15-year experience with rotavirus vaccination in Mexico: a systematic literature review

ABSTRACT A systematic review was conducted in Mexico to consolidate and evaluate evidence after 15 years of rotavirus vaccination, according to the National Immunization Program. Five databases were screened to identify published articles (January 2000–February 2020) with evidence on all clinical and epidemiological endpoints (e.g. immunogenicity, safety, efficacy, impact/effectiveness) of rotavirus vaccination in Mexico. Twenty-two articles were identified (observational studies including health-economic models: 17; randomized controlled trials: 5). Fourteen studies evaluated a human attenuated vaccine (HRV), four studies evaluated both vaccines, and only two evaluated a bovine-human reassortant vaccine, with local efficacy data only for HRV. Local evidence shows vaccines are safe, immunogenic, efficacious, and provide an acceptable risk-benefit profile. The benefits of both vaccines in alleviating the burden of all-cause diarrhea mortality and morbidity are documented in several local post-licensure studies. Findings signify overall benefits of rotavirus vaccination and support the continued use of rotavirus vaccine in Mexico.


Introduction
Diarrhea is among the leading causes of mortality in young children under 5 years of age, especially in low-and middleincome countries. 1,2 Every year, estimated 1.5 million children under 5 years of age die of diarrhea worldwide. 3 Vaccinepreventable rotavirus infection is one of the most frequent causes of gastroenteritis and diarrhea, which can lead to rapid dehydration in children under 5 years of age. 4,5 Since 2006, two live, orally administered rotavirus vaccines have been available and licensed for the prevention of rotavirus gastroenteritis: Rotarix® (GlaxoSmithKline Biologicals, Rixensart, Belgium), a two-dose human attenuated vaccine (HRV), and RotaTeq® (Merck & Co. Inc., West Point, PA, USA), a three-dose bovinehuman reassortant vaccine (BHRV). 6 Based on the World Health Organization (WHO) recommendation, rotavirus vaccines have been introduced into the National Immunization Programs (NIP) of several countries, and licensed in more than 100 countries. 6 Since then, diarrhea-associated mortality has decreased markedly over time, attributable to the widespread use of rotavirus vaccines with other contributing factors such as improvements in diarrhea treatment, sanitation and provision of safe drinking water, and other aspects related to nutrition (breastfeeding practices, vitamin A supplementation). 7 Yet, rotavirus is still responsible for high levels of diarrhearelated morbidity globally, especially in low-and middleincome nations. 8 In Latin America and the Caribbean (LAC) region, during the pre-vaccination era, it was estimated that rotavirus caused about 75,000 hospitalizations and 2 million clinic visits per year. 9 The majority of this burden peaked during the cool winter months. 9,10 After rotavirus vaccination, in the LAC region, it was observed that in children <5 years of age the number of diarrheal deaths decreased from 32,780 in 2000 to 8,750 in 2013; and deaths due to rotavirus decreased from 11,631 in 2000 to 2,288 in 2013. 9 In the LAC region, the most common G type of rotavirus is G1, which is responsible for almost half of the rotavirus diarrhea burden, followed by G4, G3, and G9, although regional and temporal variations are significant. 11 The Pan American Health Organization (PAHO) Technical Advisory Group on vaccine-preventable diseases recommends that countries in this region should continue making efforts to administer rotavirus vaccines as part of their routine vaccination schedules, at the recommended ages according to the vaccine used, usually at 2 and 4 or 2, 4 and 6 months of age. Both of these schedules, particularly the twodose with HRV which can be completed by 24 weeks of age, foster the early protection for children at the highest risk of severe disease due to rotavirus diarrhea. 10 Several Latin American countries participated and led the way in pivotal pre-licensure clinical trials. This led to a comprehensive evidence base from a substantial number of rotavirus-specific studies that are available to guide and inform vaccine policy development in the region. 12 In July 2004, HRV was first registered in Mexico after which it was introduced for routine use into the NIP of several countries in the region. 11 Brazil and Mexico were among the first to implement childhood rotavirus vaccination into their NIP. 13 In the Mexican NIP, the two-dose HRV was used from 2006 to 2011 and the three-dose BHRV has been offered since 2011. 14 The Mexican Social Security Institute (IMSS) partially re-introduced HRV into the NIP in 2019, and distributed both the vaccines through the NIP. It has been over a decade since the licensing and first introduction of rotavirus vaccination into the NIP of Mexico. Numerous studies have been published since, on the impact of vaccination on diarrheal disease burden and safety, testifying to the success of the vaccination programs. Therefore, we conducted a systematic literature review to appraise the available evidence on rotavirus vaccination in Mexico: First we describe findings on the clinical effectiveness, safety, burden of disease, costeffectiveness of vaccination, and compliance to the recommended vaccine schedule. Then, the results are evaluated to assess the overall impact of rotavirus vaccination on diarrheaassociated mortality, morbidity, and hospitalization since the implementation of the vaccine in the NIP of Mexico (see Figure 6 Plain Language Summary).

Methods
This review was conducted according to the Preferred Reporting Items for Systematic Literature Reviews and Meta-Analyses (PRISMA) guidelines. 15,16 In line with these guidelines, we developed a search strategy and established study eligibility criteria prior to conducting the review. Following this, searches were performed and retrieved articles were assessed for eligibility in a two-phase screening process and full-text review by two reviewers. From the final list of eligible publications, data were extracted based on the scope which was established a priori. A risk of bias assessment was conducted for all included studies independently by two authors.

Article selection, data extraction, and reporting
The identified articles were screened in two phases by two reviewers using the inclusion and exclusion criteria provided in Table 1. The retrieved articles were initially screened by title and abstract for eligibility by two reviewers followed by a second step which included screening of the full-text of articles using For interventional studies. *References cited by screened articles were manually reviewed for relevance (i.e. snowballing) **References of included articles in these systematic reviews/metaanalyses were manually screened for additional relevant original articles (as deemed necessary by the reviewer).
the eligibility criteria specified in Table 1. Any discrepancies were discussed and resolved with the other review authors. From each of the eligible articles, relevant information established a priori with all authors was extracted using a customized extraction form that included the following items: reference, author, journal and year, region/city, main study objectives, study type/design, study period, sample size, age group, clinical outcomes, and measures of vaccine impact.
A descriptive analysis of the extracted data was performed to summarize the main outcomes of this review.

Risk of bias assessment
A risk of bias assessment was conducted for all included observational studies and randomized controlled trials (RCTs). The risk of bias for observational studies was assessed using the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) 17 Table 2). 18,19 An algorithm programmed into a spreadsheet was used to estimate a summary assessment of risk of bias considering five criteria: methods for selecting study participants, methods for measuring exposure and outcome variables, and methods to control confounding, designspecific sources of bias, and statistical methods. The risk of bias of each study was rated as high, moderate, low, or doubtful. The Cochrane risk of bias tool was used to assess RCTs and clinical controlled trials (Supplementary Table 3). 20 The criteria for judging risk of bias were adequate sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting, and other sources of bias. The risk of bias of each study was rated as high, moderate and low. The Cochrane Effective Practice and Organization of Care (EPOC) quality criteria were used to assess the risk of bias of the controlled before and after studies and interrupted time series. 21 The risk of bias assessment was conducted independently by two authors and any disagreements were resolved by consensus through discussion with the authors.

Characteristics of included studies
The literature search yielded 294 articles; of these 114 articles were screened at the title, the abstract phase, and finally 32 articles were screened at the full-text phase ( Figure 1). After full-text screening, 22 articles were included in this review ( Figure 1; Table 2). 14,22-43 Among these 22 articles, 14 studies were conducted only in Mexico, whereas eight studies were conducted in the Latin America region and included Mexico ( Figure 1; Table 2).
An overview of the study characteristics is presented in Figure 1 with individual study details provided in Table 2. Of the 22 studies, 17 were observational studies and 5 were RCTs. A majority of the studies included children ≤5 years of age (n = 19 studies) among which the distribution is as follows: infants <2 years (n = 6 studies), <1 year (n = 2), and <3 and <4 years each (n = 1 each).
Fourteen and two studies evaluated HRV and BHRV, respectively, and four studies evaluated both vaccines. The distribution of studies by vaccine and type of study outcome is provided in Figure 2 and Figure 3. Both vaccines, HRV and BHRV, had data for all outcomes with the exception of local efficacy data which were not identified for BHRV.

Burden of disease
One study reported estimates of the burden of disease prior to the implementation of rotavirus vaccination in Mexico in 2006. It estimated the percentage of rotavirus gastroenteritis cases among all-cause acute gastroenteritis cases in children <3 years of age at 59% (2003). 30 A second study estimated the effect of rotavirus diarrhea on disability-adjusted life-years (DALYS) and diarrhea treatment costs in hypothetical cohorts of infants who were followed from birth up to 5 years of age. 28 From birth to the age of 5 years, the estimated DALYs were 19,426 in 2001 and decreased by 28.9% in 2006, meanwhile costs of treatment were relatively constant, estimated at US$ 38.7 million and increased only by 5% ( Table 2). 28

Immunogenicity and efficacy of rotavirus vaccine
Five RCTs with Mexican participants 24,31,38-40 provided evidence on the immunogenicity (n = 3) 24,38,40 and efficacy (n = 4) 31,38-40 of rotavirus vaccination (Table 2; Figure 3). The immunogenicity of HRV was reported in two studies which showed that most of the infants had seroprotective levels of antibodies when co-administered with the oral polio vaccine and other routine vaccinations. 39,40 In the first phase 2b, randomized, dose-response study, range of seroconversion rates was 34.2-63.9% 2 months after the first dose and increased to 50%-70.6% two months after second dose in infants 6-12 weeks of age. Geometric mean titers were high and sustained after the completion of two doses. 38 In a second randomized, placebo-controlled study, the efficacy of different concentrations (10, 4.7 10 5.2 or 10 5.8 focus-forming units [FFU]) of HRV was evaluated in infants of 6-13 weeks of age. The study reported seroconversion rates of 38% (10 4.7 ) to 43% (10 5.8 ) two months after the first dose and ranged between 61% (10 4.7 ) and 65% (10 5.8 ) two months after the second dose. 40 The immunogenicity of BHRV was reported in one study which lends support to the concomitant use of BHRV and the oral poliovirus vaccine. 24 While the immunogenicity of OPV did not change when co-administered with BHRV, there was a reduction in the anti-RV IgA titers when given at the same time as OPV, yet children still met the criteria for seroconversion.
Vaccine efficacy was investigated in four studies for HRV. 31,[38][39][40] Overall, the evidence from RCTs shows that HRV is efficacious in preventing against severe and any rotavirus gastroenteritis, with efficacy ranging from 77%-100% and 70%-80%, respectively. 31,[38][39][40] The efficacy of rotavirus vaccination against hospitalizations due to any cause of gastroenteritis and severe rotavirus gastroenteritis was 42% and 85%, respectively. 39 In one study among infants <2 years of age vaccine efficacy was high against severe rotavirus gastroenteritis and sustained up to the third year of life (82.1%-100%). 31

Safety of rotavirus vaccine
A total of seven studies carried out in Mexico provided local evidence of an adequate safety profile of both rotavirus vaccines. While six of the seven studies reported safety data for HRV (randomized [n = 5]; non-randomized[n = 1]), 22,31,[38][39][40]42 only one provided safety data for BHRV (randomized) 24 (Table 2). Overall, both vaccines were well tolerated among vaccine recipients with low rates of serious adverse events including a low risk of intussusception. Both vaccines showed an acceptable safety profile when co-administered with the oral polio vaccine and other routine vaccinations.
Significantly fewer serious adverse events were reported among infants who received HRV compared to those who did not receive the vaccine (i.e. placebo). 22,38,40 While the majority of studies showed that HRV was not associated with an increased risk of intussusception during a 31-day window after administration of the first or second dose versus placebo, 31,38,39 other studies indicate a low risk of intussusception, specifically a temporal increase in the risk for intussusception within 7 days of administration of the first vaccine dose. 22,42 In the largest surveillance study for intussusception after rotavirus vaccination to date, the relative incidence of intussusception within 31 days of vaccination was 1.75 (p = .001) after the first dose and 1.06 (p = .75) after the second dose; and within 7 days of vaccination, the relative incidence was 6.49 (p < .001) after the first dose and 1.29 (p = .29) after the second dose. 42 The health benefits of vaccination, in terms of absolute number of deaths and hospitalizations averted, far outweigh the risk of short-term probable side effects which rarely have complications. 44 A randomized study that evaluated the concomitant use of BHRV with the oral poliovirus vaccine compared to BHRV alone in infants showed a similar safety and tolerability profile between both regimens (Table 2). 24

Health economics of rotavirus vaccination
The cost-effectiveness of rotavirus vaccination in Mexico has been elucidated in four publications (HRV [n = 2]; BHRV [n = 1]; both vaccines [n = 1]). 23,25,34,41 Overall, the two-dose vaccination schedule with HRV or the three-dose vaccination schedule with BHRV was associated with higher net savings and gain in quality-adjusted life-years (QALY) compared with no vaccination (Table 2). 25,34,41 Only one analysis directly compared HRV and BHRV. 23 For both vaccines, the economic evaluation projected a reduction in rotavirus events by 39% for HRV and 30% for BHRV, a reduction in the frequency of cases seeking medical advice by 58% for HRV and 45% for BHRV, and a decrease in hospital admissions by 67% for HRV and 53% for BHRV. The two-dose vaccination schedule with HRV was associated with a net savings of 74 million Mexican pesos (MXN) plus a gain of 553 QALY when compared with the three-dose schedule, with BHRV indicating that vaccination with HRV was the most cost-effective strategy ( Table 2). 23

Impact/effectiveness of rotavirus vaccination on mortality and morbidity
Evidence on the impact of rotavirus vaccination on acute diarrheal disease mortality was reported in five studies (HRV [n = 4]; both vaccines [n = 1]) ( Table 2). 26,27,32,35,37 Overall, a substantial decline in all-cause diarrhea mortality rate was observed in children under 5 years of age after the implementation of rotavirus vaccination in Mexico, 26,27,32,35,37 regardless of the choice of vaccine. The majority of the studies provide evidence of vaccine-specific impact for HRV and one study assessed the overall impact on mortality for a period of 10 years without a differentiation in the vaccine used. During the time when HRV was implemented in the NIP in Mexico, a significant decline in all-cause diarrhea mortality and deaths due to acute diarrheal disease among children under 5 years of age was observed. 26,27,35,37 Only one study provided evidence of the impact of vaccination with HRV in the different regions of Mexico: across the regions, mortality due to all-cause diarrhea among children aged under 5 years of age declined by 43%-55% in all regions after the implementation of vaccination with HRV (2003-2006) ( Table 2). 27 Evidence on the impact of rotavirus vaccination on all-cause diarrhea morbidity was reported in three studies (HRV [n = 2]; both vaccines [n = 1]) ( Table 2). 29,33,37 Overall, the numbers of new cases and hospitalizations due to all-cause diarrhea including acute diarrhea were reduced during 2006-2017. The first evidence of this comes from a 10-year observational study which showed that rotavirus vaccination resulted in a 15.5%-46%  • Reduction in the anti-RV IgA titers when given at the same time as OPV, yet children met criteria for seroconversion.
(Continued)  . 29 These trends of declining levels of morbidity due to rotavirus vaccination were confirmed in a second study that assessed the vaccine-specific impact of HRV vaccination: a decline of 11%-40% in all-cause diarrhea hospitalizations was observed during 2008-2009 with the greatest reduction reported in infants <12 months of age (25%-52%). In addition, among children 12-23 months of age, a 43% decline in all-cause diarrhea hospitalizations was reported during the 2009 season. 33 Vaccine effectiveness data were identified only for HRV. In an observational case-control study, a completed 2-dose schedule with HRV resulted in an effectiveness of 94% against hospitalization due to laboratory-confirmed G9P [4] rotavirus infection. 43

Compliance of rotavirus vaccination
Evidence on compliance with the recommended vaccination schedule, including timeliness of vaccination, was reported in one study based on a registry provided by the IMSS. In this registry, there were 659,249 and 780,483 infants eligible for HRV (2010) and BHRV (2012), respectively. 14 Among these infants, compliance with full vaccine series was reported in 93.7% of infants who received HRV compared to 71.1% who received BHRV (p < .001). Likewise, the percentage of infants who completed the full vaccination series according to the recommended schedule (age and interval between doses) was higher with HRV (75.5%) compared to BHRV (70.9%) (p = .105). 14

Risk of bias
The results of the risk of bias appraisal for observational studies are shown in Figure 4. The majority of studies (10/17) were regarded as presenting a moderate risk of bias and the remaining studies presented a low risk of bias. The moderate risk of bias of individual studies was driven mainly by a lack of methods to control confounding and design-specific source of bias which can be attributed to the nature of observational studies, specifically those using passive surveillance and laboratory data (with non-probabilistic sampling methods). Observational studies have inherent biases, particularly since they are not randomized. Yet we classified most as having lowto-moderate risk of biases overall. The specific categories that contained higher bias were mostly around design-specific sources of bias (i.e. recall bias, loss to follow-up, no blinding, retrospective databases from passive surveillance systems, underreporting) and in most studies the methods for controlling confounding (i.e. appropriate design or analytical methods) were unclear/not reported. For almost all of these studies, most endpoints were descriptive with no adjustment for multiple comparisons.
The results of the risk of bias appraisal for RCT studies are shown in Figure 5. The majority of studies (4/5) were regarded as presenting a low risk of bias and one study was associated with a high risk of bias. 24 This was driven by the fact that the concealment of allocation was unclear, and it was not a blinded trial. Adding to this was the small sample size considering loss of follow-up and adherence. 24

Discussion
In this review, we summarize evidence on the burden of rotavirus gastroenteritis in Mexico, regional and local immunogenicity, efficacy and safety data of the available rotavirus vaccines, health economics, and the impact of the rotavirus vaccination program in Mexico.
In 2006, rotavirus vaccination for children was added into the Mexican NIP; HRV was used from 2006 to 2011 and partially re-introduced in 2019 until present, and BHRV has been used since 2011. Along with several Latin American countries, Mexico was one of the countries that led the accelerated clinical development of rotavirus vaccines. Since the beginning of the rotavirus vaccination program in Mexico, several studies have been conducted to assess the local immunogenicity, efficacy, and safety of rotavirus vaccines. Local immunogenicity data available only for HRV show that infants had seroprotective levels of antibodies after both vaccine doses, 38,40 whereas immunogenicity data to support the use of rotavirus vaccines with the oral polio vaccine and other routine vaccinations were available for both HRV and BHRV. 24,38,40 Local efficacy data were reported in five studies, all of which were specific to HRV. 31,[38][39][40]43 Overall, the local efficacy of HRV among children <5 years of age is high against severe (77%-100%) and any rotavirus gastroenteritis (70%-80%), 31,[38][39][40] including hospitalizations (all-cause: 42%; severe rotavirus gastroenteritis-related hospitalizations: 85%). 39 According to local studies, both rotavirus vaccines show an acceptable safety profile without a severe risk of intussusception. However, a temporal increase in the risk for intussusception was observed within 7 days of receipt of the first vaccine dose. 22,42 Whether rotavirus vaccination has any impact on the overall incidence of intussusception is yet to be  determined. 22,42 Importantly, this finding should be interpreted along with the well-documented benefits of rotavirus vaccination, demonstrating a high benefit versus risk profile.
Over more than 15 years after implementation of the childhood rotavirus vaccination program in Mexico, a substantial reduction in the diarrheal disease burden primarily among children <5 years of age has been documented. These findings correspond with the trends observed from other Latin American countries such as Brazil and Panama which were also early in their implementation of a national rotavirus vaccination program. 45,46 In Mexico, G9, a strain fully heterotypic from the vaccine strain, has emerged as an important serotype causing severe rotavirus gastroenteritis. 11,47 We identified one study that showed high vaccine effectiveness (94%) against laboratory-confirmed G9P [4] rotavirus infection, 43 indicating that the strain predominance in Mexico was unrelated to vaccine pressure. Because variations in rotavirus types can occur independently of vaccination, the role of vaccination in observed strain changes requires cautious interpretation. 11 In 2017, a systematic review and meta-analysis was conducted to analyze efficacy, safety, and effectiveness of BHRV and HRV rotavirus vaccines used in the LAC region. This review highlights that the risk of any-severity rotavirusrelated gastroenteritis was reduced by 65% following rotavirus vaccination; both vaccines significantly reduced the risk of hospitalization and emergency visits by 85%-90% and did not increase the risk of death, intussusception, or severe adverse events. 48 Our review reaffirms these previous findings from the region that rotavirus vaccination was effective with a good riskbenefit profile in children. Evidence on compliance to the HRV and BHRV vaccination schedule shows a better compliance (age and interval between doses) with two-dose HRV throughout Mexico, while regional differences were observed with BHRV. 14 In the majority of health economic evaluations for Mexico, rotavirus vaccination was compared with no vaccination. Only one study that directly assessed the cost-effectiveness of HRV and BHRV was identified in this review; this analysis suggests that vaccination with HRV is a much more cost-effective strategy when compared to vaccination with BHRV. 23 Findings from health economic evaluations of the rotavirus vaccination program in Mexico underscore the benefit of continuing the rotavirus vaccination program in Mexico. Notably, extensive economic evaluations were performed in the LAC region during the time vaccine introduction decision-making processes were ongoing. 11 As more data on the vaccine-specific effectiveness of rotavirus vaccination programs become available, further economic analyses are needed to make evidencebased decisions for universal use of rotavirus vaccinations. These analyses would support the ongoing discussions on changing vaccine policy in Mexico based on new epidemiological data or the availability of new rotavirus vaccines. 49,50 With regard to new rotavirus vaccines, recently, ROTAVAC™ (Bharat Biotech, Hyderabad, India) and RotaSIIL, (Serum Institute of India, Pune, India) received WHO prequalification. 51 These vaccines are anticipated to expand the global reach of rotavirus vaccines by improving on certain programmatic aspects of HRV and BHRV, like heat stability, reduction of cold-chain footprint, and potentially providing more cost-effective options. 52 The initial Phase 3 clinical studies of both ROTAVAC and RotaSIIL reported no intussusception events in the first month following any dose of vaccine or placebo; however, these studies are of limited size and geographic scope and thus do not have extensive safety results nor an established risk-benefit profile. 53,54 These efficacy results are similar to the results from the clinical trials of BHRV and HRV which showed a lower efficacy in low-and middle-income nations with high diarrhea-related mortality. Based on the limited clinical trial data available for these new vaccines, the vaccine efficacy against severe rotavirus disease was 56% for ROTAVAC (in 3 sites in India) and ranged from 37% (in 6 sites in India) to 67% (1 site in Niger) for RotaSIIL. [53][54][55] Currently, these vaccines have not been evaluated in Latin American or Mexican populations, and their three-dose schedule might limit their utilization in these countries.
A few limitations of this review are worth noting in the interpretation of the overall findings. Systematic reviews are high in the hierarchy of evidence generation, but they always have specific (inherent) biases such as publication bias. To deal with these biases, we had two reviewers during the screening, and eligibility process and all discrepancies were discussed among the reviewers to reach consensus on the outcome. Additionally, the risk-of-bias evaluations were done as part of the quality assessment of each article in order to reduce biases during the interpretation (i.e. putting less weight on the articles with high risk of bias/lower quality). For this review, we had a wide scope covering a diverse array of clinical and epidemiological endpoints with different time periods considered in the studies. This may have led to the dilution of the individual findings. However, the focus on a single country which has licensed use of both rotavirus vaccines allowed us to meet our review objective to consolidate and integrate all existing evidence on the situation of rotavirus diarrhea in Mexico. Consequently, generalizability to other countries in the region or middle-income countries is limited.

Conclusions
This systematic review underscores the documented benefit of the childhood rotavirus vaccination program in Mexico more than 15 years after its implementation, specifically in terms of good efficacy/immunogenicity, clinical and real-life effectiveness, a favorable safety and tolerability profile, and substantial reductions in diarrhea-related mortality and hospitalizations. Both HRV and BHRV vaccines have been widely used, and this review highlights that rotavirus vaccines have a large and robust evidence base in Mexico, extending from clinical trials to real-world evidence, and the high compliance rate of HRV with the two-dose schedule, provides confidence in its continued use in all Mexican infants.

Disclosure of potential conflicts of interest
AGH, EO, MYCA, DC, and RC are employed by the GSK group of companies. MYCA, EO, and RC hold shares in the GSK group of companies. AAF is an external GSK employee hired by Randstad. All authors declare no non-financial relationships and activities.

Trademark
Rotarix is a trademark owned by or licensed to the GSK group of companies.
RotaTeq is a trademark of Merck Sharp & Dohme Corp. ROTAVAC is a trademark of Bharat Biotech International Limited. RotaSIIL is a trademark of the Serum Institute of India.