Inter-arm differences in blood pressure and mortality: individual patient data meta-analysis and development of a prognostic algorithm (Iinterpress-IPD collaboration)
Résumé
Objective: Study-level meta-analyses have associated inter-arm differences (IAD) in systolic blood pressure (SBP) with increased mortality. However key areas of investigation remain, such as determining the additional contribution of IAD to prognostic risk estimation for cardiovascular and all-cause mortality, and determining the minimum value for IAD that defines elevated risk. We are conducting an individual participant data (IPD) meta-analysis to address these questions.
Design and method: Systematic review and IPD meta-analysis; we identified prospective studies that measured SBP in both arms during recruitment, and invited contribution of eligible datasets to the Collaboration. Study datasets were cleaned and combined into a single dataset for analyses. A non-random sample of four datasets were reserved for model validation; remaining data were analysed in fixed effect 1-stage meta-analyses with multivariable time-to-event regression modelling. Analyses of IAD dichotomised using cut-offs by 1mmHg increments from 0 to 20 mmHg, adjusted for age, gender and baseline SBP, were compared to identify a lower limit of IAD associated with increased all-cause mortality using random effects 2-stage models.
Scientific Data: Searches to January 2017 yielded 4448 unique citations, 152 full texts were screened, 60 potentially relevant datasets were identified and their authors contacted. Data from 24 studies (57,434 eligible individual adult patients) were received.
Results: In complete case analysis of 35,900 records (the derivation cohort), absolute systolic IAD was associated with increased all-cause mortality: fully adjusted hazard ratio (HR) 1.01, 95% confidence interval (95%CI) 1.00 to 1.02 per mmHg of IAD. Other significant model variables were age, gender, baseline SBP, current smoking, total cholesterol, ethnicity, and diagnoses of hypertension or diabetes. Incremental analyses of 50,661 records showed increasing HRs associated with rising IAD cut-offs, and suggested that an IAD of 7 mmHg or more is associated with increased risk of all-cause mortality.
Conclusions: This IPD meta-analysis confirms the role of systolic IAD as an independent risk marker for all-cause mortality, with a threshold of 7 mmHg as a lower limit for increased mortality risk. We continue to model cardiovascular outcomes and will present full results to the meeting.