Treatment Effect Heterogeneity in Acute Kidney Injury Incidence Following Intravenous Antihypertensive Administration for Severe Blood Pressure Elevation During Hospitalization

RATIONALE & OBJECTIVE: Severe hypertension (HTN) that develops after hospital admission is prevalent in 10% of patients admitted for reasons other than HTN. Severe HTN is commonly treated with intravenous (IV) antihypertensives and is associated with a greater risk of acute kidney injury (AKI). We explored whether there is heterogeneity in IV antihypertensives' effect on AKI incidence among patients who develop severe HTN during hospitalization. STUDY DESIGN: Heterogeneity of treatment effect analysis. SETTINGS & PARTICIPANTS: Patients who developed severe HTN, defined as systolic blood pressure (BP)>180 or diastolic BP>110mm Hg, during hospitalization and did not have kidney failure. EXPOSURE: Treatment with IV antihypertensives within 3 hours of BP elevation. OUTCOME: Time to developing AKI. ANALYTICAL APPROACH: An accelerated failure time Bayesian additive regression trees (BART) model to capture the association between the time to develop AKI and predictors. Individual treatment effects were estimated for each participant using a counterfactual outcome framework, and these estimates were used to identify patient characteristics associated with treatment effect heterogeneity in response to IV antihypertensives. LIMITATIONS: Data-driven, hypothesis-generating approach. Findings were not validated with external data sources. CONCLUSIONS: These exploratory findings suggest that most patients who develop severe HTN will not benefit from IV antihypertensive treatment. Future studies should assess for heterogeneity when identifying treatment options, if any are needed, for severe HTN. PLAIN-LANGUAGE SUMMARY: Patients who develop severe blood pressure elevation during hospitalization are commonly treated with intravenous antihypertensives; however, this could lead to acute kidney injury (AKI). We wanted to assess whether this is consistent across all patients, using a new statistical approach that predicts what would happen if patients who were treated had not been treated and those who were not treated had been. We found that most patients will develop AKI, and only a small subset of patients might not. This exploratory study can help inform future studies on the treatment of hypertension that develops during hospitalization.