DOI: 10.1161/circ.148.suppl_1.12210 ISSN: 0009-7322

Abstract 12210: Repeated Systolic Blood Pressure Measures During Intensive Treatment Show a Loss Of Protective Effect Over Time. A Secondary Analysis of the SPRINT Trial

Lyda M Herrera-Sanchez, Oscar L Rueda-Ochoa
  • Physiology (medical)
  • Cardiology and Cardiovascular Medicine

Introduction: The SPRINT trial showed a 25% reduction in the primary outcome with SBP intensive treatment using Cox proportional risk analysis. A recent secondary analysis by the SPRINT group assessed the post-trial maintenance of this protection by considering repeated SBP measurements and using a mixed linear regression model. They found that the intervention lost its protective effect post-trial, possibly due to the loss of intensive blood pressure control. Could this loss of protective effect also be found during the clinical trial, when considering repeated measurements of systolic blood pressure in the analysis?

Methods: Secondary analysis of the SPRINT study database included 9,068 patients with repeated measures of SBP and complete baseline information. Cox proportional hazards and mixed linear models were used to analyze the primary endpoint and repeated SBP measures, respectively. A Cumulative Joint Model Analysis was conducted to combine the direct and indirect effects of the intervention on the primary outcome. R Studio V12.0 software was used for the analyses.

Results: The protective effect is lost during the trial, both in the total population (3.4 years) as well as in subgroups such as women (1.1 years), patients with CVD (1.3 years), CKD (1.3 years), black patients (1.8 years), younger than 75 years (2.1 years) and in patients who present serious adverse events (3.4 years) compared to those who do not present SAEs (4.2 years). Beyond the loss of blood pressure control, other factors such as cumulative BP, adverse events, and the increase in SBP variability could be considered.

Conclusions: Including repeated measures of SBP revealed a loss of protective effect in the intervention during the trial and generated Simpson's paradox. This highlights the need to consider alternative statistical methods in clinical trial analysis beyond Cox proportional hazard analysis for a comprehensive perspective on intervention efficacy.

More from our Archive