Potential Causes and Biases Explaining Heterogeneity in Rotavirus Vaccine Effect Estimates Restricted; Files Only

Amin, Avnika B. (Fall 2021)

Permanent URL: https://etd.library.emory.edu/concern/etds/fq977w00c?locale=en
Published

Abstract

Of the estimated 128,500 rotavirus-associated deaths in 2016, 82% occurred in sub-Saharan Africa. Most rotavirus gastroenteritis morbidity and mortality occurs in high child mortality settings, where rotavirus vaccines appear 20-40% less effective. In some post-introduction settings, vaccine effects appear time-varying. We examined population-level factors that could explain geographic and temporal rotavirus vaccine effect heterogeneity.

Genotypic diversity may explain some heterogeneity if vaccine-induced protection is lower against genotypes more common in high child mortality settings. In Aim 1, we pooled individual-level data from clinical trials conducted in South America, Europe, Africa, and the Western Pacific. Using multinomial logistic regression, efficacy against severe rotavirus gastroenteritis was lower against the G2P[4] genotype (VE: 71%; 95% CI: 43-85%) than against the G1P[8] genotype (VE: 96%; 95% CI: 89-98%) homotypic with the vaccine formulation.

Differences in the baseline force of infection may also influence heterogeneity in vaccine effects. In Aim 2, we pooled individual-level data from clinical trials in 23 countries and estimated efficacy using Poisson regression with and without adjustment for country-specific forces of infection. Unadjusted country-specific estimates of efficacy against severe rotavirus gastroenteritis ranged from 10% (95% CI: <0-94%) to 100% (95% CI: <0-100%). Adjustment attenuated estimate heterogeneity (lowest VE: 73% (95% CI: 61-81%); highest VE: 85% (95% CI: 78-90%)).

A time-varying force of infection may underlie observed post-introduction temporal fluctuations in rotavirus vaccine effectiveness. In Aim 3a, we fit logistic regression models to six years of post-introduction U.S. rotavirus surveillance data. Effectiveness against moderate-to-severe rotavirus gastroenteritis biennially fluctuated between 70% (95% CI: 45-84%) and 81% (95% CI: 72-87%). Fluctuations attenuated after adjustment for the annual force of infection (lowest VE: 75% (95% CI: 61-85%); highest VE: 77% (95% CI: 70-82%).

Time-varying force of infection may also drive time-varying diagnostic misclassification bias. In Aim 3b, we bias-adjusted effectiveness estimates from Aim 3a. Bias adjusted generally improved effectiveness estimates, but also eliminated lingering fluctuations in force-of-infection-adjusted effectiveness estimates.

Genotypic diversity, force of infection, and diagnostic misclassification each explain portions of observed heterogeneity in rotavirus vaccine effects. Rotavirus vaccine evaluations should incorporate these factors for valid vaccine effect estimation and comparisons across settings and years.

Table of Contents

1 - Background... 1

1.1 - Diarrheal disease burden... 1

1.2 - Rotavirus... 1

1.3 - Immunity... 3

1.4 - Rotavirus vaccines and their performance... 4

2 - Motivating rationale and specific aims... 8

2.1 - Genotype Diversity... 8

2.2 - Force of Infection... 10

2.3 - Diagnostic Misclassification... 12

2.4 - Specific Aims... 13

3 - Data sources... 15

3.1 - Rotarix Clinical Trials... 15

3.2 - Global Age Distribution of Rotavirus Hospitalizations... 19

3.3 - New Vaccine Surveillance Network Data... 20

3.4 - National Respiratory and Enteric Virus Surveillance System... 22

4 - Aim 1: Genotype-specific Vaccine Efficacy... 23

4.1 - Abstract... 23

4.2 - Main Manuscript... 24

4.3 - Supplemental Material... 36

5 - Aim 2: Local Rotavirus Incidence... 47

5.1 - Abstract... 47

5.2 - Main Manuscript... 48

5.3 - Supplemental Material... 71

6 - Aim 3: Time-Varying Rotavirus Incidence and Diagnostic Misclassification... 81

6.1 - Abstract... 81

6.2 - Main Manuscript... 82

6.3 - Supplemental Material... 97

7 - Public Health Implications... 109

7.1 - Aim 1: Genotypic Differences... 109

7.2 - Aim 2: Local Incidence Rates... 111

7.3 - Aim 3: Time-Varying Incidence and Misclassification... 113

7.4 - Broader Implications... 115

8 - References... 118

About this Dissertation

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Keyword
Committee Chair / Thesis Advisor
Committee Members
Last modified Preview image embargoed

Primary PDF

Supplemental Files