The reported annual incidence of RVGE managed at the primary care level ranges from 15.4 to 19.5 cases per 1000 children under 5 years and 20 cases per 1000 children under 3 years (Table 1). This variability could be explained by differences between the studies in methodology and the period under study. It is believed that active efforts to detect cases are generally less likely to be affected by biases. The reported proportion of RVGE cases relative to the total cases of AGE also varies between studies.

These differences may be explained by: (a) health care factors (patients seeking care for illness of varying severity), (b) the duration of follow-up (analysis of the entire year versus epidemic seasons), (c) the methods used for diagnosis, each with a different sensibility and sensitivity, with the gold standard being polymerase chain reaction (PCR), whose use is currently restricted to the context of research, and (d) the natural seasonal fluctuations in the circulation of RV.

A European study conducted before the introduction of the rotavirus vaccines found an incidence of RVGE managed at the primary care level of 1.45–4.18 per 100 children under 5 years, and the proportion of total cases of AGE with positive detection of RV by enzyme-linked immunoassay (ELISA) ranged from 25.5% to 41.3%, depending on the country.7

Since these data are based only on the cases of patients that sought medical care, the documented incidence must be lower than the actual incidence, as it is estimated that a significant proportion of affected children do not seek care from a paediatrician.8

In Spain, the annual incidence of hospitalization due to RV in children ranges between 120 and 480 cases per 100000 children under 5 years, and the proportion of cases of AGE attributable to RV ranges between 11% and 55% (Table 2). These differences are mainly explained by differences in methodology. Most studies are retrospective with collection of data from discharge summaries, the Conjunto Mínimo Básico de Datos (CMBD, “minimum basic data set”) and in some instances additional information from the Sistema de Información Microbiológica (SIM, Microbiological Information System). Furthermore, the quality of the documentation and the coding schemes used in the CMBD of inpatients in different hospitals could also be a source of discrepancies in these results.9,11,16 Other factors possibly at play are the patient selection criteria, differences in the protocols for hospital admission and the performance of virologic tests with different diagnostic yields.14

In Europe, the rate of hospitalization and the proportion of AGE cases due to RV out of all cases of AGE are higher compared to the figures reported for Spain (300–600 cases/100000 children under 5 years and up to 69% of hospital admissions due to AGE) (https://ecdc.europa.eu/en/publications-data/expert-opinion-rotavirus-vaccination-infancy). Differences between health care systems and surveillance systems preclude the direct comparison of the data, but their presence and the fact that the impact of vaccination against RV in Spain has been similar on AGE of all causes and on RVGE suggests that the identification of hospital admissions attributable to RV by means of the CMBD underestimates the burden of hospitalization.1

Two studies have assessed the incidence of nosocomial infection by RV obtaining data from the CMBD database (Table 2), with the limitations inherent in the use of this source to analyze nosocomial infection. The incidence ranged from 0.25 to 0.45 cases per 100 hospitalised children under 5 years. On the other hand, another study with a prospective design found an incidence that was 6–11 times higher.17 In the latter, the authors performed an active search for cases extending to up to 72h after discharge, which allowed a more specific identification of cases and a more accurate estimation of the frequency of nosocomial infection.

A systematic review conducted in Europe found an estimated median incidence of 6.2 cases per 100 hospitalizations (range, 0.3–27.7).18 The European Centre for Disease Prevention and Control (ECDC) has estimated that 25% to 30% of infections by RV in hospitalised children are acquired in hospital, and that these infections are the leading cause of nosocomial diarrhoea in the paediatric population (https://ecdc.europa.eu/en/publications-data/expert-opinion-rotavirus-vaccination-infancy). This adds to the burden of the disease, as it prolongs hospitalizations due to other causes.

In agreement with what has been described in other countries, the clinical presentation of RVGE in Spain is more severe and associated with a higher incidence of complications compared to diarrhoeas of other causes (Table 3), thus involving a greater use of health care resources, higher costs and a greater psychosocial impact.6

In 2006, on the occasion of the introduction of the rotavirus vaccines, the EuroRotaNet surveillance network was established in Europe to track the circulation of the different RV genotypes in participating countries (http://www.eurorota.net). Spain has participated in this network, contributing approximately 700 samples each year.

Overall, genotypes G1P[8], G2P[4], G3P[8], G4P[8], G9P[8] and G12P[8] have circulated with a frequency of 1% or greater. In Spain, these 6 genotypes account for 98% of samples in which only 1 genotype is detected. There are seasonal fluctuations in the genotype distribution (Fig. 1). This variability has been described in similar terms in Spanish studies.13,21–24

Figure 1.

Distribution of rotavirus genotypes circulating in Spain, 2006–2016. Adapted from EuroRotaNet: Annual Report 2016 (available at: http://www.eurorota.net/).

(0.11MB).

Based on the reports of EuroRotaNet, there is no evidence to date that rotavirus vaccination programmes are causing an emergence of new genotypes that are not covered by the vaccines. Changes in genotype distribution must be interpreted in the context of the natural fluctuations that are also observed in countries without vaccination programmes and that also existed prior to the introduction of the vaccines.

Two studies have evaluated the economic impact of RVGE on families (Table 4). Their results are not comparable, as they were conducted in different periods in relation to the introduction of vaccination.

A systematic review of the economic and psychosocial impact of infection by RV in Europe that includes Spanish studies estimated the cost of RVGE episodes treated on an inpatient basis, in the emergency department or in primary care at 1000–1700 euro, 409 euro and 166 euro, respectively, resulting in a total cost for the health care system of 28 million euro a year. The review found that 68% of parents of children hospitalised due to RV missed work, for a mean of approximately 4 days. The estimated annual cost of RVGE to society would rise to approximately 50 million euro.27

Infection by RV also has a considerable psychosocial impact, causing stress on the parents, whose quality of life is affected27 (Table 4). When it comes to prevalent diseases, variables related to quality of life are considered very relevant outcome measures, especially in relation to the interference of paediatric diseases in the everyday activity of households.28 This emotional and social impact may be an intangible cost and pose challenges to its standardization and quantification, but it should nevertheless be considered in the evaluation of health related quality of life in the patients and caregivers affected by the disease.28

There is evidence of an economic and emotional impact in the parents of children with RVGE, either hospitalised38 or managed at the primary care level, and this impact is greater compared to the impact found in parents of children with AGE of other aetiologies.6,29

The reviewed studies found that RV vaccines were very effective in preventing episodes of RVGE and related hospitalizations, and the observed effectiveness was consistent with the efficacy observed in clinical trials32 (Table 5).

The effectiveness in preventing hospitalizations due to RV ranged between 83% and 96%, with disparities probably stemming from differences in methodology, the post-vaccination periods under analysis, the protocols applied to hospital admission in cases of AGE and the different tests used for determining the aetiology, which differ significantly in their sensitivity.

Studies in Europe have found a similar effectiveness of vaccination, with reductions in the cases of RVGE managed in primary care or at the inpatient level of 68% to 98% depending on the control group used for comparison.35

We assessed the impact of vaccination against RV by estimating the reduction in the rate of hospitalization using data obtained mainly from hospital databases or electronic health records (Table 6). When comparing data from different sources, it is important to take into account differences in study design, geographical area and vaccination coverage (including how coverage has been estimated), as well as potential differences between hospitals in clinical practices.

Thus, in their study, Orrico et al. estimated vaccine coverage based on the regional vaccination register of the Valencian Community, unlike other studies in which it has been estimated based on the number of doses of vaccine distributed in the corresponding province, which could be a source of bias. Orrico et al. also fitted a model to control for factors that could modify outcomes (for instance, changes in admission protocols) and calculated the decrease in the rate of hospitalization taking into account the vaccination coverage, establishing the correlation between these two variables and finding that with vaccination coverages of less 20%, the decrease in the risk of hospitalization due to RV was of 37% and 45% in children aged 1 and 2 years, whereas coverage rates of 40% or greater achieved decreases in hospitalization of nearly 70 in the same age group.1 This suggests that intermediate vaccination rates achieve a herd immunity effect. The authors estimated a reduction in the costs associated to hospitalization due to AGE of 6 million euro per 100000 children under 5 years over a period of 7 years.1

Since infection by RV is not a notifiable disease in Spain and its diagnosis does not lead to changes in the treatment of AGE, hospitals may not routinely perform diagnostic tests for its detection. Furthermore, the introduction of the vaccines may have led to an increased use of tests for detection of RV, which in turn could have increased the proportion of cases of RVGE that are actually detected, which would have a negative impact when the effectiveness of vaccination in reducing hospitalizations is assessed based on case counts. Thus, it is important to assess the impact of vaccination not only in terms of hospital admissions due to RVGE, but also taking into account admissions due to AGE of all causes. A study by Redondo et al.38 found that in an estimated 66% of hospital admissions due to AGE, the diagnosis was coded as unspecified AGE. Evidence of age and seasonal distributions of aetiology suggests that a significant proportion of these cases could be cases of RVGE. The studies summarized in Table 6 found a significant reduction not only in cases of RVGE but also in the frequency of hospitalization due to AGE of all causes.

The results of a model to assess the economic impact of universal vaccination against RV with the pentavalent vaccine suggested based on data from a birth cohort suggest that the costs associated with RVGE would decrease by 76%, similar to the decrease observed in other European countries.8 This reduction would amount to annual savings of 22 million euro for the health care system and 38 million euro for society overall. We ought to emphasize that since these are estimates obtained from a model, they must be interpreted with caution.

Despite the significant estimated economic burden of RVGE and the impact of vaccination, cost-effective analyses conducted in Spain39,40 have not found vaccination to be efficient, unlike in other countries.41 Some authors attribute this discrepancy to the cohort study designs used to develop these models in Spain, which lead to underestimation of the indirect impact of vaccination, since they do not take into account herd immunity or undiagnosed cases of RV infection. In Australia, a cost-effectiveness analysis of vaccination that took into account data on herd immunity and the reduction in hospital admissions for AGE of undetermined aetiology observed after the introduction of routine vaccination found a substantially greater reduction in cost compared to the cost estimated in models of disease prior to the introduction of vaccination.42 On the other hand, these studies did not consider changes in the cost of vaccination, which is one of the variables with the highest impact on the analysis of cost-effectiveness.27

Vaccines against RV have proven to be safe and well tolerated.43 The summaries of product characteristics describe an increased risk of intussusception that was detected in post-authorization surveillance, estimated to amount to 6 additional cases per 100000 children in the 7 days after administration of the first dose. The risk seems to be associated with the age at which the vaccine is administered, so the European Society for Paediatric Infectious Diseases recommends administration of the first dose of vaccine between age 6 and 12–15 weeks, preferably at 6–8 weeks.43

In a study conducted in Valencia that included 136 confirmed cases of hospitalization due to intussusception, 35 (26%) occurred in vaccinated children, and 3 of the latter occurred between 1 and 7 days after administration of the first dose of RV vaccine, corresponding to an adjusted incidence rate ratio of 4.7 (95% CI, 0.3–74.1).44 The authors concluded that the small number of cases of intussusception detected despite the long period under study (5 years) in a country where based on previous studies the baseline incidence of intussusception is high was encouraging. In addition, the incidence of hospital admission due to intussusception has been declining in Spain after the introduction of the vaccines against RV.44

Other aspects related to RV infection and vaccination against RV that have been studied in Spain include:

• –

  The potential impact of vaccination on the microbiota: a comparison of the composition of the microbiota in children aged 12–15 months vaccinated with 3 doses of RV5 and the microbiota of unvaccinated children did not find any differences in the long term.45

• –

  The impact of vaccination against RV on hospital admissions due to seizures: the studies conducted in Spain have obtained contradictory results,46,47 so this is an area that requires continued investigation through high-quality studies designed to appropriately assess effect size.

• –

  Reliability of diagnosis of RV: one study estimated that there are up to 51% of false-positive results in vaccinated children when immunochromatography is used for diagnosis during periods when the prevalence of disease is low, due to a reduction in the positive predictive value. Interpreting such false-positive results as cases of vaccine failure may lead to a lack of confidence in the vaccine and therefore to a decrease in vaccination coverage.48 It may also have an impact in the estimation of the effectiveness of the vaccine, which would be artificially deflated.