Mediation Analysis to Identify Individual Component Effects of the Care Bundle to Improve Functional Outcome After ICH in the INTERACT3 Study

BACKGROUND AND OBJECTIVES: The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT3) showed that the implementation of a care bundle improves the functional outcome of patients who experience acute intracerebral hemorrhage (ICH). However, uncertainty exists over the relative contribution of each component of the care bundle for the benefit. METHODS: INTERACT3 used a stepped-wedge, cluster randomized controlled trial design, which was conducted in 5 lower and 4 upper middle-income countries and 1 high-income country. Compared with usual care, there was a statistically significant beneficial effect of the intervention, a care bundle comprising early control of elevated blood pressure (BP), glucose, and temperature and reversal of warfarin-related anticoagulation, on the primary outcome of functional recovery of ICH. We performed a model-based causal mediation analysis to assess the contribution of each component of the care bundle to the overall effect, as measured by the modified Rankin Scale (mRS) at 6 months after randomization. The mediation analysis considered whether protocol-specified treatment targets were reached as well as the actual achieved levels of physiologic control according to the summary measures of systolic BP (mean, variation over 1-24 hours, and reduction in 1 hour) and mean of blood glucose, body temperature, and international normalized ratio over 24 hours. The analyses were performed in the modified intention-to-treat population with available mRS data. RESULTS: A total of 6,225 patients (mean age 61.9 years [SD 12.6], 2,284 women [36.5%]) with available primary outcome data were included in these analyses. Overall, only the control of BP and blood glucose contributed positively to the beneficial effect, with mediated proportions of 8.9% (95% CI 4.8-20.0) and 7.0% (1.1-17.0) for achieved systolic BP and blood glucose over 24 hours after randomization, respectively, and 4.0% (1.2-14.0) and 7.6% (2.2-15.0) for reaching the specified targets for systolic BP and blood glucose, respectively. DISCUSSION: The major contributors to the effectiveness of the care bundle in INTERACT3 for improved functional outcomes after acute ICH were the control of systolic BP and blood glucose, as indicated by both the achieved levels and protocol targets being met. TRIAL REGISTRATION INFORMATION: The name of the registry is "the Third, Intensive Care Bundle With Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT3)." This trial was submitted for registration at ClinicalTrials.gov (NCT03209258) on July 1, 2017, and the Chinese Clinical Trial Registry (ChiCTR-IOC-17011787) on June 28, 2017. The first patient was enrolled on December 12, 2017.