PEDIATRICS Vol. 118 No. 2 August 2006, pp. e258-e264 (doi:10.1542/peds.2005-2874)
ARTICLE |
Rotavirus Vaccination and Intussusception: Can We Decrease Temporally Associated Background Cases of Intussusception by Restricting the Vaccination Schedule?
a Viral Gastroenteritis Section
b Office of the Director, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
OBJECTIVE. The first rotavirus vaccine that was licensed in the United States, RotaShield, could have prevented the enormous burden of rotavirus diarrhea in American children but left instead the unfortunate legacy that live oral rotavirus vaccines may be associated with a serious but rare adverse event: intussusception. Although large trials indicate that the next generation of rotavirus vaccines is not associated with this complication, many children likely will develop intussusception by chance alone in the 2-week window after immunization, raising concerns about whether these cases might be "caused" by the vaccine. Our objective for this study was to model and compare the number of temporally associated intussusception events that are expected by chance alone under 2 rotavirus vaccination strategies.
METHODS. We used national vaccine coverage rates and age-specific incidence of intussusception by months to model the number of temporally associated cases of intussusception that are expected by chance alone for 2 potential vaccination strategies: a strict schedule, limiting immunization to children within 1 month of the designated age for each dose (ie, 60–89, 120–149, and 180–209 days for doses 1, 2, and 3, respectively) versus a free schedule whereby infants are immunized whenever they appear for their routine vaccines up to 1 year of age.
RESULTS. The number of intussusception events during the 2-week postvaccination window was 24% lower for the strict versus the free schedule (138 vs 182, respectively). This reduction was attributable largely to the smaller number of infants who were immunized fully under the strict schedule (vaccine coverage for 3 doses, 67% vs 89%). The cumulative risk for intussusception's occurring by chance in the 2-week postvaccination window essentially was the same between schedules (4.59 vs 4.76 per 100000 doses). Most cases occurred after the second or third dose.
CONCLUSIONS. Although an age-restricted vaccination schedule substantially reduced the number of intussusception events that were observed in the 2-week postvaccination window when compared with a schedule with fewer restrictions, this decrease was attributable to a lower rate of vaccine coverage rather than a safer schedule of vaccination. The risk for intussusception did not differ significantly between vaccination strategies. Public health policy and education messages for physicians and parents should be developed to anticipate intussusception events that will occur by chance alone but are temporally related to rotavirus vaccination.
Key Words: rotavirus • vaccine • intussusception • vaccine safety • vaccine adverse event
Abbreviations: KID—Kids' Inpatient Database • DTP—diphtheria-tetanus toxoids-pertussis
Rotavirus, the most common cause of severe diarrhea in children, may soon be prevented with live oral vaccines that will be licensed in the United States and abroad.1,2 The first vaccine against rotavirus, RotaShield, was licensed by the US Food and Drug Administration in 1998 and was recommended by the Advisory Committee on Immunization Practices of the Centers for Disease Control and Prevention and the American Academy of Pediatrics for the routine immunization of children in 3 oral doses at ages 2, 4, and 6 months.3 Nine months later, after 
600000 infants had received >1.2 million doses of the vaccine, intussusception was identified as a rare adverse event of the vaccine that occurred primarily in the 2 weeks after the first dose. This association with intussusception led to the withdrawal of RotaShield, the legacy of which remains a major impediment for the acceptance of the next generation of live oral vaccines by pediatricians and the community.4–6
Manufacturers of the new rotavirus vaccines, Merck and GlaxoSmithKline, are keenly aware that their vaccines might encounter the same problem.1,2 After consultation with the Food and Drug Administration, each mounted large safety trials with >60000 infants to ensure the safety of their vaccines within 30 to 42 days after each dose, a period defined from the RotaShield experience as the high-risk window.7 These studies found no difference in intussusception rates between recipients of vaccine versus placebo, suggesting that the adverse event might be specific to the RotaShield vaccine.1,2
After withdrawal of RotaShield, epidemiologic studies of natural intussusception identified another potential clue to the problem. The incidence of intussusception is low in infants in the first 3 months of life and then rises 10-fold to peak in infants at 6 months of age.8 Although the first dose of the RotaShield vaccine was recommended for infants at 2 months of age to prevent disease from occurring early in the first year of life, "catch-up" vaccination was permitted up to 6 months of age. Simonsen et al9 found that these older catch-up infants, who represented 38% of those who were immunized with the first dose, experienced 80% of the intussusception events, whereas only 20% of intussusception events occurred among infants who were immunized at <3 months of age. They concluded that the risk for intussusception might have been substantially reduced if infants had been immunized strictly on schedule at 2, 4, and 6 months of age.
Although the new vaccines have clean profiles in large-scale safety trials to date, the perception that live oral rotavirus vaccines might cause intussusception remains a concern for physicians, parents, and the immunization community. Because intussusception is the most common cause of intestinal obstruction in infants at an incidence of 25 to 50 cases per 100000 infants in the first year of life, we anticipate that some infants will develop intussusception within the 2 weeks after a vaccination dose by chance alone. Given the public awareness of this severe adverse event, however, a parent whose child develops intussusception in this period may believe that his or her child's illness was caused by rotavirus vaccine instead of being a temporal, chance event.
Because the risk for natural intussusception is substantially lower in younger infants, we evaluated whether the number of these "chance" cases could be reduced by using an age-restricted vaccination strategy (strict schedule) compared with a strategy with few age restrictions (free schedule). The strict schedule potentially would reduce the number of intussusception events after immunization but decrease the number of infants who receive the vaccine, whereas the free schedule would provide greater vaccine coverage but could lead to a greater number of intussusception events in the follow-up period. We therefore modeled these 2 possible strategies for immunization to gauge the impact of each on the number and the rate of intussusception events observed during follow-up and on vaccine coverage of the target population.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vaccination Strategies
Because Merck's pentavalent rotavirus vaccine has been approved for use in the US market, we modeled different strategies for administration of a rotavirus vaccine on a similar 3-dose schedule. The 2 immunization strategies evaluated, the strict schedule and the free schedule, differed only in the age at which immunizations could be administered. In the strict schedule, a vaccine dose would be administered no later than 29 days after the currently recommended immunization visits (ages 2, 4, and 6 months) and would be withheld outside these ages; that is, infants would not receive dose 1 if they were older than 89 days, dose 2 if older than 149 days, or dose 3 if older than 209 days. For the free schedule, the vaccine would be recommended for administration at ages 2, 4, and 6 months, but infants would receive vaccine whenever they arrived for immunizations up to age 1 year. For both strategies, we limited our analysis to infants who were younger than 1 year and assumed that the new vaccines would not cause intussusception (ie, all intussusception events are background in the population and occur by chance alone).
The Model
We modeled the expected numbers of intussusception events that would occur by chance alone and the fraction of the birth cohort that would receive the vaccine according to our 2 vaccination schedules. We assumed the US birth cohort to be 4.1 million per year and that there is no significant seasonality in either the occurrence of intussusception events or birth rates.10–12 We chose our period of interest to be the 2 weeks after receipt of any dose of the vaccine because most cases of intussusception that were temporally associated with RotaShield occurred within 14 days of the first or second dose.7 Although immunization with RotaShield was halted before enough children received the third dose to identify a risk if it existed, we evaluated the third dose for a possible risk association.
Sources of Data
We estimated the incidence of intussusception by determining the average number of intussusception cases by 1-month age groups using results from published studies from California for 1990–1997, New York State for 1989–1998, a survey in 10 states, and hospital discharge estimates from the 2000 Kids' Inpatient Database (KID).10–12 KID is a large, national, all-payer, hospital, pediatric discharge database that is designed to generate national pediatric hospitalization estimates for rare and common childhood conditions.13 Because rotavirus vaccine would be administered at the same time as diphtheria-tetanus toxoids-pertussis (DTP) vaccine, we based estimates of US vaccine coverage on data reported by the 2003 National Immunization Survey.14 We used these inputs to calculate the expected number of intussusception events by chance alone during the 2-week period after receipt of any dose of rotavirus vaccine. The 95% confidence intervals for the expected number of cases by month of age were calculated by assuming a Poisson distribution.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We first estimated the mean incidence of intussusception-associated hospitalizations among children in their first year of life (Table 1). We selected rates from the most recent year available from 3 previous studies (31, 39, and 42 hospitalizations for intussusception per 100000 infants) to obtain an overall mean of 37 hospitalizations per 100000 infants. We then plotted the age-specific data from these studies and the 2000 KID national database to obtain the median incidence by age in months (Fig 1). These data exhibited parallel trends: the incidence of intussusception-associated hospitalizations was low during the first 2 months of life, rose steadily to peak between months 5 and 6, and then decreased slowly each month for the remainder of the first year of life.
|
|
We then calculated the month-specific rates of intussusception by applying the median proportion of intussusception cases that would be expected for each month of age to the total number of intussusception hospitalizations that would be expected during the first year of life (Fig 2). Because rotavirus vaccine will be administered at the same visits as DTP immunizations, we assumed that for a mature rotavirus program, the levels of coverage would be the same as for DTP in the United States (Fig 3). As expected, vaccination coverage decreases slightly for each of the recommended doses at ages 2, 4, and 6 months. Overall, 88.1%, 77.1%, and 66.6% of children are currently immunized with doses 1, 2, and 3 on schedule, respectively.
|
|
By using the anticipated 3-dose vaccination schedule and the month-specific incidence rates for intussusception, we calculated the estimated number of infants who would be immunized with 1, 2, or 3 doses and the number of intussusception cases that would be anticipated for each strategy of vaccination (Table 2). For the free schedule, 182 cases of intussusception would be expected by chance alone within 2 weeks of the administration of any dose. The majority of these cases would occur in infants of 120 to 149 days of age (49 cases) and 180 to 209 days of age (60 cases). Doses 1, 2, and 3 of rotavirus vaccine would be administered to 98%, 96%, and 89% of the birth cohort, respectively. Under the strict schedule, a total of 138 intussusception cases per year would be anticipated in the 2 weeks after any dose of rotavirus vaccine. Again, most cases would occur in infants with the second and third doses at ages 120 to 149 days (48 cases) and 180 to 209 days (54 cases). Vaccination coverage under this strategy would be considerably lower, with the first dose reaching 88% of the birth cohort, 77% with the second dose, and 67% with all 3 doses.
|
We then calculated the risk for an intussusception event per 100000 doses per year (Table 3). When compared with the free schedule, we would expect under the strict schedule 44 fewer intussusception events but also 934800 fewer infants to be immunized fully and 2123800 fewer doses administered. The per-dose risk (eg, 0.61 vs 0.70 per 100000 doses for dose 1) and cumulative dose risk (4.59 vs 4.76 per 100000 doses) of intussusception would be comparable, although the risk for intussusception per dose would be slightly higher in the infants who were on the free schedule (1 event in 63888 vs 1 in 68868). Of note, relatively few of the intussusception events would follow the first vaccine dose. The largest number would be expected after the third dose because the incidence of intussusception is relatively high among infants aged 5 to 6 months.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The withdrawal of RotaShield from the US market in 1999 left a heavy legacy of concern about the risk for intussusception that might occur with the next generation of live oral rotavirus vaccines. Despite evidence from large safety trials of the new bovine pentavalent vaccine (RotaTeq) and the attenuated human vaccine strain (Rotarix) that neither is associated with intussusception, this concern will not be put to rest until additional data are gathered from postlicensure surveillance.1,2 In the interim, cases of intussusception are likely to occur in the 2-week period of suspected high risk after each dose, and it initially will be impossible to distinguish whether these events are or are not associated with the vaccine. Assuming that these events occur by chance alone and are unrelated to the vaccine itself, different vaccination strategies that are based on age at the time of immunization might lower the incidence of observed intussusception events, thereby facilitating the acceptance of rotavirus vaccine by communities and pediatricians and limiting investigation of purported vaccine-associated cases. Our model compared 2 strategies of immunization, a strict versus a free schedule, with regard to the risk for temporally associated background intussusception's occurring by chance and the level of vaccine coverage. Our findings provide insights into recommendations for using the next generation of rotavirus vaccines.
First, because the incidence of intussusception rises markedly with increasing age during the first year of life, we expected that the strict schedule would substantially reduce a child's risk for intussusception for all 3 doses as well as the risk per dose. Although the number of intussusception cases anticipated was 24% less (n = 44) with the strict schedule, this was attributable mostly to lower vaccine coverage and not to any substantial difference in the risk for intussusception after each dose. In fact, the cumulative risk for intussusception among infants who received all 3 doses in the 2 schedules was nearly the same (4.59 vs 4.76 events per 100000 infants fully vaccinated). With the strict schedule, 23% fewer infants would receive their full complement of 3 doses of vaccine.
A second finding of the model was that for both strategies, the smallest number of intussusception events occurred in the 2 weeks after the first dose, with at least twofold more expected after a second dose. Therefore, if intussusception were to be associated with the new vaccine, then efforts to identify vaccine-associated cases should focus on events that occur after a first dose, not only because most intussusception cases that were associated with RotaShield occurred after the first dose but also because the rate of background disease is lowest in this period, thereby amplifying the impact of vaccine.
Our findings provide important insights into beginning the next immunization program with a live oral rotavirus vaccine. Even if the new vaccine were free of the risk for intussusception, each year the program still would contend with 20 to 30 cases after the first dose and >100 intussusception cases in the 2-week window after any dose. Until we can conclude that the vaccine is not associated with intussusception, we will need an epidemiologic strategy to measure this low-frequency risk in postlicensure surveillance and a communications strategy to address the concerns of parents whose children might be affected before all data are assessed. Several statistical models have been suggested to detect and assess these rare adverse events. For example, the case-series approach of Farrington15 can be used to identify each intussusception case among vaccinees and to evaluate its relationship with the time of vaccine administration. In the absence of a causal relationship, the cases should occur randomly over time and not be clustered around the high-risk period in the 2 weeks after administration of a rotavirus vaccine. Alternatively, sequential analysis can be done (as used by Merck) to assess events that occur among vaccinees versus placebo recipients to determine whether threshold limits of confidence are exceeded, suggesting a causative relationship. The data presented here provide rates of intussusception that would be expected by chance alone and provide an index to monitor cases that might be expected as more doses are administered.
One option not considered here is a mixed schedule, in which an infant would receive the first dose before age 89 days (as on the strict schedule) but would be allowed to receive any of the next 2 doses whenever he or she appeared for the next immunization (as on the free schedule). One reason for this approach is that because RotaShield was associated with risk for intussusception primarily after the first dose, physicians and parents are likely to be most concerned about complications after this dose. In addition, the mixed schedule reflects that the cumulative risk for natural intussusception did not differ significantly between the strict and the free schedules and would allow for greater vaccine coverage.
Several caveats must be considered in the interpretation of these data. First, our results cannot be generalized to settings where children do not routinely arrive for immunizations on schedule and vaccines are given at an older age than recommended. Particularly in the developing world, where rotavirus diarrhea often is fatal, limiting immunization exclusively to infants who arrive on schedule could dramatically decrease the number of children who receive these life-saving vaccines, an unacceptable scenario. We believe, however, that this study provides a model for countries with different vaccine coverage patterns to assess the risk for baseline intussusception in their populations and to identify a favorable age profile for administration of rotavirus vaccine. Second, we have taken the high-risk period of intussusception to be the 2 weeks after any of the 3 doses of rotavirus vaccine. In fact, for RotaShield, the adverse events occurred largely after the first dose, with lower risk after the second dose and none after the third. Clearly, fewer children were immunized with 2 or 3 doses, making it more difficult to establish an excess risk in these groups if it existed. In our analysis, fewer background cases would be observed if the time period were limited to the 2 weeks after the first dose alone (22 cases under the strict schedule; 28 cases under the free schedule). Finally, although we modeled intussusception events on the basis of the Merck vaccine that will be administered as 3 oral doses, the GlaxoSmithKline vaccine requires only 2. Many excess cases in our model occurred after the third dose, so a vaccination schedule that require only 2 doses would lower the number of natural intussusception cases observed.
Until a final conclusion can be drawn on the safety of a new rotavirus vaccine, policy makers and manufacturers will have to decide how best to deal with these severe adverse events in the interim period while postlicensure surveillance monitors for any potential association between the vaccine and intussusception. These analyses should provide some insights into these deliberations and how intussusception events that are temporally associated with rotavirus vaccination should be handled once the next generation of vaccines is introduced.
|
ACKNOWLEDGMENTS |
|---|
We thank Robert Holman for helpful comments and Kathleen Murray for editorial assistance.
|
FOOTNOTES |
|---|
Accepted Feb 15, 2006.
Address correspondence to Roger I. Glass, MD, PhD, Centers for Disease Control and Prevention, Viral Gastroenteritis Section (Mailstop GO4), 1600 Clifton Rd, NE, Atlanta, GA 30333. E-mail: rig2{at}cdc.gov
The authors have indicated they have no financial relationships relevant to this article to disclose.
The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the funding agency.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Ruiz-Palacios GM, Perez-Schael I, Velazquez FR, et al. Safety and efficacy of an attenuated vaccine against severe rotavirus gastroenteritis.
N Engl J Med. 2006;354
:11
–22
[Abstract/Free Full Text] - Vesikari T, Matson DO, Dennehy P, et al. Safety and efficacy of a pentavalent human-bovine (WC3) reassortant rotavirus vaccine.
N Engl J Med. 2006;354
:23
–33
[Abstract/Free Full Text] - Centers for Disease Control and Prevention. Rotavirus vaccine for the prevention of rotavirus gastroenteritis among children. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep. 1999;48(RR-2) :1 –20
- Centers for Disease Control and Prevention. Intussusception among recipients of rotavirus vaccine: United States, 1998–1999. MMWR Morb Mortal Wkly Rep. 1999;48 :577 –581[Medline]
- Centers for Disease Control and Prevention. Withdrawal of rotavirus vaccine recommendation. MMWR Morb Mortal Wkly Rep. 1999;48 :1007[Medline]
- Iwamoto M, Saari TN, McMahon SR, et al. A survey of pediatricians on the reintroduction of a rotavirus vaccine. Pediatrics. 2003;112 :6 –10[CrossRef]
- Murphy TV, Gargiullo PM, Massoudi MS, et al. Intussusception among infants given an oral rotavirus vaccine.
N Engl J Med. 2001;344
:564
–572
[Abstract/Free Full Text] - Hutchison IF, Olayiwola B, Young DG. Intussusception in infancy and childhood. Br J Surg. 1980;67 :209 –212[Web of Science][Medline]
- Simonsen L, Viboud C, Elixhauser A, Taylor RJ, Kapikian AZ. More on RotaShield and intussusception: the role of age at the time of vaccination. J Infect Dis. 2005;192 :S36 –S43
- Parashar UD, Holman RC, Cummings KC, et al. Trends in intussusception-associated hospitalizations and deaths among US infants.
Pediatrics. 2000;106
:1413
–1421
[Abstract/Free Full Text] - Chang HG, Smith PF, Ackelsberg J, Morse DL, Glass RI. Intussusception, rotavirus diarrhea, and rotavirus vaccine use among children in New York State.
Pediatrics. 2001;108
:54
–60
[Abstract/Free Full Text] - Simonsen L, Morens DM, Elixhauser A, Gerber M, Van Raden M, Blackwelder WC. Effect of rotavirus vaccination programme on trends in admission of infants to hospital for intussusception. Lancet. 2001;358 :1224 –1229[CrossRef][Web of Science][Medline]
- Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract. 2002;5 :143 –151[Medline]
- National Center for Health Statistics and National Immunization Program. The 2003 National Immunization Survey. Hyattsville, MD: Centers for Disease Control and Prevention; 2004. Available at: ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/nis/nispuf03dat.zip. Accessed March 23, 2005
- Farrington CP. Control without separate controls: evaluation of vaccine safety using case-only methods. Vaccine. 2004;22 :2064 –2070[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
