A study to identify risk factors that modify Rotarix efficacy and influence the time to rotavirus gastroenteritis among vaccinated and unvaccinated African infants.

Proposal
1088

Title of Proposed Research

Lead Researcher

Joann Gruber

Affiliation

University of North Carolina at Chapel Hill

Funding Source

None

Potential Conflicts of Interest

Research Assistant at Merck (Jan-Dec 2014) working on analysis related to Merck's rotavirus vaccine (RotaTeq)

Data Sharing Agreement Date

04 September 2015

Lay Summary

Rotavirus infections are the world's leading cause of severe diarrhea in children under five years of age [1, 2]. Each year there are 140 million cases of rotavirus infection in children under five with 26 million of these cases requiring clinic visits or hospitalizations [1]. Despite similar rates of infection in developed and developing countries, rotavirus-associated mortality disproportionately affects children in developing countries, where 80-90% of rotavirus-associated deaths occur [1, 6].

Vaccination against rotavirus is considered the most promising intervention to prevent infection and reduce the burden of disease [4, 11, 12]. As of 2009, universal rotavirus vaccination of infants was recommended by the World Health Organization [13]. Two live, oral rotavirus vaccines are available for use across the globe—Rotarix and RotaTeq. Doses are administered two or three times, depending on the vaccine type, around 2, 4, and 6 months of age [11]. While rotavirus vaccine trials have shown high protective efficacy (85-96%) against severe rotavirus gastroenteritis in high income countries [14-16], the efficacy has been much lower (40-59%) in trials conducted in GAVI-eligible countries in Africa and Asia [17-19]. This lower vaccine efficacy reported in developing countries is similar to study results for other live oral vaccines [20-22]. Although this is a common theme in developing countries, the reason for lower efficacy is not well understood.

I purpose to analyze risk factors that modify Rotarix vaccine efficacy and influence time to rotavirus gastroenteritis among African infants. I am a PhD student in epidemiology at the University of North Carolina and propose to do this work as my dissertation. The study participants will be 2,089 infants who participated in a clinical trial of Rotarix sponsored by GlaxoSmithKline. For the first research aim, I will utilize the randomized design of the clinical trial to identify risk factors that modify the effect of Rotarix in developing countries. The second research aim will be time-to-event analysis. Each risk factor will be analyzed as a potential exposure variable influencing the time to occurrence of breakthrough case of rotavirus (any severity and severe). Aim 3 will use the results of Aim 1 and Aim 2 to predict the effectiveness of Rotarix in developing countries. Potential risk factors analyzed will depend on the data collected during the study; however, I propose the following as a guide. Risk factors will be age at vaccination, breastfeeding status, nutrition status, birth weight, gestational age at birth, and co-administration of vaccines and medicines at time of vaccination. Two peer reviewed publication will be submitted summarizing the results. Understanding how clinical and demographic risk factors influence rotavirus vaccine efficacy can help to improve efficacy in developing countries.

Study Data Provided

[{ "PostingID": 1296, "Title": "GSK-102248", "Description": "Multi-Center Study to Assess the Efficacy, Safety and Immunogenicity of 2 or 3 Doses of GSK Biologicals' Oral Live Attenuated Human Rotavirus (HRV) Vaccine Given Concomitantly With Routine EPI Vaccinations in Healthy Infants" },{ "PostingID": 1297, "Title": "GSK-111274", "Description": "Multi-Center Study to Assess the Efficacy, Safety and Immunogenicity of 2 or 3 Doses of GSK Biologicals' Oral Live Attenuated Human Rotavirus (HRV) Vaccine Given Concomitantly With Routine EPI Vaccinations in Healthy Infants" }]

Statistical Analysis Plan

Aim 1: To identify demographic and clinical characteristics that modify the effect of Rotarix among infants in Malawi and South Africa. I will determine if there is a significant interaction (a = 0.1) between treatment type (Rotarix versus placebo) and baseline infant characteristics on the effect of vaccine. The effect of the vaccine will be analyzed for all rotavirus gastroenteritis cases and severe rotavirus gastroenteritis cases. Modifiers assessed will depend on the data available; proposed modifiers to be assessed are age at vaccination, breastfeeding status, nutrition status, birth weight, gestational age at birth, and co-administration of vaccines and medicines at time of vaccination. Aim 1 will utilize the randomized design of the clinical trial and will be completed by conducting Poisson regression. A Wald test will be used to determine if there is a significant interaction between vaccine type (Rotarix versus placebo) and baseline infant characteristics on the additive and multiplicative scales, rate differences and rate ratios, respectively. Vaccine efficacy will be calculated as 1 - IRR where IRR is the rate ratio comparing the rate of rotavirus gastroenteritis among those given Rotarix to the rate among those given placebo. Stratum specific estimates will be calculated for risk factors that modify the effect of the vaccine. I will compare the absolute (additive scale) and relative (multiplicative scale) population effects. I will also analyze the immunogenicity data to determine if the statistical interactions may be signaling a true biological effect.Aim 2: To identify demographic and clinical characteristics associated with time to breakthrough cases of rotavirus gastroenteritis after Rotarix vaccination. A breakthrough case of rotavirus gastroenteritis was defined as a case occuring 14-days after an infant had received a full vaccine series (2 Rotarix/placebo doses). I will first analyze the median time to the occurrence of breakthrough cases of rotavirus (any severity and severe) comparing the vaccinated to unvaccinated group. Next, I will conduct a cohort study of only vaccinated infants in the clinical trial. Each risk factor will be analyzed as a potential exposure variable influencing the time to occurrence of breakthrough case of rotavirus. I will conduct a crude and adjusted analysis for each variable using an accelerated failure time model to calculate the time ratios comparing levels of each risk factor. I will report the median time ratio of time to breakthrough case of rotavirus. I will also create weighted Kaplan-Meier cumulative incidence curves. In addition, I will compare the time-to-event results in the treatment (vaccine) group to the results from the placebo group to better understand how the vaccine works in these populations. Risk factors assessed will depend on the data available; proposed risk factors to be assessed are age at vaccination, breastfeeding status, nutrition status, birth weight, gestational age at birth, and co-administration of vaccines and medicines at time of vaccinationAim 3: To determine the predicted effectiveness of Rotarix when introduced into routine vaccination schedules in developing countries in Africa. Aim 3 will use the results of Aim 1 and Aim 2 to predict the effectiveness of Rotarix in developing countries. I will use baseline data of the risk factors analyzed to determine how the vaccine would be expected to perform in selected countries based on the results of these analyses. I will also utilize the parametric g-formula to predict vaccine efficacy under potential exposure contrasts. The parametric g-formula is a type of marginal structure model. These types of models are used to replicate the finding of a randomized trial in observational data. Although these data are randomized for vaccine treatment, I will use the parametric g-formula to calculate the effects of joint randomization (treatment and risk factor) to estimate the efficacy under different exposure scenarios.All analyses will be conducted using SAS software (version 9.3 for Windows, SAS Institute Inc., Cary, NC, USA). My committee will advise me throughout the data analysis process. The methodological approaches to each aim may evolve as my committee advises. My tentative committee will be Dr. Michele Jonsson-Funk (chair and pharmacoepidemiologist); Dr. Sylvia Becker-Dreps (clinician and rotavirus expert); Dr. Michael Hudgens (biostatician and vaccine expert); Dr. Daniel Westreich (infectious disease epidemiologist); and Dr. David Weber (vaccine expert/infectious disease epidemiologist).This committee is subject to the availability of the persons listed, but will include Dr. Jonsson-Funk, Dr. Becker-Dreps, and Dr. Hudgens.

Publication Citation

Gruber JF, Becker-Dreps S, Hudgens MG, Brookhart MA, Thomas JC, Jonsson Funk
M. Timing and predictors of severe rotavirus gastroenteritis among
unvaccinated infants in low- and middle-income countries. Epidemiol Infect
2018;146(6):698-704.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5895523/

Gruber JF, Becker-Dreps S, Hudgens MG, Brookhart MA, Thomas JC, Jonsson Funk
M. Timing of Rotavirus Vaccine Doses and Severe Rotavirus Gastroenteritis
among Vaccinated Infants in Low- and Middle-Income Countries. Epidemiology
2018.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6167161/