3+1 schedule as part of a routine vaccination program for all infants will result in reduction of carriage and disease in age groups not targeted to receive PCV. Of 12 studies that we identified evaluating the 3+1 schedule for VT-IPD, only 2 showed no evidence of reductions in VT-IPD in unimmunized age groups; both took place in countries without catch-up campaigns and vaccine coverage in the population may have been insufficient to demonstrate indirect effects. VT-NP carriage studies have also shown indirect effects with 3+1 schedules. The impact on VT-NP carriage was observed in high-risk populations; no NP carriage studies with indirect effects were conducted in general populations. Studies of syndromic pneumonia only showed an impact with 3+1 schedules. Among 6 studies evaluating the 3+1 schedule, only 2, from Spain and the United States, observed increases in overall trends of pneumonia (pneumococcal, clinical and radiologically confirmed), with authors of both studies speculating that the overall increases were due to increases in nonvaccine serotype disease, although other secular trends could have contributed.
A smaller but growing number of studies have examined 3-dose (2+1 and 3+0) schedules. Most policy makers recently adopting PCV have used 1 of these schedules and the World Health Organization recently updated their recommendation for PCV to be used on either of these 3-dose schedules. Our review did not find sufficient data to directly compare these 2 schedules or to make conclusions regarding the impact of these schedules on VT-NP carriage or syndromic pneumonia; statistically significant indirect effects for pneumonia and VT-NP carriage using 2+1 and 3+0 schedules were not observed in any of the 6 studies of these outcomes identified by our review, although many of these studies were conducted early in the immunization programs or evaluated nonspecific endpoints. Despite these limitations, both 3-dose schedules appear to have indirect effects on VT-IPD when introduced nationally. Substantial reductions in VT-IPD were observed among young adult groups in 5 European countries using a 2+1 national immunization schedule. In countries with catch-up campaigns, this reduction was observed as early as 1 year after vaccine introduction. Two studies evaluated 3+0 schedules for indirect effects on VT-IPD and both found significant reductions. One 3+0 study in Australia did find 3–11% reductions in pneumonia; however, these findings were only borderline significant. Additionally, a 3+PPV23 schedule in Australia showed a decrease in VT-NP carriage among older children but not among adults. However, other studies suggest that PPV23 boosters do not affect VT-carriage, and thus a 3+PPV23 schedule likely approximates a 3+0 schedule in terms of benefits against VT-carriage.
This review also found studies of indirect effects of PCV on high-risk populations, including 9 studies evaluating PCV on either VT-NP carriage or VT-IPD; no studies evaluated pneumonia and all used 3+1 or 3+PPV23 schedules. Seven studies focused on the impact of PCV on indigenous populations, including Australian Indigenous, Alaskan Native and American Indian populations, and 2 focused on HIV-infected populations. Despite these varying populations, the findings were consistent. All studies noted reductions in disease in older children and adults, suggesting indirect impact of PCV on high-risk populations. These observations may be of relevance to countries with a high burden of HIV or vulnerable populations at higher risk of pneumococcal disease.
While the strength of this analysis is the diversity of settings and study designs included, both for high-risk and nonhigh-risk populations, there are some limitations to our analysis. The heterogeneity of the data did not allow for direct comparisons among schedules and since many factors contribute to the indirect impact of a vaccine schedule, this analysis was unable to fully address the wide variability in study settings and factors that may contribute to the relative impact of PCV schedules (eg, vaccine coverage, presence of a catch up campaign, proportion of the population under 5 years of age, HIV prevalence). Additionally, few data points exist for most of the outcomes we evaluated. Only 1 study directly compared impact among dosing schedules; this VT-NP study from the Netherlands showed no impact of either a 2+0 or a 2+1 schedule on NP carriage in older siblings and parents of vaccinated and unvaccinated children participating in an individual RCT. This study was the only study to evaluate a 2+0 schedule. Furthermore, some of the studies presented here were small and/or were substudies of clinical trials and therefore may not accurately represent the herd protection of vaccine introduction in a broad population. Many of the studies took place over relatively short periods; since full realization of herd effects in a population may take years, study periods of just a few years likely underestimate the measured herd effects in some studies. As PCV introductions in lower- and middle-income countries have only recently occurred, almost all data on impact from routine use came from high-income, early introducing countries with more mature immunization programs, which may be more likely to show indirect effects; however, a number of studies are ongoing and data will likely be available soon on the impact of routine use of PCV on unvaccinated older children and adults in lower- and middle-income settings.
The findings of our review suggest to policy makers that, should they adopt either a 3- or 4-dose PCV schedule, indirect effects are likely to add to the overall benefits seen from their program. The evidence to date is strong for the 3+1 schedule and is growing for the 3-dose schedules (2+1 and 3+0). More data to support evidence of herd effects from countries using either the 2+1 or 3+0 schedule would be useful, in particular for the outcomes of VT-NP carriage and pneumonia, from developing country settings where transmission may be more intense and across a wider age range than in high-income populations and for the new generation of conjugate vaccines (PCV10 and PCV13). Because studies of PCV effect on NP carriage in vaccinated children show that 3- and 4-dose schedules reduce colonization, we anticipate that with time and more study, vaccination of infants using all of these schedules will be found to prevent a variety of disease syndromes and colonization in unvaccinated age groups. For policy makers trying to determine the best schedule to adopt for their national PCV program, the evidence summarized here on indirect effects should provide an adjunct to data on the direct benefits of various PCV schedules for infants and to programmatic and epidemiologic factors specific to their situation that would drive their decisions on PCV use.
Discussion
Our review identified a substantial body of research evaluating whether PCV use in young children leads to indirect effects in other age groups, although there are more data supporting some schedules than others. Most of the data were from studies evaluating 2 or 3 primary dose schedules with a booster dose (2+1, 3+1 or 3+PPV23), and among these, studies evaluating a 3+1 dosing schedule were most common. While studies have evaluated pneumonia, VT-NP carriage and VT-IPD, the demonstration of indirect effects was most consistent across studies and for all schedules for VT-IPD.
Because the first countries to introduce PCV used a 3+1 schedule, most of the available literature on indirect effects is for that schedule. The weight of evidence suggests that the use of a 3+1 schedule as part of a routine vaccination program for all infants will result in reduction of carriage and disease in age groups not targeted to receive PCV. Of 12 studies that we identified evaluating the 3+1 schedule for VT-IPD, only 2 showed no evidence of reductions in VT-IPD in unimmunized age groups; both took place in countries without catch-up campaigns and vaccine coverage in the population may have been insufficient to demonstrate indirect effects. VT-NP carriage studies have also shown indirect effects with 3+1 schedules. The impact on VT-NP carriage was observed in high-risk populations; no NP carriage studies with indirect effects were conducted in general populations. Studies of syndromic pneumonia only showed an impact with 3+1 schedules. Among 6 studies evaluating the 3+1 schedule, only 2, from Spain and the United States, observed increases in overall trends of pneumonia (pneumococcal, clinical and radiologically confirmed), with authors of both studies speculating that the overall increases were due to increases in nonvaccine serotype disease, although other secular trends could have contributed.
A smaller but growing number of studies have examined 3-dose (2+1 and 3+0) schedules. Most policy makers recently adopting PCV have used 1 of these schedules and the World Health Organization recently updated their recommendation for PCV to be used on either of these 3-dose schedules. Our review did not find sufficient data to directly compare these 2 schedules or to make conclusions regarding the impact of these schedules on VT-NP carriage or syndromic pneumonia; statistically significant indirect effects for pneumonia and VT-NP carriage using 2+1 and 3+0 schedules were not observed in any of the 6 studies of these outcomes identified by our review, although many of these studies were conducted early in the immunization programs or evaluated nonspecific endpoints. Despite these limitations, both 3-dose schedules appear to have indirect effects on VT-IPD when introduced nationally. Substantial reductions in VT-IPD were observed among young adult groups in 5 European countries using a 2+1 national immunization schedule. In countries with catch-up campaigns, this reduction was observed as early as 1 year after vaccine introduction. Two studies evaluated 3+0 schedules for indirect effects on VT-IPD and both found significant reductions. One 3+0 study in Australia did find 3–11% reductions in pneumonia; however, these findings were only borderline significant. Additionally, a 3+PPV23 schedule in Australia showed a decrease in VT-NP carriage among older children but not among adults. However, other studies suggest that PPV23 boosters do not affect VT-carriage, and thus a 3+PPV23 schedule likely approximates a 3+0 schedule in terms of benefits against VT-carriage.
This review also found studies of indirect effects of PCV on high-risk populations, including 9 studies evaluating PCV on either VT-NP carriage or VT-IPD; no studies evaluated pneumonia and all used 3+1 or 3+PPV23 schedules. Seven studies focused on the impact of PCV on indigenous populations, including Australian Indigenous, Alaskan Native and American Indian populations, and 2 focused on HIV-infected populations. Despite these varying populations, the findings were consistent. All studies noted reductions in disease in older children and adults, suggesting indirect impact of PCV on high-risk populations. These observations may be of relevance to countries with a high burden of HIV or vulnerable populations at higher risk of pneumococcal disease.
While the strength of this analysis is the diversity of settings and study designs included, both for high-risk and nonhigh-risk populations, there are some limitations to our analysis. The heterogeneity of the data did not allow for direct comparisons among schedules and since many factors contribute to the indirect impact of a vaccine schedule, this analysis was unable to fully address the wide variability in study settings and factors that may contribute to the relative impact of PCV schedules (eg, vaccine coverage, presence of a catch up campaign, proportion of the population under 5 years of age, HIV prevalence). Additionally, few data points exist for most of the outcomes we evaluated. Only 1 study directly compared impact among dosing schedules; this VT-NP study from the Netherlands showed no impact of either a 2+0 or a 2+1 schedule on NP carriage in older siblings and parents of vaccinated and unvaccinated children participating in an individual RCT. This study was the only study to evaluate a 2+0 schedule. Furthermore, some of the studies presented here were small and/or were substudies of clinical trials and therefore may not accurately represent the herd protection of vaccine introduction in a broad population. Many of the studies took place over relatively short periods; since full realization of herd effects in a population may take years, study periods of just a few years likely underestimate the measured herd effects in some studies. As PCV introductions in lower- and middle-income countries have only recently occurred, almost all data on impact from routine use came from high-income, early introducing countries with more mature immunization programs, which may be more likely to show indirect effects; however, a number of studies are ongoing and data will likely be available soon on the impact of routine use of PCV on unvaccinated older children and adults in lower- and middle-income settings.
The findings of our review suggest to policy makers that, should they adopt either a 3- or 4-dose PCV schedule, indirect effects are likely to add to the overall benefits seen from their program. The evidence to date is strong for the 3+1 schedule and is growing for the 3-dose schedules (2+1 and 3+0). More data to support evidence of herd effects from countries using either the 2+1 or 3+0 schedule would be useful, in particular for the outcomes of VT-NP carriage and pneumonia, from developing country settings where transmission may be more intense and across a wider age range than in high-income populations and for the new generation of conjugate vaccines (PCV10 and PCV13). Because studies of PCV effect on NP carriage in vaccinated children show that 3- and 4-dose schedules reduce colonization, we anticipate that with time and more study, vaccination of infants using all of these schedules will be found to prevent a variety of disease syndromes and colonization in unvaccinated age groups. For policy makers trying to determine the best schedule to adopt for their national PCV program, the evidence summarized here on indirect effects should provide an adjunct to data on the direct benefits of various PCV schedules for infants and to programmatic and epidemiologic factors specific to their situation that would drive their decisions on PCV use.