ariaBarsoum Zak

Abstract Background Rotavirus is the primary cause of gastroenteritis in children worldwide and is a leading cause of gastroenteritis in children, with a significant burden. Rotavirus vaccine became available in Ireland in 2016. This study aimed to investigate hospital admissions and seasonal characteristics of rotavirus gastroenteritis in a pre- and post-vaccination period in a single district general hospital.Methods In the post-vaccination year, from November 18th 2016 to November 18th 2017, all children up to 3 years of age who presented to Mayo University Hospital with vomiting and diarrhea, were recruited and had their stool tested for rotavirus. Retrospective analysis of hospital data of children of the same age during pre-vaccination years (2014–2016) were used for comparison.Results Compared with the pre-vaccination years (2014–2016), the median percentages of reduction of rotavirus positive stool requests and hospital admissions were high, 48.5% and 73%, respectively. In the post-vaccination year, the median percentage of reduction of emergency department presentation (stool requests) with gastroenteritis was 9%. No delay in the onset of RV season or reduction of the peak of RV infection was noted in the post-vaccination year. The duration of rotavirus season in 2016/2017 was short.Conclusions Compared with 3 pre-vaccination years, the total number of gastroenteritis presentations, gastroenteritis hospital admissions, and rotavirus positive gastroenteritis cases were all reduced, and the duration of the rotavirus season was shorter.

Keywords Gastroenteritis · Rotavirus · Vaccination

Introduction

Worldwide vaccine efficacy trials demonstrated that rotavirus (RV) vaccine was effective in reduction of RV hospitalizations by 65–84% (US, Latin America) [ 1, 13] and by 85% (Australia) [ 8, 9] and RV incidence and hospitalization in areas with high vaccine coverage by 84% and 60%, respectively (Germany) [ 2, 3]. In addition, severe RV-induced gastroenteritis (GE) requiring hospitalization was virtually eliminated in vaccine-eligible children during the 3 years following vaccination in Finland [ 4]. Stopping the use of RV vaccine led to increased hospitalization in Spain in 2010 which was caused by a quality problem due to the detection of DNA fragments of porcine circovirus [ 5]. A decrease in the number of nosocomial RV infections after RV vaccine introduction in several countries was noted [ 6, 8, 9]. RV vaccine (Rotarix?) was added to the primary immunization schedule in Ireland in 2016 to be given at 2 and 4 months of age to all babies born in Ireland since October 2016. This vaccine may prevent approximately 2000 GP visits, 3280 A&E attendances and 2490 hospitalizations [ 7]. Across Europe and worldwide, following its introduction, RV vaccine has led to a significant reduction of RV GE–associated disease [ 8– 11, 15], not only for the age group eligible for RV vaccine, but also for children born before [ 8, 9, 12]. Since the introduction of RV vaccination, RV vaccination impact studies demonstrated a significant reduction in RV disease burden in terms of hospitalisations and emergency department visits (reduced by up to 90%) and in terms of the magnitude and duration of the RV season [ 13].

Currently, two vaccines, Rotarix? and RotaTeq, have been licensed for use globally. Their use could significantly reduce the mortality associated with RV disease [ 14 ]. RV vaccine (Rotarix?) was added to the primary immunization schedule in Ireland in 2016 to be given at 2 and 4 months of age to all babies born in Ireland since October 2016. This vaccine may prevent approximately 2000 GP visits, 3280 A&E attendances and 2490 hospitalizations [ 7]. The introduction of RV vaccine would have had only a limited impact on gastroenteritis presentations in 2017, because the first doses of the vaccine were administered in December 2016 and the first infants completed their course in February 2017. The vast majority of children under 3 years of age in 2017 were unvaccinated. This study analyzed the data related to children aged less than 3 years who presented to Mayo University Hospital (MUH) with GE from 2014 to November 18th 2017.

The objectives of this study were to perform retrospective and prospective examinations of the rate of hospitalization with GE, total number of RV stool requests, total number of RV gastroenteritis (GE) positive stool samples and seasonal regional trends of RV in pre-vaccination years (2014–2016) and in the post-vaccination year (2016/17) within an Irish context (district regional hospital).

Methods

The period of prospective recruitment was November 18th 2016 to November 18th 2017. The period of retrospective analysis was January 1st 2014–18th November 2016. Retrospective data were obtained from a secure electronic database from the microbiology laboratory at Mayo University Hospital and were analysed on a weekly basis similar to the prospective data following approval of hospital management.

Inclusion criteria (prospective and retrospective periods) were as follows: (1) All children up to 3 years old that attended the Emergency Department (ED) or were admitted to hospital with vomiting and diarrhea (loose stool) or diarrheal symptoms; (2) all children on the paediatric ward who developed diarrhea 3 days (72 hours) after their admission, possible nosocomial RV (NRV); (3) all children up to 3 years old readmitted to the paediatric ward within 48 hours following recent discharge, possible NRV. Exclusion criteria (prospective recruitment only) were as follows: (1) Parents not willing to participate; (2) children presenting with chronic diarrhea due to other disease (e.g., immunodeficiency or inflammatory bowel disease); (3) children presenting with the same diagnosis within a 48-hour period.

Stool samples were tested in the National Virus Reference Laboratory (NVRL), Dublin for RV by reverse transcription real-time –(RT) PCR of RNA genome. Samples were extracted on the Roche MagNA Pure 96 as per manufacturer’s protocol 10 , and were analysed on the Applied Biosystems ABI Viia7 (RV) instrument using the ABI software version, 2.3 (vaccine strain, Rotarix ?) or version 2.1.1 (RV).

Regardless of RV vaccine status, all RV positive stool samples were tested in the NVRL for vaccine strain to determine if a wild type or a vaccine strain was present using by real-time (RT) PCR testing [as described in WHO Press Manual of rotavirus detection and characterization methods, (2009), Order number WHO/IVB/08.17].

Written consent was obtained from the parents of the children. Information leaflets were given to parents (See supplementary material 1). Ethical approval was obtained from the Clinical Research Ethical Committee at Mayo University Hospital (MUH) before study commencement (Ref TOM/DP). Approval was obtained from the hospital management to access retrospective hospital data (2014–2016).

Data management

During the prospective study period, data related to hospital outcome of disease were collected confidentially using electronic hospital laboratory system. Data were verified on a case to case basis. Internationally recognised Vesikari Scoring System assessed disease severity in the prospective phase of the study (See supplementary material 2). Assessment of severity in the retrospective analysis was not applicable. Data related to vaccine uptake rate for all eligible infants in CO Mayo following vaccine introduction were obtained from Mayo Immunisation Office. Data included the two doses of vaccine, the total number of children vaccinated and the total number of infants eligible for RV vaccine in Co Mayo, Ireland.

Definitions

Data were gathered in relation to the following: (1) total number of all RV stool samples that were requested from both the Pediatric Ward and Emergency Department from all children ≤ 3 years old, from November 18th 2016, date of the start of the study to November 18th 2017, date of the end of the prospective study; (2) total number of all RV stool samples that were requested from both the Paediatric Ward and Emergency Department from all children ≤ 3 years old from January 1st to December 31st (2014–2015), and from January 1st 2016 to November 18th 2016 (period of retrospective analysis of pre vaccination years); (3) weekly analysis of prospective and retrospective data regarding the following: the total number of RV stool samples requested, the number of positive stool samples in this week, the calculated mean or the median of positive stool samples in every 2 consecutive week period (the total number of positive stool samples in these 2 consecutive weeks divided by two), the calculated median percentage of RV positive stool samples in each 2 consecutive week period (i.e., the mean or the median of RV positive stool samples in 2 consecutive weeks divided by the total number of stool samples requested for RV infection in these 2 consecutive weeks.) Hospital stay was defined as more than one night stay in hospital.

RV positive results were noted and seasonal characteristics were analysed using the following standard definitions of RV season: (1) Onset of RV season: any 2 consecutive weeks in any month when the median percentage of RV positive stool samples was 10% or more; (2) peak of RV Season: any 2 consecutive weeks in any month when the median percentage of RV positive stool samples was the highest percentage; (3) end of RV season: any 2 consecutive weeks in any month with the median percentage of RV positive stool samples was less than 10%; (4) a short Episode of RV Infection (EOI): any short period of RV infection with no ‘peak’ of RV infection, not longer than 3 weeks with a rapid onset and a rapid end (with a median percentage of RV infection in the immediate 2 consecutive weeks following the first week of the onset of RV infection was less than 10%).

Statistical analysis

Descriptive statistics using chi-squared tests to compare proportions of RV detected and not detected pre- and postvaccination and to compare proportions of GE admitted and not admitted cases for the same period of analysis.

Results

Rotavirus epidemiology pre- and post-vaccine

RV season: pre-vaccination (2014–2016). The RV season was studied for 3 consecutive years prior to the introduction of RV vaccine in Ireland October 2016. Retrospective analysis of hospital data for 2014, 2015 included the period January 1st to December 31st, and in 2016, included the period January 1st to November 18th 2016, (date of the start of prospective analysis) on a weekly basis until the study ended (18th November 2017). RV vaccine in Ireland became available in December 1st 2016, the start of analysis of the RV season (S) for 2016/2017 (post-vaccination year).

2014 season (January 1st–December 31st)

A total of 515 stool samples were requested; 269 were male (52%), 70 were RV positive (RV + ve) (13.6%). Three seasons of RV GE and one short episode of RV infection were noted. Onset, peak, median percentage, end of season and short episodes of RV GE are shown in Table 1.

Table 1 2014 RV season & short episode of infection (EOI)

Table 1 (continued)

Table 2 2015 RV season (S) & short episode of infection (EOI)

Table 2 (continued)

Table 3 2016 (Jan. to Dec 1st) RV season (S) & short episode of infection (EOI)

Table 3 (continued)

2015 season (January 1st–December 31st)

A total of 586 stool samples were requested, 301 were male (51%). A total of 118 samples were RV positive (20%). One long season of RV and one short episode of RV infection were detected. Onset, peak, median percentage of RV infection, and end of season are shown in Table 2.

RV Season 2016 (January 1 to December 1-exclusive)

During 2016, 430 stool samples were requested; 235 were male (54.6%); 68 samples were RV positive (15.8%). Three seasons of RV GE and three short episodes of RV infection were noted in 2016. Onset, peak, median percentage of RV infection, end of season and short episodes of RV GE are shown in Table 3.

RV season post vaccination (December 1st 2016 to November 18th 2017)

From December 1st 2016 (when RV vaccine was available) to November 18th 2017, 157 stool samples were requested for testing from both paedatric ward (PW) and emergency department (ED). Of 29 samples from ED, 9 samples were excluded from testing because the samples were either missing or mislabeled. A total of 148 stool samples from 146 patients were tested (2 patients presented twice on two separate occasions). Eighty patients were male (54%), and 36 samples were RV positive (24.3%). Totally 128 children were admitted. The majority of cases were severe (28 cases, 78%), with 5 moderate (14%) and 3 mild GE (8%).

Total attendances of children with GE who were not eligible for RV vaccine (between 1 to 3 years old) were 317. One child with GE was transferred to another hospital to be admitted and was excluded because the child did not stay more than one night at MUH. Three children presented with the same diagnosis within a 48-h period and were excluded. In total, four children were excluded.

Since December 2016 when RV vaccine was available, 2,410 infants were administered vaccine in our region and 19 received RV vaccine and presented with GE (about 13% of all 146 patients who were tested for RV). Among all 148 stool samples tested for RV, vaccine strain was confirmed in only 6 RV vaccinated infants cases (4%, 2 severe, 3 mild, 1 moderate, 1 admitted) because RV first dose of vaccine was given, RV vaccine strain was not isolated from the stool of non-vaccinated cases of GE in our study (129 cases).

Four seasons of RV GE and five short episodes of RV infection were noted. Onset, peak, median percentage of RV infection, end of season and short episodes of RV GE are shown in Table 4. Seasonal variations are first outlined below followed by effect on RV testing and infection for comparisons.

Table 4 (continued)

Table 4 RV season in post-vaccination year from December 2016 to November 18, 2017 & short episode of infection (EOI)

Table 5 Rotavirus (RV) stool requests and admission rates pre-vaccination (2014-Dec 2016) and post-vaccination (Dec 2016 to Nov 2017)

RV season pre-vaccination (January 1st 2014 to December 1st 2016)

The longest season was noted in 2015 and lasted 19 weeks, with two long and high peaks of positive RV infection. The highest peak of RV positive infection was 30.8% and was noted in October 2016 following the introduction of RV vaccine in Ireland and prior to the first dose of vaccination (December 2016). The RV season started early in February and March in 2014 and 2015. The RV season onset was delayed until April in 2016. Short episodes of RV infection were noted during the three seasons (2014–2016). The highest number of these short episodes was detected in 2016, with three episodes of RV infection detected.

RV season post vaccination (December 1st 2016 to November 18th 2017)

Four seasons of RV GE and five short episodes of RV infection were noted in 2016/2017 from December 1st 2016 to November 18th 2017. Onset, peak, median percentage of RV infection, end of season and short episodes of RV GE are shown previously (Table 4).

During the study period, 128 children ≤ 3 years of age were admitted with GE, 27 patients were RV positive (21%); 15 were males (56%). Length of hospital stay was between 1 and 2 days in 17 cases (63%). Monthly figures, showing reduced total presentation with GE and peak of RV GE during post vaccination year 2016/2017 in contrast to the pre-vaccination years (2014–2016) (Figs.1 and 2. Further statistical comparisons are described below.

Pre- versus post-vaccine comparisons

From January 1st 2014 to December 2016 (pre-vaccination years), 515, 586 and 430 RV stool samples were requested, respectively. The percentage of RV positive stool samples in those years were 13.6%, 20%, and 15.8%, respectively. From December 2016–2017, 148 stool samples were requested, and 24.3% were RV positive (Table 5; Fig.3).

The total number of RV positive stool samples or requests, post-vaccination dropped significantly compared with pre-vaccination years. Reduction in the percentage of the total number of positive RV stool samples of 48.5%, 69% and 47%, compared with pre-vaccination years, with a median percentage of reduction of positive RV stool requests of 48.5%, The total number of RV stool requests from the PW or hospital admissions due to GE, dropped to 128 postvaccination compared with 483, 556 and 391 in pre-vaccination years with a percentage of reduction of GE-related hospital admissions of about 73%, 77% and 67% and with a median percentage of reduction of GE-related hospital admission of 73%.

The total number of stool requests or samples from ED, which also reflects ED presentations, dropped to 29 compared with 32, 30 and 39 in the pre-vaccination years with percentages of reduction of about 9%, 3% and 26% and with a median percentage reduction of ED presentation with GE of 9%.

Discussion

RV occurs in autumn and winter in temperate climates and throughout the year in tropical climates [ 16– 22]. Our study demonstrated that the seasonal trend of RV was variable and that RV GE may occur during summer and throughout the year during the pre- and post-vaccination periods. Various short episodes of RV infection were noted throughout the year. RV vaccine may shorten the duration of RV season [ 13]. In our study, similarly, all four RV seasons in the postvaccination year (December 2016 to November 2017) were short and did not extend beyond 7 weeks in contrast to prevaccination years (2014–2016) when most of the seasons were long. We acknowledge the limitation of not collecting data earlier than 2014 or longer than 1 year post-vaccine for further extrapolation of this trend. In our study, 19 infants received RV vaccine and presented with GE (about 13% of the 146 patients). Among all 148 stool samples tested for RV, since RV first dose of vaccine was given, vaccine strain was confirmed in only 6 RV-vaccinated infant cases (4%). Since the first dose of RV vaccine administration occurred on December 1 2016, 84 of 317 children aged 1–3 years (ineligible for RV vaccine) presented to hospital with GE (26% of the total number of all hospital presentations of children for this age group) during the study [ 12, 13, 25] period. Children may benefit from RV vaccine herd protection [ 2, 3, 6, 9, 12, 13, 25]. In our study, RV vaccine strain was not isolated from the stool of non-vaccinated cases of GE (129 cases). No cases of vaccine strain in those unvaccinated demonstrate that no episode of transmission of rotavirus from vaccination was documented. However, our study was conducted for only 1 year following RV vaccine introduction, so additional prospective study may address this question.

RV vaccine may delay the onset of RV infection [ 25] and may reduce the seasonal peak of infection [ 13]. Following vaccination in 2013 in the UK, RV season in 2014/2015 started late in April and May in 2016 in contrast to the years preceding vaccine introduction when the peak occurred in March [ 25]. During the post-RV vaccination year (December 2016 to November 18 th 2017) in our study, four seasons of RV GE and five short episodes of RV infection were noted and RV season was not delayed, started in late February-early March, almost similar to the onset of RV season in 2014 and 2015 and high peaks of RV infection continued in the immediate 1 post vaccination year. We acknowledge the limitation of this observation due to a limited sample size in a small geographical location resulting in the possible overestimation of analysis. Additional research for years in the post-vaccination era with larger geographical locations and with data from GE cases within primary care units will enable further observation of the possible effect of RV vaccine on seasonal trends of RV. RV vaccine strain was isolated from only 6 vaccinated infants, and only 19 patients presented to our local hospital with GE following their first and/or second dose of RV vaccine (0.8% of all RV vaccinated infants in Co Mayo in Ireland during the study period).Whether this was due to RV vaccine effect remains to be seen. However, we cannot assume that all 2,410 children who received RV vaccine would have presented to our hospital if they had developed GE. Several children may have presented to GPs or may have been managed at home. Moreover, the isolation of vaccine introduction strain in the stool of patients following vaccination may not necessarily be the cause of their presentation with GE. Future research is warranted to include GE cases in primary care and larger geographical locations to study the vaccination failure rate of RV vaccine.

Our study demonstrates a significant reduction of the total number of GE presentations where the total number of requested stool samples for RV testing during the post-RV vaccination year, has decreased to 157 compared with 515, 586, 430 in pre-vaccination years, both from Pediatric Ward, (hospital admissions) and from the ED (Table 5). Studies worldwide showed RV vaccine reduced the total number of presentations with GE [ 2– 6, 8, 9, 13, 23– 26]. Compared with pre-vaccination years, the total number of RV positive stool samples or requests, post-vaccination also dropped significantly, with a median percentage of reduction of positive RV stool requests of 48.5%, similar to other studies [ 9, 11, 23, 24, 27]. A statistically significant difference in the proportion of detection of RV between pre- and post-vaccination status was identified in our study. Based on statistical analysis For GE-related admission proportions, it was noted that a statistically significant lower proportion of children were admitted to the hospital after the RV vaccine was introduced [128, (81.5%)]. Admissions rate (%) due to GE infection pre-vaccination was 11.9% higher than admissions during the post-vaccination months ( supplementary material 3, 4).

However, it merits highlighting that the percentage of detection was higher in the post-vaccination group, possibly due to the smaller sample size (only 1 year post-vaccination), which is less than 10% of the pre-vaccination cohort. Sampling variability may have played a role. The significant difference is not as expected likely due to the small sample size of the post-vaccination group and also to a reduction of the total number of presentations with GE with a subsequent reduced number of total requests for RV tests may result in a slight increase of the proportion of RV positive cases in the post-vaccination year.

Vaccination uptake rate varies across Europe [ 25], and efficacy studies require high vaccine uptake rates. In Germany, 90% and 64% vaccine coverages resulted in 84% and 60% reductions of RV GE and RV hospitalisations, respectively, and a 19% reduction in the low vaccination coverage area [ 2, 3]. Recent figures from the UK show high uptake rates for first dose (94%) and second dose (90%) in 2016 [ 25].Our recent figures of RV vaccine uptake from the Co. Mayo Immunization Office, west of Ireland, revealed that 2,410 infants in this region were administered RV vaccine during the study period, with a high uptake rate of vaccine first dose (94.2%) and second dose (92.8%). With those figures, the total number of RV stool requests from the PW (GE hospitalization) dropped to 128 compared to 483, 556, 391 in prevaccination years with a median percentage of reduction of RV GE related hospital admission of 73% (Table 1). A similar trend was noted in other studies worldwide [ 1– 6, 8, 9, 13, 24, 26].The total number of stool requests from ED also dropped to 29 compared with 32, 30 and 39 in the pre-vaccination years with a percentage of reduction of about 9%, 3% and 26% and with a median percentage of reduction of ED presentation with GE of 9% similar to other studies [ 13, 26]. RotaTeq? vaccination-impact Studies in the USA, Europe and Australia between 2006 and February 2010 report high vaccine effectiveness of up to 100% with decreased RV GE hospitalizations and emergency department visits by up to 90% [ 13].Our study is not a RV vaccine efficacy study; however, our figures from the RV post-vaccination year show a reduction in the number of presentations with GE to ED, PW compared with 3 pre-vaccination years, similar to RV vaccine efficacy studies worldwide [ 1– 6, 8, 9, 11, 13, 23, 26, 27].

Our study was the first study in a regional Irish pediatric department to evaluate seasonal trends of RV, RV GE severity in vaccinated infants, and GE hospitalisation by comparing data from the immediate 3 pre-vaccination years (2014–2016) to data from the first post-RV vaccination year (2016/17). Our research has shown a reduction of hospital admissions with GE and total number of ED attendance in the post-vaccination year. This study will provide a reference point for further researches.

Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s12519- 022- 00546-0.

Acknowledgements I greatly thank Ms Zoe Yandle (Clinical Scientist at the National Virus Reference Laboratory-Dublin) for her kind assistance in relation to performing laboratory investigations and was happy with submission for publication.

Author contributions ZB was responsible for conceptualization, study design, study protocol, data collection and management, literature search, writing original draft, review & editing, and submitting the final version of the manuscript.

Funding None.

Data availability statement Available upon reasonable request.

Declarations

Conflict of interest No financial or non- financial benefits have been received or will be received from any part related directly or indirectly to the subject of this article.

Ethics approval Obtained from Ethical Committee at Mayo University Hospital prior to study commencement (Ref TOM/DP).

Consent to participate Obtained from carers of children.