Monovalent rotavirus vaccine effectiveness and long-term impact among children <5 years old in Antananarivo, Madagascar, 2010-2022

Abstract

Background: Monovalent rotavirus vaccine substantially reduced rotavirus disease burden after introduction (May 2014) in Madagascar. We examined the effectiveness and long-term impact on acute watery diarrhea and rotavirus-related hospitalizations among children <5 years old at two hospitals in Antananarivo, Madagascar (2010-2022).

Methods: We used a test-negative case-control design to estimate monovalent rotavirus vaccine effectiveness (VE) against laboratory-confirmed rotavirus hospitalizations among children age 6-23 months with documented vaccination status adjusted for year of symptom onset, rotavirus season, age group, nutritional status, and clinical severity. To evaluate the impact, we expanded to children age 0-59 months with acute watery diarrhea. First, we used admission logbook data to compare the proportion of all hospitalizations attributed to diarrhea in the pre-vaccine (January 2010-December 2013), transition period (January 2014-December 2014), and post-vaccine (January 2015-December 2022) periods. Second, we used active surveillance data (June 2013-May 2022) to describe rotavirus positivity and detected genotypes by vaccine introduction period and surveillance year (1 June-31 May).

Result: Adjusted VE of at least one dose against hospitalization due to rotavirus diarrhea among children age 6-23 months was 61 % (95 % CI: -39 %-89 %). The annual median proportion of hospitalizations attributed to diarrhea declined from 28 % in the pre-vaccine to 10 % in the post-vaccine period. Rotavirus positivity among hospitalized children age 0-59 months with acute watery diarrhea was substantially higher during the pre-vaccine (59 %) than the post-vaccine (23 %) period. In the pre-vaccine period, G3P[8] (76 %) and G2P[4] (12 %) were the dominant genotypes detected. Although genotypes varied by surveillance year, G1P[8] and G2P[4] represented >50 % of the genotypes detected post-introduction.

Conclusions: Rotavirus vaccine has been successfully implemented in Madagascar's routine childhood immunization program and had a large impact on rotavirus disease burden, supporting continued use of rotavirus vaccines in Madagascar.

Keywords: Genotype; Madagascar; Rotavirus; Rotavirus surveillance; Rotavirus vaccine effectiveness; Rotavirus vaccine impact.

Fig. 1.

Vaccine effectiveness, vaccine impact, and genotype distribution analyses flowchart using active rotavirus surveillance data, Madagascar, 1 June 2013–31 December 2022. Abbreviations: G, glycoprotein (VP7); P, protease-cleaved protein (VP4). ^a We restricted the vaccine effectiveness (VE) analysis children enrolled at the Centre Hospitalier Universitaire Mère Enfant Tsaralalàna (CHUMET) and Centre Hospitalier Universitaire Joseph Raseta Befelatanana (CHUJRB) from 1 September 2018 (when sites began collecting written vaccine records) to 31 December 2022. ^b We restricted the VE analysis further to children age 6–23 month because rotavirus burden was highest in this age group. ^c We included children who were age-eligible to receive a rotavirus vaccine (born ≥1 May 2014) with verified vaccination status via a copy of their immunization card or registry or those who were reported to never have received any vaccines. We classified children as vaccinated if a dose of rotavirus vaccine was administered at least 14 days before symptom onset. ^d We restricted the vaccine impact analysis to data collected at CHUMET, where active surveillance data were available from pre-vaccine and post-vaccine periods. ^e Pre-vaccine introduction period: 1 June 2013–31 May 2014 period; post-vaccine introduction period: 1 June 2014–31 May 2022. ^f The genotype distribution analysis was restricted to rotavirus positive stool specimens with G or P strain detected.

Fig. 2.

Monovalent rotavirus vaccine effectiveness (at least one dose) for the prevention of rotavirus-associated hospitalizations among children age 6–23 months, 1 September 2018–31 December 2022. Abbreviations: AGE, acute gastroenteritis; VE, vaccine effectiveness. ^a VE adjusted for age group (6–11 and 12–23 months), year of admission, rotavirus season (April–July and August–March), Vesikari score (<11 and ≥11) and chronic-malnutrition status (stunting and normal height for age). If the models stratified by one of the adjustment variables, then this variable was excluded from the model. For example, in the age stratified model, we included hospital admission year, rotavirus season, Vesikari score, and chronic-malnutrition status as adjustment variables. ^b Chronic malnutrition was based on stunting defined as >2 standard deviations below the median height for age. ^c Clinical severity classification was based on total modified Vesikari Severity Score (MVSS): mild disease (0–10) and moderately-very severe disease (11–20).

Fig. 3.

Three-month rolling average of the number of all-cause and diarrhea-associated hospitalizations at CHUMET among children age 0–59 months, 1 January 2010–31 December 2022. Abbreviations: CHUMET, Centre Hospitalier Universitaire Mère Enfant Tsaralalàna.

Fig. 4.

Distribution of rotavirus positivity among children age 0–59 months enrolled in active surveillance stratified by vaccine introduction period, CHUMET, 1 June 2013–31 May 2022.^a Abbreviations: CHUMET, Centre Hospitalier Universitaire Mère Enfant Tsaralalàna. ^a Pre-vaccine introduction period: 1 June 2013–31 May 2014 period; post-vaccine introduction period: 1 June 2014–31 May 2022. Surveillance year: each year ranged 12 months starting on 1 June of an earlier calendar year and ending on 31 May of the following calendar year later year. For example, surveillance year 2013 ranged from 1 June 2013 through 31 May 2014, and surveillance year 2021 ranged from 1 June 2021 through 31 May 2022.

Fig. 5.

Age distribution of children age 0–59 months hospitalized with all-cause diarrhea based on logbook data (Panel A) and children enrolled in active surveillance and tested positive for rotavirus (Panel B) stratified by vaccine introduction period, CHUMET, 1 June 2013–31 May 2022^a. Abbreviations: CHUMET, Centre Hospitalier Universitaire Mère Enfant Tsaralalàna. ^a Pre-vaccine introduction period: 1 January 2010–31 May 2014 period; post-vaccine introduction period: 1 June 2014–31 May 2022.

Fig. 6.

Distribution of common rotavirus genotypes among children age 0–59 months enrolled in active surveillance by vaccine introduction period and surveillance year, CHUMET and CHUJRB, 1 June 2013–31 May 2022^a. Abbreviations: CHUMET, Centre Hospitalier Universitaire Mère Enfant Tsaralalàna. ^a Pre-vaccine introduction period: 1 June 2013–31 May 2014 period; post-vaccine introduction period: 1 June 2014–31 May 2022. Surveillance year: each year ranged 12 months starting on 1 June of an earlier calendar year and ending on 31 May of the following calendar year later year. For example, surveillance year 2013 ranged from 1 June 2013 through 31 May 2014, and surveillance year 2021 ranged from 1 June 2021 through 31 May 2022.