Respiratory Syncytial Virus in Indonesian Children

To the Editors:

Globally, respiratory syncytial virus (RSV) is estimated to cause between 66,000 and 199,000 deaths among young children every year.¹ Several factors, including a better understanding of RSV disease burden, improvements and vaccine technology and the remarkable success of vaccines such as rotavirus vaccines, have led to a renaissance in interest in RSV vaccine development. Currently, there are more than a dozen RSV vaccine candidates in preclinical or early clinical development.

One of the critical questions about RSV vaccines is whether they can be effective early enough in life to prevent the most severe disease. Studies in industrialized as well as developing nations suggest that the highest incidence of hospitalization for RSV is in the first 6 months of life, and in particular, during the second and third month of life.^2–5 Developing vaccines to induce active immunity at such a young age would be challenging.

The study by Simões et al⁶ in September’s Pediatric Infectious Disease Journal presents an intriguing finding from a study conducted in Indonesia a decade earlier—that fewer than 5% of cases of RSV lower respiratory infection among periurban and semirural Sundanese infants occur in the first 6 months of life. If true, and if this pattern occurs in other developing world settings, this opens the possibility that an RSV vaccine that does not become effective until age 6 months could prevent the bulk of serious RSV-related disease in those settings.

Given this remarkable finding and implications for RSV vaccination, some additional clarification would help readers to better evaluate the study:

• The methods section states that infants 0–2 months of age were not recruited in the longitudinal cohort. Was this the case and what was the rationale for the exclusion? Also, note that the age groups in the first year of life are at 3-month intervals, not 2-month intervals, as stated throughout the manuscript.
• The study appears to have lasted 28 months (as stated in the methods section, “Longitudinal Cohort”), but parts of the manuscript suggest it lasted 24 months (as stated in the methods section, “Newborn Cohort”) or 3 years (as stated in the discussion).
• The data in Table 2 are internally inconsistent in that the incidence rates are not equal to what one would calculate based on the child-years of observation and the number of RSV lower respiratory infection cases in the table. Could this discrepancy be explained or the numbers corrected?
• The results section states that in “the second and third years, more RSV was detected because of the use of PCR.” The methods section describes technical details of the polymerase chain reaction (PCR) testing but does not state that the testing was used for only part of the study. When did PCR testing begin in the study, and what testing was used before PCR was used in the study?
• The recruitment rate of newborn infants during the latter part of the study (382 in 16 months) was about half that in the first part of the study (558 per 12 months). As it is unlikely that the birth rate fell by half during a 1-year time period, what is the explanation for the lower recruitment rate in the second half of the study, and is it true, as stated in the discussion, that the study “recruited virtually all newborn babies in the first 3 years in cohort (almost 1000)…” ?
• In Figure 2, because of the study design, there may have been substantial differences in the proportion of children in the various age groups in the 7 time periods. One way to account for this would be to adjust each observation for the number of child-years of observation represented in the particular time point and age group.

In general, because risk of RSV disease can vary greatly during the first year of life and because RSV can be very seasonal, results of cohort studies such as the one presented here may be influenced by study design. For example, in the United States, one could consider a study that recruited a cohort of infants during the month of November, the start of the annual RSV season, and followed this cohort for a year. Such a study would tend to find most RSV infections occurring at a young age, as there would be little transmission of RSV during the second half of the study (ie, May through October). In contrast, a similar study that recruited infants only during the month of May would find infections occurring at an older age. Either study would be misleading.

Although this phenomenon may have played a role in skewing the age distribution of RSV lower respiratory infection cases in this study, and the use of PCR testing only in the latter half of the study may have had a similar effect, neither of these would explain the low number of cases during the first 6 months of life in this group. Confirming this finding would have important implications for RSV vaccine development.

Gregory L. Armstrong, MD

Division of Viral Diseases

National Center for Immunization and Respiratory Diseases

Centers for Disease Control and Prevention

Atlanta, GA