Rotavirus infection is the main cause of acute gastroenteritis in infancy but its pathogenic role is not limited to the intestine.1 Among the systemic manifestations of the disease, rotavirus infection has been specifically linked to childhood seizures (CS) by different mechanisms.2–5 Although there is an increasing body of evidence supporting the role of rotavirus infection as the cause or trigger of clinical neurologic illness,1–5 it is difficult to estimate the actual proportion of cases that are explained by this entity. Rotavirus is usually underestimated as a pathogen in children without diarrhea and it does not constitute part of the routine diagnostic work-up of children with seizures.
The existing rotavirus vaccination has shown consistent direct and indirect benefits in terms of effectiveness and impact against intestinal disease, namely acute gastroenteritis.6 The protective association between rotavirus vaccination and CS has been suggested recently, showing an average 20% reduction in the risk of seizure requiring hospitalization or emergency care in the year after a full course of rotavirus vaccination.7
Rotavirus vaccines were licensed in Spain between late 2006 and early 2007, reaching moderate coverage in the Galicia region (North West Spain) despite not being reimbursed.8 This moderate coverage—ranging from 29% to 41%—showed a high effectiveness and significant impact on acute gastroenteritis admissions in children in this region.8–10
We have hypothesized that rotavirus vaccination might have also impacted the incidence of hospitalizations for CS among children <5 years of age in the Spanish region of Galicia.
PATIENTS AND METHODS
This is a retrospective, observational and hospital-based surveillance study conducted in Galicia (North West Spain). Total population for the region as a whole (January 2006 census) is 2,767,524 people (approximately 100,000 under 5 years of age). This study was performed using a previously reported method.8–10 In brief, the study period, between 2003 and 2013, included pre- and post-rotavirus vaccination seasons. The number and rate of hospitalizations of children <5 years of age in the Galicia hospital network (SERGAS) due to any kind of CS were calculated for the study period. The primary objective of the study was to identify and assess any changes in the incidence of seizure hospitalizations and to analyze their relation to rotavirus vaccine coverage and rotavirus acute gastroenteritis (RAGE) hospitalization.
Data Collection and Study Procedures
The official surveillance system for hospital data, known as Conjunto Mínimo Básico de Datos de Hospitalización de Agudos (Minimum Data Set of Acute Hospitalizations) was used as the information source. This database includes personal, administrative and medical data for all patients admitted to any hospital in Galicia, with diagnoses codified according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). The Conjunto Mínimo Básico de Datos de Hospitalización de Agudos has maintained a coded valid discharge rate and since 2003, 100% of the hospitalizations are coded.8 A search strategy was applied based on disease category codes corresponding to convulsions (780.3* ICD-9-CM code), convulsions in newborn (779.0*), myoclonus (333.2*) and epilepsy (345*) diagnosis at discharge in any diagnostic position. These ICD-9-CM code combinations have been previously shown to have a high positive predictive value for seizures.7,11 The diagnosis, age and date of admission were collected for each case identified. The study protocol was reviewed and approved by the Clinical Research Ethics Committee of Galicia (Comité Ético de Investigación Clínica de Galicia).
This was a hospital-based surveillance study and therefore no sample size was calculated. The annual and monthly hospitalization rates for convulsions (780.3*), “any kind of CS” (780.3* + 779.0* + 333.2* + 345*) during the study period were calculated both for the whole study population (children <5 years of age) and by age groups (<1, 1–2, 2–3, 3–4, 4–5 years). Vaccination coverage was also compared with yearly admission rates due to the considered pathologies and using the Pearson’s r coefficient. Normality and homoscedasticity of the data were verified using the Lilliefors and Breusch–Pagan test, respectively. Admission rates due to the mentioned pathologies were compared with those of RAGE (008.61) using the Spearman’s ρ statistic. For this purpose, quarterly periods were considered to reduce the number of ties, because otherwise it could affect the performance of the statistic. Rotavirus vaccination coverage was estimated according to the number of doses of each vaccine distributed per month (Source: IMS Health) divided by 2 or 3 doses needed depending on the specific vaccine used (Rotateq®: three doses, Rotarix®: two doses), and by the registered number of children born for each age interval in that period (Source: Spanish National Statistics Institute): 29% (2007), 49% (2008), 49% (2009), 22% (2010), 28% (2011) and 41% (2012). Case-population coverage was calculated through the individual electronic vaccination record. The population data used as denominator for the calculation of rates was obtained from the Galician Institute of Statistics. Variations in the hospitalization rates for each individual diagnosis during the post-vaccination period (2008, 2009, 2010, 2011 and 2012) as compared with the rate of the pre-vaccination period (2003–2006) were also calculated with 95% confidence intervals. An additional confirmation analysis making this pre/post-vaccination comparison using the least favorable year of the pre-vaccination period (ie that with the lowest rate) was also performed. An alternative analysis was also performed comparing the general rotavirus vaccination coverage with the vaccination coverage in the case-population using the asymptotic normality of the sample coverage.
A P-value <0.05 was considered significant. The statistical analyses were performed using R version 3.0.2.
Our study cohort totaled 6149 children <5 years of age admitted to hospital between 2003 and 2013 with any kind of CS. The correlations between the yearly admission rates of the seizure-related studied pathologies and the rotavirus vaccine coverage are shown in Table 1. The correlations between the monthly admission rates due to considered pathologies and the RAGE admission rates, as well as with the vaccine coverage are shown in Table 2. The results are not shown for myoclonus, because the number of cases is too small, and we only have considered this pathology in the category “all kind of CS.”
The changes in incidence rates at different age intervals for those diagnostic codes that showed significant correlation with vaccination are summarized in Table, Supplemental Digital Content 1, http://links.lww.com/INF/C113.
The annual hospitalization rate for any kind of CS in children <5 years is negatively correlated with the vaccination coverage (r = −0.673; P = 0.033) (Table 1, Fig. 1) and positively correlated with the admission due to RAGE (ρ = 0.506; P = 0.001; Table 2). These decrease rates (Table, Supplemental Digital Content 1, http://links.lww.com/INF/C113) ranged from 16.2% [95% confidence interval: 8.3–23.5%] in 2007 to 34.0% (27.3–40.1%) in 2010, as compared with the median rate of the pre-vaccination period (2003–2006). In children 1–2 years of age, there was also a significant decrease in hospitalization rates ranging from 28.5% (16.2–39.0%) in 2008 to 48.4% (38.2–56.9%) in 2012, significantly correlated with rotavirus vaccination (r = −0.798; P = 0.006) and RAGE hospitalization rate (ρ = 0.366; P = 0.020). There was a positive and significant correlation for children 2–3 years of age (ρ = 0.421; P = 0.007) but without evidence of concomitant reduction in seizures admission rates, which could be explained by the lack of power for a low number of cases in this age group. In the remaining age groups, changes were either not statistically significant or not related to vaccination changes and/or RAGE admission.
The hospitalization rate for convulsions (780.3* ICD-9-CM code) is also correlated with rotavirus vaccination coverage (Table 1, Fig. 1) and RAGE hospitalization rate (Table 2) and a significant decrease was found (Table, Supplemental Digital Content 1, http://links.lww.com/INF/C113) in: (1) infants 1–2 years of age (coverage r = −0.813, P = 0.004; RAGE rate ρ = 0.395, P = 0.012) with decrease rates from 26.6% (12.9–38.1%) in 2007 to 51.8% (41.1–60.6%) in 2009; (2) children <5 years (coverage r = −0.747, P = 0.013; RAGE rate ρ = 0.543, P < 0.001) with decrease rates from 18.7 % (9.6–26.8%) in 2007 to 42.5% (35.3–48.9%) in 2012. Once again a significant and positive correlation for the children of 3 years of age was found (ρ = 0.249; P = 0.006), but no change was observed in seizure hospitalizations rate for this age group. The decrease rates obtained for the remaining age groups cannot be explained as a result of rotavirus vaccination and/or RAGE admission rates. Note that for these 2 pathologies there are other significant values in Table 2, however only those significant in both tables have been commented. For epilepsy (345* ICD-9-CM codes), there were no significant correlations (Tables 1 and 2) for any age group. For convulsions in newborns (779.0* ICD-9-CM codes), there was no significant correlation with RAGE hospitalization rate (ρ = −0.010; P = 0.950) and correlation with vaccination coverage was r = −0.175 (P-value was not computed because lack of homoscedasticity of the data). Therefore, variations obtained for both epilepsy and convulsions in newborns cannot be related to rotavirus vaccination.
To corroborate our findings an additional comparison was performed taking into account the lowest rate of the analyzed seizure code in pre-vaccination period instead of the average rate. These rates were selected according the age group and the corresponding pathology. Our positive findings for children <1 year of age and younger than 5 were maintained (see Table, Supplemental Digital Content 1, http://links.lww.com/INF/C113). Of note, as in the previous analysis, we have considered that the post-vaccination period started in the same year for all age groups, because there could be a herd effect from the vaccinated infants in non-vaccinated children. According to Table, Supplemental Digital Content 2, http://links.lww.com/INF/C114, there is a significant reduction in hospitalization rates only in children who actually received the vaccine.
Finally, an alternative analysis was performed comparing the vaccination coverage estimated in the whole population with the real coverage observed among “seizures” patients, obtained through the individual electronic vaccination record. (Table, Supplemental Digital Content 3, http://links.lww.com/INF/C115). The coverage in cases were significantly lower than global coverage for all the pathologies (“any kind of CS” and “convulsions”) for all the years except 2010—the year when rotavirus vaccines were temporarily withdrawn from the Spanish market.12
This study shows for the first time that rotavirus vaccination even with moderate coverage may have a significant impact in seizure-related hospitalizations in childhood. This additional benefit of rotavirus vaccination appears more marked in the youngest infants, namely during the first 2 years of age. This reduction may mean not only additional prevention of rotavirus disease burden but also further cost saving and avoidance of infant and family suffering.
Our results are aligned with the recent Centers for Disease Control and Prevention report on the benefits of rotavirus vaccines in terms of effectiveness against CS.7 Payne et al7 found that a full course of rotavirus vaccination significantly reduced the risk of CS at least during the year following the last rotavirus vaccination, with a 18–21% decrease in risk seizures requiring emergency care attendance or hospitalization, as compared with children not receiving the vaccine. This study confirms the findings of Payne et al using a different analytical method in a different country, focusing on the more severe forms of CS, that is, those that require hospitalization. Interestingly, in our series no displacement of cases to later ages is found, which might suggest a longer duration of this protective effect of vaccination and/or the cumulative effect of vaccinated cohorts and increasing coverage.
The rotavirus vaccine impact found is driven mainly through its effect over hospitalizations due to convulsions (ICD-9-CM code) but not to other seizure-related conditions. This agrees well with the rationale behind its beneficial effect: the vaccination prevents the stimulus (ie, rotavirus infection) in infants otherwise susceptible to either febrile seizures or neurological tropism of rotavirus. This impact on CS hospitalizations is more obvious in the first 2 years of age, which may simply reflect that protection is provided soon after vaccination and that both febrile and non-febrile seizures have their maximum incidence around that age. Our design cannot distinguish whether the contribution is mainly driven by the decrease of non-febrile or febrile seizures. Non-febrile seizures incidence usually peaks before 1 year of age and is more frequently linked to rotavirus etiology. However, we consider that the benefit exerted by rotavirus vaccines is mainly due to the prevention of febrile seizure-related hospitalizations. Febrile seizures are the most frequent cause of seizure hospitalization in childhood and their incidence peaks also in this age range—around 1.5 years of age—when the magnitude of the hospitalization decrease found is larger.13,14
Rotavirus can act as a seizures trigger in susceptible children or get directly involved in the pathogenesis.1–5,15 Antigenemia is frequently found in children with rotavirus gastroenteritis, showing that the infection is not locally restricted to the intestine and that systemic illness can be produced.16 The specific relation between rotavirus infection and seizures has been described in the literature.1–5,15 Rotavirus-associated seizures are more frequently reported in clusters and in afebrile children.15,17 Whatever the case, rotavirus assessment is not part of the routine work-up of infants with febrile or afebrile convulsions, and thus, its actual burden is probably underestimated.18 Several mechanisms have been suggested to explain the relation between rotavirus infection and seizures, from direct neurotoxicity to specific neurotransmitter interferences at calcium channel levels.19,20
Our findings need to be carefully interpreted. Other uncontrolled factors (prematurity, any other infectious trigger…), the variability in management and admission policy in children with seizures or changes in the coding practices, may have also influenced our findings. The strengths and weakness of the national administrative database for hospital admissions used as data source in our study have been previously discussed.8 Even though the ICD-9-CM code combination applied in our study has been previously shown to have a high positive predictive value for seizures,11 it is also true that the positive predictive value was lower for inpatient admissions—the population in our study—than for emergency department visits. The sustained global reduction in CS hospitalization of “any cause” and specifically of “convulsions” observed is strongly correlated with the use of rotavirus vaccines and rotavirus gastroenteritis but this is not the case for other kind of childhood convulsions classified under the myoclonus, epilepsy or neonatal seizures codes. Further differentiation in other more specific clinical subgroups of seizures was not possible with the current design. In addition, although estimation of vaccination coverage using sales is a usual practice, the percentage obtained may be overestimated, as the distributed doses may not necessarily reflect the actual number of doses that have been administered. However, the strong correlation of our findings with rotavirus vaccine coverage and patient age reinforces our interpretation of the data. The impact increases with vaccine coverage and with the accumulation of vaccinated children in the cohort, suggesting also that this protective effect of rotavirus vaccines persist over time: there seems to be no displacement of cases to later ages. Also, the vaccine impact is driven mainly through its effect in infant convulsions (nonfebrile and febrile seizures) but not in other seizure-related conditions, which agrees well with the rationale behind its beneficial effect: the vaccination prevents the stimulus (ie, rotavirus infection) in otherwise susceptible infants. For all these reasons, continued surveillance and further population-based studies are needed to corroborate and follow-up our findings.
Rotavirus vaccine may reduce the risk of hospitalization due to CS, mainly in infants. These benefits can arise with just moderate rotavirus vaccination coverage. Additional studies are required, but rotavirus vaccination benefits seem to be broader than just prevention of severe forms of acute gastroenteritis.
1. Candy DC.. Rotavirus infection: a systemic illness? PLoS Med. 2007;4:e117
2. Yeom JS, Kim YS, Park JS, et al. Role of Ca2+
homeostasis disruption in rotavirus-associated seizures
. J Child Neurol. 2014;29:331–335
3. Lloyd MB, Lloyd JC, Gesteland PH, et al. Rotavirus gastroenteritis and seizures
in young children. Pediatr Neurol. 2010;42:404–408
4. DiFazio MP, Braun L, Freedman S, et al. Rotavirus-induced seizures
in childhood. J Child Neurol. 2007;22:1367–1370
5. Pang XL, Joensuu J, Vesikari T.. Detection of rotavirus RNA in cerebrospinal fluid in a case of rotavirus gastroenteritis with febrile seizures
. Pediatr Infect Dis J. 1996;15:543–545
6. Cortese MM, Immergluck LC, Held M, et al. Effectiveness of monovalent and pentavalent rotavirus vaccine. Pediatrics. 2013;132:e25–e33
7. Payne DC, Baggs J, Zerr DM, et al. Protective association between rotavirus vaccination and childhood seizures
in the year following vaccination in US children. Clin Infect Dis. 2014;58:173–177
8. Martinón-Torres F, Martinón-Torres N, Bouzón Alejandro M, et al. Acute gastroenteritis hospitalizations among children aged < 5 years before and after introduction of rotavirus vaccines
: a hospital-based surveillance study in Galicia, Spain. Hum Vaccin Immunother. 2012;8:946–952
9. Martinón-Torres F, Aramburo A, Martinón-Torres N, et al. A reverse evidence of rotavirus vaccines
impact. Hum Vaccin Immunother. 2013;9:1289–1291
10. Martinón-Torres F, Bouzón Alejandro M, Redondo Collazo L, et al.ROTACOST Research Team. Effectiveness of rotavirus vaccination in Spain. Hum Vaccin. 2011;7:757–761
11. Shui IM, Shi P, Dutta-Linn MM, et al.Vaccine Safety Datalink Research Team. Predictive value of seizure ICD-9 codes for vaccine safety research. Vaccine. 2009;27:5307–5312
12. Bouzón Alejandro M, Diez Domingo J, Martinón-Torres F.. Circovirus and impact of temporary withdrawal of rotavirus vaccines
in Spain. Hum Vaccin. 2011;7:798–799
13. Verity CM, Golding J.. Risk of epilepsy after febrile convulsions: a national cohort study. BMJ. 1991;303:1373–1376
14. Reid AY, Galic MA, Teskey GC, et al. Febrile seizures
: current views and investigations. Can J Neurol Sci. 2009;36:679–686
15. Dickey M, Jamison L, Michaud L, et al. Rotavirus meningoencephalitis in a previously healthy child and a review of the literature. Pediatr Infect Dis J. 2009;28:318–321
16. Blutt SE, Matson DO, Crawford SE, et al. Rotavirus antigenemia in children is associated with viremia. PLoS Med. 2007;4:e121
17. Martin ET, Kerin T, Christakis DA, et al. Redefining outcome of first seizures
by acute illness. Pediatrics. 2010;126:e1477–e1484
18. Steering Committee on Quality Improvement and Management, Subcommittee on Febrile Seizures
American Academy of Pediatrics. . Febrile seizures
: clinical practice guideline for the long-term management of the child with simple febrile seizures
. Pediatrics. 2008;121:1281–1286
19. Kawashima H, Inage Y, Ogihara M, et al. Serum and cerebrospinal fluid nitrite/nitrate levels in patients with rotavirus gastroenteritis induced convulsion. Life Sci. 2004;74:1397–1405
20. Tian P, Ball JM, Zeng CQ, et al. The rotavirus nonstructural glycoprotein NSP4 possesses membrane destabilization activity. J Virol. 1996;70:6973–6981