Respiratory syncytial virus (RSV) is the leading cause of pneumonia and bronchiolitis in infants and young children worldwide.1,2 In the United States, RSV infections have been estimated to cause between 75,000 and 125,000 hospitalizations annually among infants and young children. Infants with cardiac or pulmonary abnormalities or with a compromised immune system are at increased risk for severe disease.3,4 Although a vaccine against RSV is not available, immune prophylaxis is effective in reducing the risk of severe RSV disease among certain groups of high-risk infants and young children. Palivizumab reduces hospitalization rates by 40% to 75% among high-risk infants.5–7 The American Academy of Pediatrics recommends immune prophylaxis to carefully defined high-risk infants and children generally starting in early November and continuing through the RSV season offset week in early March.6
The risk of serious RSV disease must be balanced against the substantial cost of palivizumab.8 The most cost-effective administration schedule would follow precisely the RSV season observed in a local community. Ideally, a better understanding of RSV seasonality would allow clinicians to administer prophylaxis during the months when RSV is circulating and ensure at-risk infants are protected while avoiding the cost of unnecessary monthly administration of prophylaxis.
Since 1983, the Centers for Disease Control and Prevention has conducted surveillance tracking the seasonal and geographic trends of RSV through the National Respiratory and Enteric Virus Surveillance System (NREVSS).9,10 These data have provided information on the timing of RSV seasons, including variation in the timing of the onset and offset weeks of the RSV season. In this report, we refer to these previous analyses and provide updated data to explore whether surveillance data can be used to more accurately define the timing of RSV immune prophylaxis.
MATERIALS AND METHODS
National Respiratory and Enteric Virus Surveillance System.
A voluntary laboratory-based surveillance system, NREVSS, collects the weekly number of specimens tested and the number of positive results for RSV, parainfluenza 1–4, influenza, respiratory and enteric adenovirus, rotavirus, and human metapneumovirus. Participating laboratories report antigen detection, virus isolation, electron microscopy, and polymerase chain reaction data. When NREVSS started in 1983, laboratories submitted data via mailed cards and later submitted data by telephone; currently, all laboratories submit data electronically. The Centers for Disease Control and Prevention compiles these data and updates the NREVSS website weekly; these data can be accessed at http://www.cdc.gov/ncidod/dvrd/revb/nrevss/rsvtre1.htm.
From July 1990 through June 2006, 79 to 100 laboratories reported to NREVSS each season. In the autumn of 2006, a data sharing agreement with Surveillance Data Inc. increased the number of reporting laboratories to 389. Figure 1 shows the locations of current participating laboratories. Data are aggregated by U.S. Census regions, with 65 (17%) in the Northeast, 148 (38%) in the South, 106 (27%) in the Midwest, and 70 (18%) in the West. Approximately 75% of laboratories are private or teaching hospitals, with children's hospitals, public health laboratories, public hospitals, and commercial laboratories comprising the remaining laboratories.
In a study by Mullins et al,10 national and regional trends were assessed using data from all laboratories reporting to NREVSS from 1990 through 2000. Local community trends were determined based on a subset of 18 laboratories with representative locations, consistent reporting, and greater than 50 positive RSV test results each season during the study period. Variation within localities was examined by Mullins et al by selecting a pair of laboratories from the following locations: Missouri (Kansas City and St. Louis), California (Long Beach and Los Angeles), and Nebraska (both in Omaha).
To assess trends and variations in RSV activity between regional, state, and local areas during the period between July 2000 and June 2006, we analyzed reports from laboratories with more than 26 weeks of consistent reporting, at least 20 specimens tested, and at least 10 positive results.
Season onset, peak, offset, and duration were calculated for each year and each laboratory, as previously described in annual NREVSS reports. Briefly, season onset was defined as the first of 2 consecutive weeks where at least 2 specimens tested positive and the percent positivity was at least 10% by antigen detection. The season offset was defined as the last of 2 consecutive weeks where at least 2 specimens tested positive and the percent positivity was at least 10% by antigen detection. Duration was defined as the difference between the onset and offset weeks. As outlined in the Mullins et al study,10 individual laboratory results were then compared with median onset and duration measurements of its U.S. Census region and national median measurements for each year. The magnitude of differences, measured in weeks, was tabulated for each season by subtracting the corresponding national or regional median seasonal measurement from the sampled laboratory measurement. The Wilcoxon rank sum test was used to test for statistical significance when comparing regional and national trends.
In the United States, a distinct seasonal winter peak of RSV activity is experienced each year, with limited variation in season onset and offset. The median week of onset reported by Mullins et al from 1990 through 2000 was week 51 (late December), ranging from week 46 (late November) to week 3 (late January).10 The season duration had less variability with a median duration of 15 weeks (range, 13–17 weeks), with an offset week dependent on the variability of the onset. The median week of offset from 1990 through 2000 was week 13 (end of March). These trends have continued as shown by data from July 2000 through 2006 based on data from laboratory/years with at least 26 weeks of continuous reporting and reports of ≥20 specimens tested and ≥10 specimens positive (Fig. 2).
Figure 3 illustrates the RSV seasonal activity by U.S. Census region from July 2000 through June 2006. RSV activity varied some by U.S. Census regions as previously noted by Mullins et al.10 The Southern region consistently exhibited an earlier onset than that of the other regions. The median onset for the South was week 47 (late November; range, week 41–51) and was the earliest (P < 0.01). This onset week occurred 1 to 8 weeks earlier than the national median onset week for all 10 seasons of RSV activity analyzed. The South also had the longest season duration at 16 weeks (range, 13–20 weeks; P < 0.05). In contrast, the Midwest region exhibited the latest median onset at week 1 of the following year (early January; range, week 50 to week 5 of the following year; P < 0.05) and the shortest duration at 13 weeks (range, 10–16 weeks; P < 0.05), when compared with the rest of the nation. Figure 4 illustrates these trends and summarizes the median onset weeks for 1990 to 2000 for the nation and each of the regions. As shown in the figure, compared with the national trend, South appears to have an earlier onset and the Midwest to have a later onset of RSV activity.
In addition to regional variation, the RSV season onset and offset periods can vary widely within regions, as noted by recent detections from Florida compared with all other states in the South region and the nation (Fig. 5). Surveillance data from Florida shows the onset of RSV outbreaks occurring typically as early as August and persisting for longer than usual (median duration, 40 weeks). Three laboratories (2 in Miami and 1 in Orlando) reported an especially early onset, as early as July 24 for the 2005 to 2006 respiratory season, whereas the median onset week for the South region (excluding the 3 Florida laboratories) was October 23 (Fig. 5). Again, in the 2006 to 2007 season, the onset for Florida occurred in early July, compared with mid-October for the South.11
The extent of variation at the level of individual laboratories is illustrated by NREVSS data from July 1990 to June 2000.10 Of 177 laboratory-seasons analyzed, 21% of the onset weeks for a given laboratory varied by 5 to 8 weeks; moreover, 8% of the laboratory seasons varied by more than 8 weeks from the median for its region. Eleven percent of the onset weeks for a given laboratory were identical to the median regional onset. Season duration was even more variable than season onset. Figure 6 illustrates the percent of laboratories varying from the regional onset and duration median.
Mullins et al10 also observed substantial variations in season onset between areas represented by laboratories in close proximity, and the extent of variation was noted to increase as distance increases. For example, 89% of season onsets were within 4 weeks for 2 laboratories located 5 miles apart, 80% for 2 laboratories located 25 miles apart, and 67% for 2 laboratories located 250 miles apart.
The timing of RSV seasons can vary substantially from year to year as indicated by detections from the same laboratory. Of 177 laboratory-seasons analyzed, 74% of the laboratories had season onset for a given year that was within 4 weeks of the median values for that laboratory. Figure 7 shows the variability in timing of RSV detections for 4 years for one Midwest laboratory.
Ideally, administration of RSV immune prophylaxis would be correlated to the actual onset and offset of the RSV season in a community. Administration of palivizumab before circulation of RSV begins or after circulation stops is costly and results in unnecessary use of health care resources. On the other hand, if prophylaxis is initiated after widespread circulation has begun or stopped before widespread virus circulation has ended, susceptible infants may be unprotected and at increased risk of complications. It was hoped that surveillance data such as those provided through NREVSS would provide a means to tailor administration of immune prophylaxis to the actual onset and offset of RSV season in the community. Unfortunately, our analysis of NREVSS data suggests that this will be a difficult goal to achieve. Although the surveillance data identifies the onset of RSV season, the onset is only identified as the season starts and not before onset, allowing little time to schedule patients for the initial dose. Furthermore, our analysis of NREVSS data suggests that the prior year's data is a fairly imprecise predictor of season onset, and even data from nearby laboratories during the same year may be a relatively poor predictor of the timing of the RSV season in a given community. In addition, timing of the RSV season may vary substantially in the same year among communities in close proximity. These data, as well as studies of RSV strains from different communities,12,13 demonstrate that RSV circulation is primarily a local and not a regional or national phenomenon. However, there are some regional differences in the timing of RSV circulation in communities, as indicated by the tendency for the Southern region to have earlier onsets, and even more dramatically by the very early onset report for some communities in southern Florida. Local community data can be used to determine whether RSV circulation patterns for that community fit within the national patterns. Currently, we are analyzing NREVSS data to better understand how local data can best be used to guide local immune prophylaxis. RSV surveillance data, such as that provided by NRVESS, should provide a means to refine timing of RSV immune prophylaxis in some settings.
1. Steiner RW. Treating acute bronchiolitis associated with RSV. Am Fam Physician
2. Stensballe LG, Devasundaram JK, Simoes EA. Respiratory syncytial virus epidemics: the ups and downs of a seasonal virus. Pediatr Infect Dis J
. 2003;22(2 suppl):S21–S32.
3. Shay DK, Holman RC, Newman RD, Liu LL, Stout JW, Anderson LJ. Bronchiolitis-associated hospitalizations among US children, 1980–1995. JAMA
4. Black CP. Systematic review of the biology and medical management of respiratory syncytial virus infection. Respir Care
5. Welliver RC. Review of epidemiology and clinical risk factors for severe respiratory syncytial virus (RSV) infection. J Pediatr
. 2003;143(5 suppl):S112–S117.
6. Meissner HC, Long SS. Revised indications for the use of palivizumab and respiratory syncytial virus immune globulin intravenous for the prevention of respiratory syncytial virus infections. Pediatrics
. 2003;112(6 pt 1):1447–1452.
7. American Academy of Pediatrics Committee on Infectious Diseases and Committee on Fetus and Newborn. Prevention of respiratory syncytial virus infections: indications for the use of palivizumab and update on the use of RSV-IGIV. Pediatrics
8. Schrand LM, Elliott JM, Ross MA, Bell EF, Mutnick AH. A cost-benefit analysis of RSV prophylaxis in high-risk infants. Ann Pharmacother
9. Gilchrist S, Török TJ, Gary HE, Alexander JP, Anderson LJ. National surveillance for respiratory syncytial virus, United States, 1985–1990. J Infect Dis
10. Mullins JA, Lamonte AC, Breese JS, Anderson LJ. Substantial variability in community RSV season timing: an analysis using the National Respiratory and Enteric Viruses Surveillance System (NREVSS). Pediatr Infect Dis J
11. Brief report: respiratory syncytial virus activity–United States, 2005–2006. MMWR Morbid Mortal Wkly Rep
12. Peret TCT, Hall CB, Hammond GW, et al. Circulation patterns of group A and B human respiratory syncytial virus genotypes in five communities in North America. J Infect Dis
13. Peret TCT, Hall CB, Schnabel KC, Golub JA, Anderson LJ. Circulation patterns of genetically distinct group A and B strains of human respiratory syncytial virus in a community. J Gen Virol
Keywords:© 2007 Lippincott Williams & Wilkins, Inc.
National Respiratory and Enteric Virus Surveillance System; immune prophylaxis; palivizumab; antigen detection; RSV seasonality