Rotavirus Vaccine Effectiveness Against Severe Acute Gastroenteritis: 2009-2022

Abstract

Background: Rotavirus was the leading cause of acute gastroenteritis among US children until vaccine introduction in 2006, after which, substantial declines in severe rotavirus disease occurred. We evaluated rotavirus vaccine effectiveness (VE) over 13 years (2009-2022).

Methods: We analyzed data from the New Vaccine Surveillance Network using a test-negative case-control design to estimate rotavirus VE against laboratory-confirmed rotavirus infections among children seeking care for acute gastroenteritis (≥3 diarrhea or ≥1 vomiting episodes within 24 hours) in the emergency department (ED) or hospital. Case-patients and control-patients were children whose stool specimens tested rotavirus positive or negative, respectively, by enzyme immunoassay or polymerase chain reaction assays. VE was calculated as (1-adjusted odds ratio)×100%. Adjusted odds ratios were calculated by multivariable unconditional logistic regression.

Results: Among 16 188 enrolled children age 8 to 59 months, 1720 (11%) tested positive for rotavirus. Case-patients were less often vaccinated against rotavirus than control-patients (62% versus 88%). VE for receiving ≥1 dose against rotavirus-associated ED visits or hospitalization was 78% (95% confidence interval [CI] 75%-80%). Stratifying by a modified Vesikari Severity Score, VE was 59% (95% CI 49%-67%), 80% (95% CI 77%-83%), and 94% (95% CI 90%-97%) against mild, moderately severe, and very severe disease, respectively. Rotavirus vaccines conferred protection against common circulating genotypes (G1P[8], G2P[4], G3P[8], G9P[8], and G12[P8]). VE was higher in children <3 years (73% to 88%); protection decreased as age increased.

Conclusions: Rotavirus vaccines remain highly effective in preventing ED visits and hospitalizations in US children.

FIGURE 1

Enrollment periods (surveillance years) included, NVSN, November 1, 2009 to November 30, 2022. Each surveillance year straddled 2 calendar years, and we refer to each of these surveillance years by the calendar year that they end. The differences in duration of enrollment were because of modifications to surveillance grant support during the study period.

FIGURE 2

New Vaccine Surveillance Network sites (November 1, 2009–November 30, 2022) and surveillance years included per site. Cincinnati, OH: Cincinnati Children’s Hospital Medical Center (November 1, 2009–November 30, 2022); Houston, TX: Texas Children’s Hospital (November 1, 2009–November 30, 2022); Kansas City, MO: Children’s Mercy Hospital (November 1, 2009–November 30, 2022); Nashville, TN: Monroe Carrell Jr Children’s Hospital at Vanderbilt (November 1, 2009–November 30, 2022); Oakland, CA: University of California at San Francisco Benioff Children’s Hospital Oakland (November 1, 2010–June 30, 2011; December 1, 2012–November 30, 2016); Pittsburgh, PA: University of Pittsburg Medical Center Children’s Hospital of Pittsburgh (December 1, 2016–November 30, 2022); Rochester, NY: UR-Golisano Children’s Hospital at the University of Rochester Medical Center (November 1, 2009–November 30, 2022); Seattle, WA: Seattle Children’s Hospital (November 1, 2009–November 30, 2022).

FIGURE 3

Case status determination based on site plus CDC rotavirus test results. (A) During November 1, 2009 to June 30, 2011, we used EIA (Premier Rotaclone; Meridian Bioscience, Inc) results from testing conducted at each NVSN site. (B) During December 1, 2011 to November 30, 2018, sites tested specimens using EIA, and all site EIA-positive specimens were genotyped at CDC using RT-PCR assays and nucleotide sequencing. Specimens that could not be genotyped were retested by real-time qRT-PCR to confirm positivity. Specimens that tested negative by quantitative qRT-PCR were tested further by EIA at CDC and classified as rotavirus positive or negative based on that result. (C) During December 1, 2018 to November 30, 2019, sites transitioned to qRT-PCR or multipathogen panel PCR assays (BioFire FilmArray Gastrointestinal Panel [BioFire Diagnostics, LLC, Salt Lake City, UT] or Luminex Gastrointestinal Pathogen Panel [Luminex Corp, Austin, TX, USA]). Sites’ positive specimens were tested at CDC using qRT-PCR to standardize the testing method and assay across sites. qRT-PCR including the multipathogen panel PCR assays are more sensitive and, thus, lead to more rotavirus detection than EIA. To ensure that the AGE episode was attributable to rotavirus and thus make the qRT-PCR and EIA results comparable, we used a C_t value cutoff of <27.0 to define rotavirus positive specimens based on findings from data collected at NVSN sites and other settings. (D) During December 1, 2019 to November 30, 2022, all sites used the same qRT-PCR assay, and we applied the C_t < 27.0 to define positive specimens. EIA has traditionally been used to detect rotavirus antigens in stool samples and EIA-positive results are correlated with clinical disease. The estimated sensitivity and specificity of the commercial EIA kit (Premier Rotaclone; Meridian Bioscience, Inc) used for this surveillance was 77% and 100%, respectively, based on stool samples that underwent genotype testing. Molecular assays are more sensitive than EIA. RT-PCR assays used for genotyping have been found to increase rotavirus detection by 15% to 27% compared with EIA. Real-time qRT-PCR assay was 99% sensitive and 77% specific compared with EIA and 100% sensitive and 86% specific compared with RT-PCR. For the multipathogen panel PCR assays, sensitivity ranged from 90% to 100% and specificity ranged from 96% to 100%. From November 1, 2009 to November 30, 2018, EIA served as the basis of our case- and control-patient definition with an additional layer of confirmation via genotyping in most years. With the increasing availability and use of PCR assays, NVSN sites transitioned to primarily using qRT-PCR on December 1, 2018. Although PCR assays are more likely to detect rotavirus, the causative pathogen of AGE symptoms in PCR-positive children may not always be rotavirus.^– To make the qRT-PCR and EIA results comparable, we used a Ct value cutoff of <27.0 to define rotavirus-positive specimens based on findings from data collected at NVSN sites and other settings.^–, b Participants were classified as rotavirus-positive case-patients or rotavirus-negative control-patients if their final rotavirus test result was positive or negative, respectively.

FIGURE 4

Application of eligibility criteria and enrollment to define study population, NVSN, November 1, 2009 to November 30, 2022. The figure represents rotavirus prevalence among study participants by surveillance years and vaccination coverage for having received at least 1 rotavirus vaccine dose in the overall study population and by case status. ^a Participants born ≥ April 1, 2006 who were hospitalized or received care at an ED for acute gastroenteritis at an NVSN site from November 1, 2009 to November 30, 2022 with complete health insurance information and a stool specimen tested for rotavirus. ^b Children age 8 months to 17 years at enrollment with confirmed rotavirus vaccination status and rotavirus test result. ^c Participants who tested rotavirus-positive and rotavirus-negative were classified as rotavirus-positive case-patients or rotavirus-negative control-patients, respectively. ^d The 2 recommended rotavirus vaccines in the United States require different dose schedules to complete a course (3 doses or 2 doses) by the age of 8 months. To account for this, we examined the effectiveness of receiving at least 1 dose of any rotavirus vaccine. Among those who received at least 1 vaccine dose, ≥80% completed a course across case-control patient status and age group.

FIGURE 5

Rotavirus prevalence by time and vaccination coverage time and cases status, NVSN, 2010 to 2022. (A) Represents rotavirus prevalence among study participants by surveillance years. (B) Represents vaccination coverage for having received at least 1 rotavirus vaccine dose in the overall study population and by case status. Each surveillance year (enrollment periods) straddled 2 calendar years, and we referred to each of these surveillance years by the calendar year that they end. For example, we refer to the surveillance year that started in November 2009 and ended in June 2010 as surveillance year 2010. The 2010 and 2011 surveillance years started on November 1 of the first calendar year and ended on June 30 of the second calendar year. The 2012 to 2022 surveillance years started on December 1 of the first calendar year and ended on November 30 of the second calendar year. The differences in duration of the surveillance years were because of modifications of surveillance grant support during the study period.

FIGURE 6

Effectiveness of receiving at least 1 dose of rotavirus vaccine against rotavirus-associated ED visits or hospitalizations by final level of clinical care, disease severity and commonly detected genotypes among children aged <5 years, NVSN, 2010 to 2022. ^a Models are adjusted for month and year of birth, quarter and year of symptom onset, and surveillance site. ^b Components of MVSS include the maximum number of diarrheal stools and vomiting episodes in 24 hours, duration of diarrhea and vomiting in days, maximum recorded fever, reception of treatment (oral or intravenous rehydration in ED or hospitalization) and levels of dehydration based on symptom information (eg, lethargy, sunken eyes, and ability to drink properly) according to the World Health Organization’s Integrated Management of Childhood Illnesses dehydration criteria. Categories based on the total MVSS. Mild: 0 to 10; moderately severe: 11 to 15; very severe: 16 to 20. ^c Genotype results were only available during 2012 to 2019 surveillance years.

FIGURE 7

Distribution of common genotypes detected among children age <5 years by surveillance year, NVSN, 2012 to 2019. Each surveillance year (enrollment periods) straddled 2 calendar years, and we referred to each of these surveillance years by the calendar year that they end. For example, we refer to the surveillance year that started in December 2012 and ended in November 2013 as surveillance year 2013. The 2012 to 2019 surveillance years started on December 1 of the first calendar year and ended on November 30 of the second calendar year.

FIGURE 8

Rotavirus vaccine effectiveness against rotavirus-associated hospitalization or ED visit by age among children aged 8 months to 17 years, NVSN, 2010 to 2022. NVSN, New Vaccine Surveillance Network; VE, vaccine effectiveness. ^a Models are adjusted for year of birth, quarter and year of symptom onset, and surveillance site. ^b We could not estimate the complete course of mixed RV5 and RV1 dose VE for the 60 to 71 months, 72 to 83 months, and ≥84 months age groups because of small sample sizes.