Vaccine 33 (2015) 1017–1020

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Intussusception risk after RotaTeq vaccination: Evaluation from worldwide spontaneous reporting data using a self-controlled case series approach Sylvie Escolano a,b,∗ , Catherine Hill c , Pascale Tubert-Bitter a,b a Inserm, Centre for Research in Epidemiology and Population Health (CESP), U1018, Biostatistics Team, 16 avenue Paul Vaillant Couturier, F-94807 Villejuif, France b University Paris-Sud, UMRS 1018, 16 avenue Paul Vaillant Couturier, F-94807 Villejuif, France c Service de Biostatistique et Epidémiologie, Institut Gustave Roussy, F-94805 Villejuif, France

a r t i c l e

i n f o

Article history: Received 13 November 2014 Received in revised form 12 December 2014 Accepted 5 January 2015 Available online 14 January 2015 Keywords: Rotavirus vaccine Rotateq Self-controlled case series method Spontaneous reports Intussusceptions

a b s t r a c t The increased risk of intussusception after vaccination with the rhesus-human reassortant rotavirus vaccine Rotashield led to its withdrawal in 2005. We assess the risk of intussusception following the pentavalent rotavirus vaccine (RV5) on the basis of worldwide reports to the manufacturer up to May 2014, using a self-controlled case series. The method had to be modified to account for the under-reporting, a specific feature of pharmacovigilance spontaneous reports. The risk of intussusception occurring in either of the 0- to 2-day, 3- to 7-day or 8- to 14-day risk periods, was compared to the risk in the 15to 30-day period. A total of 502 cases occurring 0–30 days after a vaccine dose were studied, including 188 cases after the first dose, 190 cases after the second dose, and 124 cases after the third dose. The incidence risk ratio relative to the control period was highest for the 3- to 7-day period and equal to 3.45 (95% CI 1.84–6.55), 1.63 (0.86–3.13) and 1.73 (0.86–3.51) after the first, second and third dose, respectively. Rotavirus vaccination with RV5 increases the risk of intussusception 3–7 days following vaccination, mainly after the first dose and marginally after the second and third doses. The risk is small and restricted to a short time window. It does not outweigh the benefit of the vaccination, but parents of vaccinated infants should be informed in order to react appropriately to the first symptoms. With appropriate assumptions about the reporting rate, spontaneous reports of adverse events after vaccination can be studied to evaluate vaccine safety. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Rotavirus infection is a major cause of infant diarrhea and therefore a frequent cause of hospitalization and outpatient visits worldwide, resulting in more than 450,000 deaths each year, mainly in low-income countries [1,2]. In 1999, the first oral rotavirus vaccine, a rhesus-human reassortant rotavirus vaccine (RRV, Rotashield, Wyeth), was withdrawn from the market because of an attributable risk of intussusception between 1 extra case per 10,000 vaccinated infants and 1 extra case per 5000 vaccinated

∗ Corresponding author at: Centre for Research in Epidemiology and Population Health (CESP), U1018, Biostatistics Team, F-94807 Villejuif, France. Tel.: +33 1 45 59 50 84. E-mail addresses: [email protected] (S. Escolano), [email protected] (C. Hill), [email protected] (P. Tubert-Bitter). http://dx.doi.org/10.1016/j.vaccine.2015.01.005 0264-410X/© 2015 Elsevier Ltd. All rights reserved.

infants [3]. Intussusception, the invagination of one segment of the intestine into another, is a medical emergency and can be fatal if untreated. The risk of spontaneous intussusception varies with age, the highest risk being observed around the age of 6 months [4]. Since 2006, two new rotavirus vaccines, a monovalent RV1 (Rotarix, GlaxoSmithKline Biologicals) and a pentavalent RV5 (RotaTeq, Sanofi-Pasteur MSD), have been authorized. Studies based on cases from Mexico [5], the United States [6,7], and worldwide cases [8] have shown that RV1 increases the risk of intussusception by one extra case per 50,000 vaccinated infants, mainly within 7 days after the first dose. Three studies conducted in the United States have recently assessed the risk of intussusception after the first dose of RV5 vaccination based on 4 and 5 cases of intussusception within 7 days after vaccination [7,9] and on 60 cases within 6 days after vaccination [6], respectively. A summary of intussusception risk estimates can be found in Table S6 of Yih’s paper [9]. Using a variant of the self-controlled case series (SCCS) method accounting for the underreporting of cases recorded

1018

S. Escolano et al. / Vaccine 33 (2015) 1017–1020

in pharmacovigilance databases [10], we present the analysis of spontaneous reports worldwide of intussusception after vaccination with RV5 on the basis of 261 cases occurring between 0 and 7 days after vaccination. 2. Methods 2.1. Spontaneous reports The line listing of the 972 reports worldwide of intussusception cases following RV5 vaccination and reported until May 2014 to the manufacturer was made available to us by the French authorities. RV5 is a three-dose vaccine to be delivered at ages 2, 4, and 6 months. For each case the dataset included gender, country of origin, date of vaccination, date of intussusception, time between most recent dose and intussusception, dose number, and either date of birth or age at time of vaccination or age at adverse event, with a substantial amount of missing data. After this phase, 22% of the cases were excluded owing to incomplete or incoherent data (for instance, when the age at first dose, calculated from age at vaccination and adverse event date was under 20 or over 180 days of age). In a second step, cases occurring more than 30 days after the most recent vaccination dose were excluded because of the 4-week minimum interval between the doses, so a case could be associated with a single dose. Finally, the working dataset contained 502 cases, 188 (37.5%) occurring after the first dose, 190 (37.9%) after the second dose, and 124 (24.7%) after the third dose. Fig. 1 details the data processing. 2.2. Statistical methods We used a validated extension of the SCCS method to analyze spontaneous reports of adverse events occurring after vaccination. This approach has previously been used to estimate the risk of intussusception after RV1 vaccination and details of the method have been described elsewhere [8,10]. Briefly, the observation period is defined as the 31 days following each vaccination dose, day 0 being the vaccination day. This period is partitioned into four risk periods: 0–2, 3–7, 8–14 and 15–30 days. This partition is equivalent for the three first periods to the choice made by Murphy et al. [3] for assessing intussusception risk after RRV vaccination. The last period is considered as a reference period, where the risk is taken to be close to baseline. For each of the first three periods, the ratio of the incidences in the period and in the reference period was estimated. With the SCCS methodology, only cases are used, the

972 Cases in the original line lisng 216 Cases ineligible because of incomplete or incoherent informaon on dates, age or dose number 756 Cases eligible 254 Cases excluded because intussuscepon did not occur within 30 days aer last vaccinaon dose 502 Cases analyzed Fig. 1. Flow chart of the intussusception cases. Worldwide spontaneous reports of intussusception to RV5 manufacturer until May 2014.

subjects acting as their own controls. Consequently, all confounding factors that are invariant with time are implicitly adjusted for. The reported events are assumed to occur according to a Poisson process with an incidence rate that is the product of the reporting probability and of the intussusception incidence rate. According to the SCCS model, the logarithm of the intussusception incidence rate is a linear function of the period and of the dose. For standard SCCS analysis to be used, case collection must be independent from vaccination, a hypothesis which is clearly violated in spontaneous reporting data. The extended SCCS used handles this reporting issue. Several parametric functions modeled the reporting probability by assuming an exponential decrease with time since vaccination. Three variants were tested, where the decrease could be either constant (1 parameter), or dosedependent (3 parameters) or age-dependent (3 parameters, age in three groups: 0–2 months, 3 months, 4+ months). These three models were compared using likelihood ratio tests. To obtain results comparable with those of a recent analysis of data from the vaccine adverse event reporting system [6], we used an alternative partitioning of the risk periods (0–2, 3–6, 7–14, and 15–30 days), taking 0–2 days as the reference period. All analyses were conducted with the R software, version 3.0.2 (www.r-project.org). 3. Results 3.1. Study population Table 1 describes the main characteristics of the 502 cases analyzed. The majority of cases (56.4%) were male, 71.5% of the cases coming from the United States and the ages at each dose being roughly in accordance with the recommended schedule. Fig. 2 describes the distribution of the reported event times for each vaccination dose. 3.2. Intussusception risk Likelihood ratio tests showed that the models with a dose- or age-dependent decrease in the reporting probability did not fit significantly better than the model with a constant decrease. The results of the retained model are given in Table 2. After the first dose, the intussusception risk was 3.45-fold higher ([95% CI, 1.84–6.55]) in the 3–7 days after vaccination as compared to the 15- to 30-day period. After the second and third doses, there were no significant Table 1 Characteristics of children having a reported intussusception after RV5 vaccination. Worldwide spontaneous reports to RV5 manufacturer until May 2014. Characteristic

Studied sample (n = 502)

Male sex – no. (%) Age at intussusception, median [range], weeks Country – no. (%) United States Germany Israel Australia Japan Otherb Age at vaccination median [range], weeks Dose 1 Dose 2 Dose 3

283 (56.4)a 17 [4–78] 359 (71.5) 31 (6.2) 25 (5.0) 23 (4.6) 19 (3.8) 45 (9.0) 9 [3–25] 17 [8–34] 26 [13–78]

RV5, pentavalent rotavirus vaccine. a Gender information was missing in 20 children. b Finland (7 cases), France (6 cases), Belgium and Malaysia (each, 4 cases), Greece and India (each, 3 cases), Austria, Canada, Nicaragua and Republic of Korea (each, 2 cases), Argentina, Czech Republic, Hong Kong, Hungary, Italy, Lebanon, Philippines, Portugal, Spain and Thailand (each, 1 case).

S. Escolano et al. / Vaccine 33 (2015) 1017–1020

1019

dose 1 was significant, with a value of 3.59 [95% CI, 2.20–6.14] which is comparable to Haber et al.’s 3.75 [95% CI, 1.90–7.39] estimation (Supplementary Table 2).

4. Discussion

Fig. 2. Distribution of time between vaccination with RV5 and intussusception by dose.

increases in any period, but the risk was about 1.5-fold higher during the first 7 days as compared to the 15- to 30-day period. The exponential decay parameter was −0.034 (95% CI, −0.07 to −0.003) per day, corresponding to a 21.2% reduction in reporting probability 7 days later. The results of the models with the three reporting probability functions can be found in Supplementary Table 1. When studying the risk period 3–6 days instead of 3–7 days and using period 0–2 days as a reference for the purpose of comparison with Haber’s work [6], the incidence risk ratio in the 3 to 6 days after

Table 2 Estimates of intussusception risk parameters with an exponential decay reporting probability model. Worldwide spontaneous reports to RV5 manufacturer until May 2014. RV5 dose

Risk window, days

No. of cases in risk window

Relative incidence (95% CI)

Dose 1

0–2 3–7 8–14 15–30

19 93 28 48

1.02 (0.44–2.38) 3.45 (1.84–6.55) 0.91 (0.51–1.62) 1a

Dose 2

0–2 3–7 8–14 15–30

34 55 41 60

1.47 (0.68–3.21) 1.63 (0.86–3.13) 1.07 (0.63–1.80) 1a

Dose 3

0–2 3–7 8–14 15–30

24 36 27 37

1.68 (0.73–3.88) 1.73 (0.86–3.51) 1.14 (0.62–2.08) 1a

RV5, pentavalent rotavirus vaccine; CI, confidence interval. a Reference period.

We observed a significant increase in the risk of intussusception 3–7 days after vaccination dose 1, with a relative risk of 3.45 as compared to the risk 15–30 days after vaccination. After doses 2 and 3, the risk increase was small (about 1.5 during both the 0- to 2- and 3- to 7-day periods) and did not reach significance. Our estimation following dose 1 in the 3- to 7-day window based on 188 cases is of the same order of magnitude as that in Haber’s study based on 60 spontaneous report cases [6], and is much lower than in two others cohort studies [9,11] based on 5 and 7 cases respectively. All these studies used self-controlled methods. The other studies were based on data from large cohorts [7,12,13]. For such a rare event, even smaller numbers of cases (4, 1 and 3, respectively) were observed and led to less reliable estimates, if any. Our study was based on a large number of cases reported worldwide, allowing evaluation of the risk of intussusception in narrow time windows and with reasonable power. The SCCS methodology is now routinely used to evaluate vaccine safety. For instance, following H1N1 vaccination, it has recently been used to evaluate the risks of Guillain–Barré syndrome [14], narcolepsy [15], epileptic seizures [16], convulsions [17], and multiple sclerosis [18]. Its efficiency compared to the cohort design has been established, which is a desirable feature when studying rare events such as intussusception [19]. The extension of the SCCS to spontaneous reports has also demonstrated its good statistical properties (control of type I error rate and power) [10]. However, spontaneous reports to vaccine-manufacturing companies or to the authorities are not always well documented: vaccination and intussusception dates are sometimes imprecise, the number of doses before intussusception is not always specified and patients’ ages are recorded in days, weeks, months, or even in years. All these approximations reduce the number of analyzable cases. More generally, the validity of information (exposure, outcome, dates) could be questioned and cannot be assessed. As it is likely to not vary markedly with time, at least for the scale time of the observation period, it should not affect our findings. Our results depend on the strong assumption of an exponential decrease in the reporting probability with time since vaccination. We have proved, in a precedent paper presenting a simulation study, that results are not much dependent on the choice of the functional form of the reporting rate [10]. This assumption seems plausible here, in view of Haber et al. [6] work, which shows (in their Fig. 1) a more or less exponential decrease in non-intussusception reports after the first dose of RV5. Furthermore, a capture-recapture analysis of the VAERS database and epidemiological studies produced a piecewise estimation of the reporting probability in VAERS of 55% in the 3- to 7-day period, 47% in the 8- to 14-day period and 28% in the 15+ day period [20]. These values give a ratio for the reporting probabilities between the two intervals 8–14 and 3–7 days of 47%/55% i.e. 0.85 and we obtained approximately the same value when calculating the ratio of our own estimates at 11 and 5 days, the mid-points of each interval. In summary, while not providing definitive evidence, our study suggests that vaccination with RV5 is associated with an increase in the risk of intussusception after the first dose. The risk is small and does not call into question the well-documented benefits of rotavirus vaccination [11,21], but parents of vaccinated infants should be informed so that they can react appropriately to the first symptoms.

1020

S. Escolano et al. / Vaccine 33 (2015) 1017–1020

Acknowledgments We thank Dr. Alexis Jacquet of the ANSM (Agence Nationale de Sécurité du Médicament et des Produits de Santé, Saint Denis, France) and the Pharmacovigilance Department of Sanofi Pasteur MSD for their valuable help. Conflicts of interest statement: We declare that we have no conflicts of interest. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.vaccine.2015.01.005. References [1] Rotavirus surveillance – worldwide, 2001–2008. MMWR Morb Mortal Wkly Rep 2008;57:1255–7. [2] Tate JE, Burton AH, Boschi-Pinto C, Steele AD, Duque J, Parashar UD, et al. 2008 estimate of worldwide rotavirus-associated mortality in children younger than 5 years before the introduction of universal rotavirus vaccination programmes: a systematic review and meta-analysis. Lancet Infect Dis 2012;12:136–41, http://dx.doi.org/10.1016/S1473-3099(11)70253-5. [3] Murphy TV, Gargiullo PM, Massoudi MS, Nelson DB, Jumaan AO, Okoro CA, et al. Intussusception among infants given an oral rotavirus vaccine. N Engl J Med 2001;344:564–72, http://dx.doi.org/10.1056/NEJM200102223440804. [4] Tate JE, Simonsen L, Viboud C, Steiner C, Patel MM, Curns AT, et al. Trends in intussusception hospitalizations among US infants, 1993–2004: implications for monitoring the safety of the new rotavirus vaccination program. Pediatrics 2008;121:e1125–32, . [5] Patel MM, Lopez-Collada VR, Bulhoes MM, De Oliveira LH, Bautista Marquez A, Flannery B, et al. Intussusception risk and health benefits of rotavirus vaccination in Mexico and Brazil. N Engl J Med 2011;364:2283–92, http://dx.doi.org/10.1056/NEJMoa1012952. [6] Haber P, Patel M, Pan Y, Baggs J, Haber M, Museru O, et al. Intussusception after rotavirus vaccines reported to US VAERS, 2006–2012. Pediatrics 2013;131:1042–9, http://dx.doi.org/10.1542/peds.2012-2554. [7] Weintraub ES, Baggs J, Duffy J, Vellozzi C, Belongia EA, Irving S, et al. Risk of intussusception after monovalent rotavirus vaccination. N Engl J Med 2014;370:513–9, http://dx.doi.org/10.1056/NEJMoa1311738. [8] Escolano S, Farrington CP, Hill C, Tubert-Bitter P. Intussusception after rotavirus vaccination – spontaneous reports. N Engl J Med 2011;365:2139, http://dx.doi.org/10.1056/NEJMc1107771.

[9] Yih WK, Lieu TA, Kulldorff M, Martin D, McMahill-Walraven CN, Platt R, et al. Intussusception risk after rotavirus vaccination in U.S. infants. N Engl J Med 2014;370:503–12, http://dx.doi.org/10.1056/NEJMoa1303164. [10] Escolano S, Hill C, Tubert-Bitter P. A new self-controlled case series method for analyzing spontaneous reports of adverse events after vaccination. Am J Epidemiol 2013, http://dx.doi.org/10.1093/aje/kwt128. [11] Carlin JB, Macartney KK, Lee KJ, Quinn HE, Buttery J, Lopert R, et al. Intussusception risk and disease prevention associated with rotavirus vaccines in Australia’s National Immunization Program. Clin Infect Dis 2013;57:1427–34, http://dx.doi.org/10.1093/cid/cit520. [12] Shui IM, Baggs J, Patel M, Parashar UD, Rett M, Belongia EA, et al. Risk of intussusception following administration of a pentavalent rotavirus vaccine in US infants. J Am Med Assoc 2012;307:598–604, http://dx.doi.org/10.1001/jama.2012.97. [13] Buttery JP, Danchin MH, Lee KJ, Carlin JB, McIntyre PB, Elliott EJ, et al. Intussusception following rotavirus vaccine administration: post-marketing surveillance in the National Immunization Program in Australia. Vaccine 2011;29:3061–6, http://dx.doi.org/10.1016/j.vaccine.2011.01.088. [14] Romio S, Weibel D, Dieleman JP, Olberg HK, de Vries CS, Sammon C, et al. Guillain–Barré syndrome and adjuvanted pandemic influenza A (H1N1) 2009 vaccines: a multinational self-controlled case series in Europe. PLOS ONE 2014;9:e82222, http://dx.doi.org/10.1371/journal.pone.0082222. [15] Miller E, Andrews N, Stellitano L, Stowe J, Winstone AM, Shneerson J, et al. Risk of narcolepsy in children and young people receiving AS03 adjuvanted pandemic A/H1N1 2009 influenza vaccine: retrospective analysis. Br Med J 2013;346:f794. [16] Arnheim-Dahlström L, Hällgren J, Weibull CE, Sparén P. Risk of presentation to hospital with epileptic seizures after vaccination with monovalent AS03 adjuvanted pandemic A/H1N1 2009 influenza vaccine (Pandemrix): self controlled case series study. Br Med J 2012;345:e7594. [17] Stowe J, Andrews N, Bryan P, Seabroke S, Miller E. Risk of convulsions in children after monovalent H1N1 (2009) and trivalent influenza vaccines: a database study. Vaccine 2011;29:9467–72, http://dx.doi.org/10.1016/j.vaccine.2011.10.029. [18] Farez MF, Ysrraelit MC, Fiol M, Correale J. H1N1 vaccination does not increase risk of relapse in multiple sclerosis: a self-controlled case-series study. Mult Scler 2012;18:254–6, http://dx.doi.org/10.1177/1352458511417253. [19] Farrington CP, Nash J, Miller E. Case series analysis of adverse reactions to vaccines: a comparative evaluation. Am J Epidemiol 1996;143: 1165–73. [20] Verstraeten T, Baughman AL, Cadwell B, Zanardi L, Haber P, Chen RT, et al. Enhancing vaccine safety surveillance: a capture-recapture analysis of intussusception after rotavirus vaccination. Am J Epidemiol 2001;154: 1006–12. [21] Glass RI, Parashar UD. Rotavirus vaccines – balancing intussusception risks and health benefits. N Engl J Med 2014;370:568–70, http://dx.doi.org/10.1056/NEJMe1315836.

Intussusception risk after RotaTeq vaccination: evaluation from worldwide spontaneous reporting data using a self-controlled case series approach.

The increased risk of intussusception after vaccination with the rhesus-human reassortant rotavirus vaccine Rotashield led to its withdrawal in 2005. ...
452KB Sizes 0 Downloads 4 Views