Editorial Commentary:Changing Epidemiology of Invasive Pneumococcal Disease: A Complicated Story

Streptococcus pneumoniae is a leading cause of pneumonia and meningitis worldwide, and >1 million children die because of pneumococcal infections globally each year [1]. Such infections are due to various pneumococcal serotypes, and the serotypes responsible for these infections vary with time and geography. Although children are generally at the highest risk for both pneumonia and invasive disease, in some populations in the world, this risk is much higher. Such populations include Native American Indians, Alaskan indigenous populations, and the Australian Aboriginals. In the White Mountain Apache (WMA) population, the risk of invasive pneumococcal disease was 8-fold higher prior to vaccine introduction, than it was in the general US population. Although the reasons for this elevated risk are still not completely understood, this elevated risk has resulted in extensive study of the WMA population. In this issue of Clinical Infectious Diseases, Lacapa et al. [2] report on the impact of the introduction of 7-valent pneumococcal conjugate vaccine into the WMA population in Arizona. This study is very beneficial to our understanding of the dynamics of pneumococcal epidemiology in general and of the impact, both positive and negative, that vaccine introduction may have in a population. The study performed by Lacapa et al. [2] was an observational study and compared rates of disease in the time period before vaccine introduction with rates in the time period after vaccine introduction. One of the key observations of the study was that the incidence of invasive disease due to the 7 serotypes covered in the vaccine was reduced substantially in this high-risk population. The observed 92% reduction in the rate of infectious pneumococcal disease among high-risk children aged ⩾5 years is consistent with the reductions observed in other studies. However, the rate of disease among vaccinated WMA children still remains higher than that among the general population. There are 2 reasons for this. First, given the higher initial rate of disease, a 92% reduction in the rate of disease still leaves a substantial disease burden in the population. Second, prior to vaccine introduction, the proportion of overall invasive disease due to the 7 serotypes covered in the vaccine was only 56.2% among WMA children, compared with ∼90% in the general population. The reasons for this difference are not known, but the difference is consistent with both the variability of coverage of the 7-valent conjugate vaccine globally and with changes in serotype distribution observed over time in the Native American population and other populations globally. Although other vaccines that contain 10 and 13 pneumococcal serotypes are in development and would provide broader coverage, until we better understand the reasons that serotype distributions vary geographically and over time, strategic vaccine design will be hampered. Of note, serotype replacement by pneumococcal serotype 19A was not observed in the WMA population in this study [2]. Unlike studies performed using the Active Bacterial Core surveillance system of the Centers for Disease Control and Prevention, in which substantial increases in the rate of disease due to serotype 19A have been seen after introduction of the vaccine, the rate of disease due to serotype 19A in the WMA population has remained the same or has decreased after vaccine introduction. This observation offers the opportunity to evaluate what factors may, in fact, be causing the increase in disease due to serotype 19A in the United States and in other populations. In studies by Whitney et al. [3] using the Centers for Disease Control and Prevention surveillance program, a 3.2-fold increase in the number of cases of invasive pneumoccocal disease due to serotype 19A was observed after vaccine introduction. In another study involving Alaskan Native children, Singleton et al. [4] observed a 140% increase in the number of pneumococcal disease cases due to non-vaccine serotypes, mostly due to serotype 19A. Both of these results have caused concern that some of the positive impact on disease burden attributable to vaccine introduction may be eroded by “replacement disease” due to serotype 19A or others. However, an increase in disease due to serotype 19A was not observed in this study by Lacapa et al. [2]. So, what is going on here? What factors exist in other populations that do not exist in the WMA population? We can begin to assess this by looking at a recent publication by Munoz-Amalgro et al. [5]. In the article, the authors report a 531% increase in the rate of invasive pneumococcal disease due to non-vaccine serotypes, including 19A and 6A, among children aged ⩾2 years in Barcelona, Spain, after introduction of vaccine. However, this dramatic change occurred despite the fact that vaccine coverage levels were quite low in Spain. The percentage of children who had received ⩾1 dose was 36% in 2005 and 47% in 2007. This low vaccine coverage rate and the fact that the change in disease epidemiology was greater than that seen in any other population argue against the vaccine as the major causative factor. Antibiotic, especially macrolide, usage patterns impact serotype distribution. In fact, the use of macrolides, especially azithromycin, for the treatment of minor infections in children has been reported to be quite common in Spain. Is it possible that nonvaccine factors, such as antibiotic use patterns, may play a role in changes of vaccine serotype distribution? Evidence for this possibility comes from studies by Dagan (R. Dagan, personal communication). In Israel, a 68% increase in the proportion of otitis media cases caused by serotype 19A was observed in the Bedouin population, associated with an increase in the use of azithromycin in that population. Similarly, in South Korea, an increase in the proportion of disease due to serotype 19A was observed prior to vaccine introduction. These changes cannot be attributed to vaccine use, because conjugate pneumococcal vaccine was not used in Korea or Israel at the time that these changes were observed. To return to the study by Lacapa et al. [2], the observation that there was no increase in pneumococcal serotype 19A in the WMA population led the authors to look at antibiotic usage patterns. Health care on the WMA reservation is provided by the Indian Health Service, and the use of antibiotics within this system is tightly monitored and controlled. Of interest, azithromycin is approved only for the treatment of sexually transmitted diseases and for children with a documented penicillin allergy. During the year 2007, only 12 children were treated with azithromycin within the Indian Health Service program. By contrast, within the Alaskan Native population, where serotype 19A “replacement” was observed, azithromycin is used more widely and is approved for the treatment of pneumonia and otitis media. These observations do not indicate that the introduction of 7-valent conjugate vaccine does not and cannot cause changes in the seroepidemiology of disease or cause “replacement” disease. However, now that vaccine has been introduced, our knowledge of such changes comes from uncontrolled observational studies. The possible influence of other factors, such as changes in antibiotic use patterns, on disease epidemiology can be large. Therefore, we must exercise caution in attributing such changes only to vaccine introduction, because other factors may play a significant or even a larger role. Potential conflicts of interest. S.B. serves on the speakers' bureau of Novartis and Merck and as a consultant for Novartis and GlaxoSmithKline.