Safety and Immunogenicity of Measles Vaccination in HIV-Infected and HIV-Exposed Uninfected Children: A Systematic Review and Meta-Analysis

Background: HIV-infected and HIV-exposed uninfected (HEU) children have an increased risk of measles that may be due to altered immune responses or suboptimal timing of measles vaccination. We aimed to evaluate the safety and immunogenicity of measles vaccination in HIV-infected and HEU children. Methods: For this systematic review and meta-analysis, we searched PubMed, Embase, Cochrane Library, CINAHL, Global Health Library and IndMED on May 9, 2018. Studies were included if they reported on safety or seroresponse (either seroprotection/seropositivity/seroconversion) after measles vaccination in HIV-infected or HEU children. We calculated pooled estimates to compare immunogenicity outcomes between HIV-infected, HEU and HIV-unexposed children, using risk ratios [RRs] (with 95%CIs). PROSPERO registration number: CRD42017057411. Findings: Seventy-one studies met the inclusion criteria (15,363 children). Twenty-eight studies reported on safety; vaccine-associated adverse events and deaths were uncommon. Sixty-two studies reported on immunogenicity, 27 were included in the meta-analysis. HIV-infected children had lower seroresponse rates after primary vaccination compared with HIV-unexposed (RR 0.74; 95%CI: 0.61–0.90, I2 = 85.9%) and HEU children (0.78; 0.69–0.88, I2 = 77.1%), which was mitigated by antiretroviral therapy and time interval between vaccination and serology. HEU and HIV-unexposed children had similar seroresponses. Vaccination at 6-months resulted in similar proportions of HIV-infected children having seroresponse compared with HIV-unexposed (0.96; 0.77–1.19) and HEU children (1.00; 0.73–1.37, I2 = 63.7%). Interpretation: Primary measles vaccination at 6-months of age may provide protection against measles during early infancy in settings with high prevalence of maternal HIV-infection, however, further studies are needed to evaluate this strategy in HEU children and HIV-infected children receiving antiretroviral therapy. Funding: South African Research Chairs Initiative of the Department of Science and Technology and National Research Foundation in Vaccine Preventable Diseases; Medical Research Council: Respiratory and Meningeal Pathogens Research Unit. This systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [27]. We searched PubMed, Embase, Cochrane Library, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Global Health Library (including African Index Medicus, Latin American and Caribbean Health Sciences), and IndMED on 9 May 2018, for articles containing (“measles” and “vaccine”) and “HIV” (Supplementary data 1). Additional studies were identified by searching reference lists of the articles included in full-text screening and ClinicalTrials.gov. Studies were eligible for inclusion in the systematic review if they reported on immunogenicity or safety of any measles vaccination strategy in HIV-infected or HEU children aged 0–18 years. For inclusion in the immunogenicity meta-analysis, studies needed to report on primary or booster vaccination and had to include a comparator group of either HIV-uninfected children (HEU/HIV-unexposed) or HIV-infected children on a different antiretroviral therapy (ART) regimen. No restrictions regarding geographical region or year of publication were applied. Eligible study designs were interventional or observational. For assessment of safety, case reports were also included. Animal studies, systematic reviews, narrative reviews, reports of proceedings and publications not written in English, French, German, Spanish, Portuguese or Dutch were excluded. The outcomes of interest were immunogenicity and safety. Immunogenicity: studies were included if data were reported as proportions of subjects with seroprotective (≥ 330 mIU/mL or as indicated by authors), seropositive, or seroconversion (4-fold rise in titre or change from seronegative to seropositive) measles antibody responses. A composite outcome for seroresponse was created using seroprotection rates post-vaccination, and if not available, seropositivity or seroconversion rates were considered. Safety: all reported safety outcomes post-vaccination were considered, including deaths, severe adverse events (SAEs) other than death and adverse events (AEs). Two independent reviewers (EM, MvR) screened titles and abstracts of identified studies. Articles were retained if they met the inclusion criteria according to one or both of the reviewers. In case of duplicate publications of the same results, the most complete reference was included. Data were extracted from manuscripts using a standardised data extraction form (Supplementary data 2) and authors were contacted in case of missing data. Data of interest included: study design, study population, vaccine type, age at vaccination, time-period between vaccination and measurement of the serological response, number of vaccine doses administered, use of ART, outcome measures, laboratory methods used to detect measles antibodies, serological cut-off values, proportions with seroresponse, and number and type of (S)AEs. The Cochrane Risk of Bias Tool was adapted to enable evaluation of observational studies (Supplementary data 3) [28]. For five categories, risk of bias was assessed as low (= 0), unclear (= 1), or high (= 2). Studies with a high summative risk of bias score (≥ 7) were excluded from meta-analysis. When multiple time-points were reported for immune responses after the same vaccine dose, the time-point closest to vaccination was reported, except for two studies that had a smaller sample size at the earlier time-point [29], [30]. For the descriptive analyses, point estimates of the proportion of seroresponders for the individual studies under each group were calculated with 95% confidence intervals (CIs) assuming an exact binomial distribution. Three different primary meta-analyses compared serological responses in HIV-infected vs. HIV-unexposed, HIV-infected vs. HEU and HEU vs. HIV-unexposed children using risk ratios (RRs) and 95%CIs stratified by vaccination dose and age at vaccination. In case of significant heterogeneity (I2 > 50%), a random-effects model was applied. To explore statistical variation and heterogeneity between trials, pre-specified subgroup analyses were performed based on outcome (seroprotection), serological test, use of ART, study design, age at vaccination and time interval between vaccination and measurement of the serological response. Meta-regression was used to explore between-study variance not explained by the covariates and risk of publication bias was assessed using normal and contour-enhanced funnel plots if ten or more articles were included in the meta-analysis. Small study effects were evaluated using Egger’s-test for asymmetry. We used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system for rating overall quality of evidence [31]. All analyses were performed using Stata, version 13 (StataCorpLP, Texas, USA). The study was prospectively registered in PROSPERO (CRD42017057411) [32]. The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and all authors had final responsibility for the decision to submit for publication.