Abstract
Background:
Rotavirus vaccines are effective in preventing severe rotavirus. Haiti introduced 2-dose monovalent (G1P[8]) rotavirus vaccine recommended for infants at 6 and 10 weeks of age in 2014. We calculated the effectiveness of rotavirus vaccine against hospitalization for acute gastroenteritis in Haiti.
Methods:
We enrolled children 6–59 months old admitted May 2014-September 2019 for acute watery diarrhea any sentinel surveillance hospital. Stool was tested for rotavirus using enzyme immunoassay (EIA) and genotyped with multiplex one-step RT-PCR assay and Sanger sequencing for stratification by genotype. We used a case-negative design where cases were children positive for rotavirus and controls were negative for rotavirus. Only children eligible for vaccination were included and a child was considered vaccinated if vaccine was given ≥14 days before enrollment. We used unconditional logistic regression to calculate odds ratios and calculated 2-dose and 1-dose vaccine effectiveness (VE) as (1 − odds ratio) * 100.
Results:
We included 129 (19%) positive cases and 543 (81%) negative controls. Among cases, 77 (60%) were positive for equine-like G3P[8]. Two doses of rotavirus vaccine were 66% (95% CI: 44, 80) effective against hospitalizations due to any strain of rotavirus and 64% (95% CI: 33, 81) effective against hospitalizations due to the equine-like G3P[8] genotype.
Conclusions:
These findings are comparable to other countries in the Americas region. To the best of our knowledge, this is the first VE estimate both against the equine-like G3P[8] genotype and from a Caribbean country. Overall, these results support rotavirus vaccine use and demonstrate the importance of complete vaccination.
Introduction
In the pre-vaccine era, rotavirus was the leading cause of diarrhea deaths among children <5 years old worldwide [1]. Oral rotavirus vaccines are effective in preventing severe rotavirus disease and deaths [2–5]. However, a gradient in oral rotavirus vaccine performance has been documented globally and among Latin American countries, where oral rotavirus vaccine is most effective in countries with low <5 year old child mortality and less so in countries with high and middle <5 year old child mortality [2, 6]. Full series vaccine effectiveness (VE) against rotavirus acute gastroenteritis hospitalizations was estimated to be 67% overall in Latin American countries and 58% in high child mortality countries globally [2, 7]. In 2015, nearly 1,000 deaths due to rotavirus among children <5 years old were estimated to have been prevented in Latin American and Caribbean countries that used rotavirus vaccines as part of their routine infant vaccination program [1, 6].
In April 2014, monovalent (G1P[8]) human rotavirus vaccine (Rotarix, GlaxoSmithKline) was introduced in Haiti following a 2009 recommendation by the World Health Organization (WHO) that all countries add rotavirus vaccine into their immunization schedule [8]. In Haiti, the 2-dose rotavirus vaccine series is recommended to be co-administered with pentavalent (tetanus, diphtheria, pertussis, Hemophilus influenzae B, and hepatitis B), pneumococcal, and oral polio vaccines to infants at 6 and 10 weeks of age. Since its introduction, national 2-dose rotavirus vaccine coverage among children <1 year old has been increasing and, by 2019, the Haitian Ministry of Health estimated rotavirus vaccination coverage to be 63% [9].
Given that Haiti is one of the countries with the lowest socio-economic status in Latin America and has high <5 year old child mortality, evaluation of rotavirus VE in this setting is of great public health interest. Monitoring rotavirus vaccine performance by ecological measures has been uniquely challenging in Haiti for two main reasons. First, a cholera epidemic that began in 2010 inflated the number of all-cause acute gastroenteritis (AGE) hospitalizations and waned through the initial 5-year surveillance period when rotavirus hospitalizations were also expected to decrease due to rotavirus vaccine introduction [10]. Aside from the number of diarrhea illnesses, the spread of cholera also had an important impact on healthcare and healthcare seeking behavior [11–13]. Thus, the percent of stool specimens that tested positive for rotavirus increased in some post-vaccine years, following the decline of the cholera epidemic, despite the declining absolute number of AGE and rotavirus hospitalizations. Second, equine-like G3P[8] reassortant rotaviruses emerged in 2013 in Australia and spread worldwide, including to the Dominican Republic, which shares the island of Hispaniola with Haiti [14, 15]. From 2015–2019, equine-like G3P[8] was detected in as many as 95% of rotavirus positive stool specimens from children <5 years old in Haiti annually [16]. Effectiveness of monovalent rotavirus vaccine against diarrhea due to the equine-like G3P[8] genotype is unknown, though rotavirus vaccines have been shown to have cross protection against other non-vaccine rotavirus genotypes [17].
To better understand rotavirus vaccine implementation in Haiti, direct measures of vaccine performance were needed. In this analysis, we evaluated rotavirus VE against hospitalizations among children <5 years old due to any strain of rotavirus and the equine-like G3P[8] rotavirus genotype.
Methods
We enrolled children 6–59 months of age who were admitted from May 1, 2014 to September 30, 2019 for acute watery diarrhea (defined as 3 or more episodes of diarrhea lasting <7 days) at any of the 16 hospitals participating in Haiti’s laboratory-enhanced surveillance system (PRESEpi), a sentinel hospital surveillance system that enrolls people of all ages for 4 syndromes (acute watery diarrhea, acute febrile illness, meningitis, acute respiratory illness) in all 10 of Haiti’s administrative departments. Children <6 months old were excluded to allow time for children to have the opportunity to have received both rotavirus vaccine doses. PRESEpi and its enrollment methodology has been described in earlier publications [18, 19]. For every child who was enrolled, we collected demographic information and information about their symptoms in a structured enrollment questionnaire. We recorded rotavirus vaccination history from the child’s vaccination card. For this analysis, we excluded children who were not age-eligible to receive rotavirus vaccine, that is children born ≥2 months before rotavirus vaccine introduction on March 1, 2014. If no date of birth was reported, we relied on parental report of the child’s age. Children without written documentation of vaccination status, those with a vaccine card that indicated rotavirus vaccine was received but did not have a date of rotavirus vaccination recorded, and those without dates of vaccination for any vaccines were excluded from the analysis due to insufficient information. A child was considered immunized if each dose of the vaccine was given ≥14 days before the date of enrollment.
Stool specimens were collected within 48 hours of admission to the sentinel hospitals and tested for rotavirus using Rotaclone (Meridian) enzyme immunoassay (EIA) at the National Public Health Laboratory (LNSP) in Port-au-Prince, Haiti. All positive specimens were genotyped using multiplex one-step Real-Time Polymerase Chain Reaction (RT-PCR) assay and Sanger sequencing for confirmation at the US Centers for Disease Control and Prevention in Atlanta, GA. Additional information about genotyping is presented in detail elsewhere [16]. Specimens that were positive by EIA but negative by RT-PCR were excluded from the analysis.
Sample size was calculated to achieve 80% power at the 5% significance level with a VE of at least 50% with two-dose rotavirus vaccination coverage of 50%. We estimated that we would need 103 rotavirus-positive cases with 2:1 ratio of controls to cases. We calculated VE using a test-negative case-control design, where children who tested positive for rotavirus were considered cases and those who tested negative were considered controls. We used an unconditional logistic regression model to estimate 2-dose and 1-dose VE among all children <5 years. VE was calculated as (1 − odds ratio) * 100. We assessed categorical age in months, enrollment year, and month of enrollment and enrollment hospital as potential confounding variables in our model by the backward elimination method; we controlled for year of enrollment in all VE models. In addition to two-dose and one-dose VE, we stratified the VE analysis by age group (6–11 months and 12–23 months) and genotype. Because there were few cases of other genotypes detected in our population, we limited the genotype-specific analysis to the equine-like G3P[8] strain and grouped the other sequenced genotypes together. Data was collected using Microsoft Access and DHIS2 and analyzed using SAS v.9.4.
Activities conducted as part of PRESEpi were considered routine public health surveillance and received a non-research determination from the Centers for Disease Control and Prevention, Atlanta, GA, and the Haitian Bioethical Review Committee, Port-au-Prince, Haiti.
Results
Overall, 3249 children aged 6–59 months were enrolled for AGE during the surveillance period, had rotavirus testing results, and were age-eligible to have received rotavirus vaccine. Of these, we excluded children with insufficient documentation of vaccination status, which included 1548 (48%) children who did not have a vaccination card and an additional 1029 (32%) who had ambiguous documentation (Supplementary Table 1). The median age was 15 months (IQR: 9–25) for children without a card and 13 months (IQR: 9–21) for children with insufficient documentation; 60% of children without a card and 62% with ambiguous documentation were enrolled at surveillance sites in Ouest Department, where Port-au-Prince, the capital and largest city of Haiti, is located. Among children who were excluded due to insufficient vaccination documentation, 379 (15%) were positive for rotavirus and 2198 (85%) were negative.
The final analysis included 129 (19%) cases and 543 (81%) controls. Among both cases and controls, approximately 50% of children were 6–11 months old and 75% were enrolled at hospitals located in Ouest Department (Table 1). The distribution of sex and year of enrollment was also similar between the two groups. Two-dose rotavirus vaccination coverage was 67% among cases and 82% among controls. Of cases, 77 (75%) were positive for equine-like G3P[8], 9 (7%) were positive for G1P[8], and 16 (16%) were positive for other genotypes (Table 2). Genotyping results were not available for 27 (21%) cases; none were non-typeable.
Table 1.
Characteristics of rotavirus-positive cases and rotavirus-negative controls aged <5 years old. Haiti, 2014–2019.
| Cases (rotavirus positive) | Controls (rotavirus negative) | |||||
|---|---|---|---|---|---|---|
| all enzyme immunoassay (EIA) positive n= 129 | Equine-like G3P[8] n=77 | n= 543 | ||||
| n | % | n | % | n | % | |
| Age in months | ||||||
| 6–11 | 63 | 49 | 36 | 47 | 298 | 55 |
| 12–23 | 53 | 41 | 35 | 45 | 210 | 39 |
| 24–59 | 13 | 10 | 6 | 8 | 35 | 6 |
| Sex | ||||||
| Female | 50 | 40 | 31 | 41 | 238 | 45 |
| Male | 75 | 60 | 44 | 59 | 292 | 55 |
| Department of enrollment facility | ||||||
| Ouest | 99 | 77 | 60 | 78 | 396 | 73 |
| Other | 30 | 23 | 17 | 22 | 147 | 27 |
| Number of rotavirus vaccine doses | ||||||
| 0 doses | 31 | 24 | 18 | 23 | 68 | 13 |
| 1 dose | 12 | 9 | 8 | 10 | 31 | 6 |
| 2 doses | 86 | 67 | 51 | 66 | 444 | 82 |
| Year of enrollment | ||||||
| 2014 | 0 | 0 | 0 | 0 | 6 | 1 |
| 2015 | 24 | 19 | 13 | 17 | 145 | 27 |
| 2016 | 51 | 40 | 35 | 45 | 215 | 40 |
| 2017 | 1 | 1 | 0 | 0 | 29 | 5 |
| 2018 | 30 | 23 | 21 | 27 | 107 | 20 |
| 2019 | 23 | 18 | 8 | 10 | 41 | 8 |
Table 2.
Rotavirus genotypes found among rotavirus-positive cases aged <5 years old. Haiti, 2014–2019
| Genotype | N=102 | % |
|---|---|---|
| Equine-like G3P[8] | 77 | 75 |
| G1P[8] | 9 | 9 |
| G12P[6] | 6 | 6 |
| G12P[8] | 4 | 4 |
| G3P[6] | 2 | 2 |
| G4P[6] | 2 | 2 |
| G2P[4] | 1 | 1 |
| G9G4P[8] | 1 | 1 |
Strain not available for 27 cases. None of the cases were non-typeable.
Two doses of rotavirus vaccine were 66% (95% CI: 44, 80) effective against hospitalizations due to any strain of rotavirus among children <5 years old (Table 3). One dose of rotavirus vaccine was 45% (95% CI: −33, 77) effective against hospitalizations due to any strain of rotavirus among children <5 years old. The VE during the first and second years of life was similar in this population; a full series of rotavirus vaccine was 65% (95% CI: 29, 83) and 69% (95% CI: 30, 86) effective in preventing rotavirus hospitalizations among children 6–11 and 12–23 months old, respectively. Rotavirus vaccine was 64% (95% CI: 33, 81) effective against hospitalizations due to the equine-like G3P[8] genotype and 62% (95% CI: −2, 86) effective against hospitalizations due to all other rotavirus genotypes.
Table 3.
Monovalent rotavirus vaccine effectiveness against hospitalization for rotavirus diarrhea among children aged <5 years old. Haiti, 2014–2019
| VE | 95%CI | |
|---|---|---|
| Overall | ||
| 1 dose | 45 | −33, 77 |
| 2 doses | 66 | 44, 80 |
| Full series by age group | ||
| 6–11 months | 65 | 29, 83 |
| 12–23 months | 69 | 30, 86 |
| Full series by genotype | ||
| Equine-like G3P[8] | 64 | 33, 81 |
| All other genotypes | 62 | −2, 86 |
Discussion
During the first five years after rotavirus vaccine introduction in Haiti, rotavirus vaccine was effective in preventing about two thirds of AGE hospitalizations due to rotavirus in the first and second years of life in this study population. To the best of our knowledge, this is the first VE estimate from a Caribbean country [2]. Earlier findings in Haiti showing a 35% reduction in rotavirus disease among children <5 years old following rotavirus vaccine introduction [10], which is consistent with vaccination coverage of 50–60% and estimated VE of 66%. Our VE estimates in Haiti, a high child mortality country, are comparable to other oral rotavirus VE findings from countries in the Americas region and are somewhat higher than other high child mortality countries globally [2, 6, 7]. Reasons for the gradient in oral rotavirus vaccine performance by child mortality setting are not well understood, but previous evaluations have hypothesized nutritional status, microbiome composition, concomitant administration with live oral polio vaccine, and maternal antibodies as possible factors in lower effectiveness [20]. Like evaluations in other countries [2], our findings show 2 doses of rotavirus vaccine were likely more effective than a single dose in Haiti in preventing hospitalizations due to rotavirus disease. Overall, these results support continued routine rotavirus vaccine use in Haiti and demonstrate the importance of complete vaccination.
We found that VE against hospitalizations due to the equine-like G3P[8] strain of rotavirus offered by Rotarix was similar to effectiveness against all other rotavirus genotypes. Previous evaluations have shown that monovalent rotavirus vaccine protects against rotavirus genotypes in addition to G1P[8], the genotype included in the live, attenuated vaccine [17]. However, this is the first assessment of rotavirus VE against the novel equine-like G3P[8] strain in any country. As this equine G3 strain has emerged to become globally widespread in recent years, our finding of comparable VE against this strain is reassuring.
In this evaluation, the large proportion of children excluded due to inadequate vaccine documentation, either lacking a card or ambiguous documentation, is an important limitation. Half of eligible children did not have a vaccination card available for review by study staff, and therefore were excluded from the analysis. The percentage of children with vaccination cards in our study was somewhat lower than what was observed in the most recent Demographic and Health Survey (DHS) conducted in Haiti in 2016–2017, where 68% of children 12–23 months and 59% of children 24–35 months had a vaccination card available [13]. In the DHS, vaccination card retention was higher in urban areas and households in the highest socio-economic quintile and 2-dose rotavirus vaccine coverage was higher among children with vaccine cards than among those whose caregivers verbally reported their vaccination status [13]. In our study, we did not collect information regarding the socioeconomic status of the children’s households, but the low percentage of eligible children with a vaccination card is a potential source of bias in this evaluation, especially as written documentation likely was not random. Indeed, children excluded because of card documentation were disproportionately enrolled at sites outside of Ouest Department (60% of excluded were enrolled in Ouest vs. 75% of included). Additionally, if vaccine card retention was greater among households with higher socio-economic status, the results may not be generalizable to the whole of the Haitian population. Excluded children were also slightly older than included children, another possible explanation for low card retention. Reassuringly, the distribution of rotavirus positive and negative children was similar among those with and without written vaccination documentation. Additionally, vaccination status was ascertained independent of and prior to knowledge of rotavirus test results for AGE subject, so systematic bias is not likely. However further research is needed to quantify and determine the direction of this potential bias in VE estimates from Haiti due to low vaccine card capture. We are unable to comment on how, if at all, children with cards with ambiguous documentation may differ from children with clear documentation on their cards. Additionally, every effort should be made to improve vaccine card use by caregivers and vaccinators in Haiti.
This analysis has a few additional limitations. Though all children enrolled were hospitalized for their illness, we are unable to describe the severity of the diarrhea illness in this population or stratify our analysis by Vesikari dehydration severity score. In some previous evaluations, researchers have documented higher protection against severe illness than milder illness and information about severity could help contextualize the findings presented here. Moreover, there were insufficient specimens of non-G3P(8) genotypes, and therefore our genotype-specific analysis was limited to the G3P8 genotype.
We found that two-doses of rotavirus vaccine were effective in preventing hospitalizations due to rotavirus diarrhea in children < 5 years old during the first five years following rotavirus vaccine introduction in Haiti. We also found that rotavirus vaccine was effective in preventing hospitalizations due to the equine-like G3P[8] strain of rotavirus. These findings should be affirming to stakeholders in Haiti to continue working toward complete, timely, and well-documented rotavirus vaccination.
Supplementary Material
Acknowledgements:
We would like to acknowledge the PRESEpi team and field staff, especially Magdaline Sam, Marc Covens Jr Jean-Baptiste, Chedelene Rivière, and the participating children and their families for their contributions to this work. Funding was provided by Gavi, the Vaccine Alliance.
Funding source: Gavi, the Vaccine Alliance targeted country funds
Footnotes
Conflicts of interest: The authors indicate that they have no conflicts of interest relevant to this article to disclose.
Disclaimer: The findings and conclusions of this report are those of the authors and do not necessarily represent the official position the US Centers for Disease Control and Prevention.
References
- [1].Tate JE, Burton AH, Boschi-Pinto C, Parashar UD, World Health Organization-Coordinated Global Rotavirus Surveillance N. Global, Regional, and National Estimates of Rotavirus Mortality in Children <5 Years of Age, 2000–2013. Clin Infect Dis. 2016;62Suppl 2:S96–S105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Burnett E, Parashar UD, Tate JE. Real-world effectiveness of rotavirus vaccines, 2006–19: a literature review and meta-analysis. Lancet Glob Health. 2020;8:e1195–e202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Linhares AC, Velazquez FR, Perez-Schael I, Saez-Llorens X, Abate H, Espinoza F, et al. Efficacy and safety of an oral live attenuated human rotavirus vaccine against rotavirus gastroenteritis during the first 2 years of life in Latin American infants: a randomised, double-blind, placebo-controlled phase III study. Lancet. 2008;371:1181–9. [DOI] [PubMed] [Google Scholar]
- [4].Ruiz-Palacios GM, Perez-Schael I, Velazquez FR, Abate H, Breuer T, Clemens SC, et al. Safety and efficacy of an attenuated vaccine against severe rotavirus gastroenteritis. N Engl J Med. 2006;354:11–22. [DOI] [PubMed] [Google Scholar]
- [5].Vesikari T, Karvonen A, Puustinen L, Zeng SQ, Szakal ED, Delem A, et al. Efficacy of RIX4414 live attenuated human rotavirus vaccine in Finnish infants. Pediatr Infect Dis J. 2004;23:937–43. [DOI] [PubMed] [Google Scholar]
- [6].Chavers T, De Oliveira LH, Parashar UD, Tate JE. Post-licensure experience with rotavirus vaccination in Latin America and the Caribbean: a systematic review and meta-analysis. Expert Rev Vaccines. 2018;17:1037–51. [DOI] [PubMed] [Google Scholar]
- [7].de Oliveira LH, Camacho LA, Coutinho ES, Ruiz-Matus C, Leite JP. Rotavirus vaccine effectiveness in Latin American and Caribbean countries: A systematic review and meta-analysis. Vaccine. 2015;33Suppl 1:A248–54. [DOI] [PubMed] [Google Scholar]
- [8].Rotavirus vaccines:an update. Wkly Epidemiol Rec. 2009;84:533–40. [PubMed] [Google Scholar]
- [9].Haiti: WHO and UNICEF coverage estimates of immunization coverage: 2019 revision. https://www.who.int/immunization/monitoring_surveillance/data/hti.pdf: World Health Organization; 2020. [Google Scholar]
- [10].Anne-Marie Desormeaux EB, Joseph Gérard, Ali Ber Lucien Mentor, Aliabadi Negar, Pierre Manise, Patrick Dély Katilla Pierre, Fitter David, Leshem Eyal, Tate Jacqueline E., Bowen Michael D., Esona Mathew, Gautier Jacqueline, Florence Siné Mark A. Katz, Yoran Grant-Greene Umesh D. Parashar, Patel Roopal, Boncy Jacques, Juin Stanley. Impact of monovalent rotavirus vaccine on rotavirus hospitalizations among children <5 years of age in Ouest and Artibonite Departments, Haiti - 2013–2019. American Journal of Hygiene and Tropical Medicine. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Cayemittes M, Busangu MF, Bizimana JdD, Barrère B, Sévère B. Haïti Enquête Mortalité, Morbidité et Utilisation des Services 2012. Calverton, Maryland, USA: Ministère de la Santé Publique et de la Population - MSPP/Haïti, l’Institut Haïtien de l’Enfance - IHE and ICF International; 2013. [Google Scholar]
- [12].Cayemittes M, Placide MF, Mariko S, Barrère B, Sévère B. Haiti Enquête Mortalité, Morbidité et Utilisation des Services 2005–2006. Calverton, Maryland, USA: Institut Haïtien de l’Enfance and Macro International; 2007. [Google Scholar]
- [13].Institut Haïtien de l’Enfance - IHE/Haiti, ICF. Haiti Enquête Mortalité, Morbidité et Utilisation des Services 2016–2017 - EMMUS-VI. Pétion-Ville/Haïti: IHE/Haiti, ICF; 2018. [Google Scholar]
- [14].Cowley D, Donato CM, Roczo-Farkas S, Kirkwood CD. Emergence of a novel equine-like G3P[8] inter-genogroup reassortant rotavirus strain associated with gastroenteritis in Australian children. J Gen Virol. 2016;97:403–10. [DOI] [PubMed] [Google Scholar]
- [15].Katz EM, Esona MD, Betrapally NS, De La Cruz De Leon LA, Neira YR, Rey GJ, et al. Whole-gene analysis of inter-genogroup reassortant rotaviruses from the Dominican Republic: Emergence of equine-like G3 strains and evidence of their reassortment with locally-circulating strains. Virology. 2019;534:114–31. [DOI] [PubMed] [Google Scholar]
- [16].Lucien MAB, Esona MD, Pierre M, Joseph G, Rivière C, Leshem E, Aliabadi N, Desormeaux AM, Andre-Alboth J, Fitter D, Grant-Greene Y, Tate JE, Boncy J, Patel R, Burnett E, Juin S, Parashar UD, Bowen MD. Diversity of rotavirus strains circulating in Haiti before and after introduction of Rotarix vaccine. In preparation 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Leshem E, Lopman B, Glass R, Gentsch J, Banyai K, Parashar U, et al. Distribution of rotavirus strains and strain-specific effectiveness of the rotavirus vaccine after its introduction: a systematic review and meta-analysis. Lancet Infect Dis. 2014;14:847–56. [DOI] [PubMed] [Google Scholar]
- [18].Juin S, Schaad N, Lafontant D, Joseph GA, Barzilay E, Boncy J, et al. Strengthening National Disease Surveillance and Response-Haiti, 2010–2015. Am J Trop Med Hyg. 2017;97:12–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Steenland MW, Joseph GA, Lucien MA, Freeman N, Hast M, Nygren BL, et al. Laboratory-confirmed cholera and rotavirus among patients with acute diarrhea in four hospitals in Haiti, 2012–2013. Am J Trop Med Hyg. 2013;89:641–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Velasquez DE, Parashar U, Jiang B. Decreased performance of live attenuated, oral rotavirus vaccines in low-income settings: causes and contributing factors. Expert Rev Vaccines. 2018;17:145–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.