Associations Between Elevated Systolic Blood Pressure and Outcomes in Critically Ill Patients: A Retrospective Cohort Study and Propensity Analysis
An-Min Hu,Chao Hai,Hai-Bo Wang,Zhongheng Zhang,Ling-Bin Sun,Zhong-Jun Zhang,Hui-Ping Li
DOI: https://doi.org/10.1097/SHK.0000000000001774
2021-01-01
Shock
Abstract:Introduction: Studies have shown nonlinear relationships between systolic blood pressure (SBP) and outcomes, with increased risk observed at both low and high blood pressure levels. However, the relationships between cumulative times at different SBP levels and outcomes in critically ill patients remain unclear. We hypothesized that an appropriate SBP level is associated with a decrease in adverse outcomes after intensive care unit (ICU) admission. Methods: This study was a retrospective analysis of data from the Medical Information Mart for Intensive Care (MIMIC) III database, which includes more than 1,000,000 SBP records from 12,820 patients. Associations of cumulative times at four SBP ranges (<100 mm Hg, 100-120 mm Hg, 120-140 mm Hg, and >= 140 mm Hg) with mortality (12-, 3-, 1-month mortality and in-hospital mortality) were evaluated. Restricted cubic splines and multivariable Cox regression models were employed to assess associations between mortality and cumulative times at SBP levels (4 levels: <2, 2-12, 12-36, and >= 36 h) over 72 h of ICU admission. Additionally, 120 mm Hg to 140 mm Hg was subdivided into <12 h (Group L) and >= 12 h (Group M) subsets and subjected to propensity-score matching and subgroup analyses. Results: At 120 mm Hg to 140 mm Hg, level-4 SBP was associated with lower adjusted risks of mortality at 12 months (OR, 0.71; CI, 0.61-0.81), 3 months (OR, 0.72; CI, 0.61-0.85), and 1 month (OR, 0.61; CI, 0.48-0.79) and in the hospital (OR, 0.71; CI, 0.58-0.88) than level-1 SBP. The cumulative times at the other 3 SBP ranges (<100 mm Hg, 100-120 mm Hg, and >= 140 mm Hg) were not independent risk predictors of prognosis. Furthermore, Group M had lower 12-month mortality than Group L, which remained in the propensity-score matched and subgroup analyses. Conclusions: SBP at 120 mm Hg to 140 mm Hg was associated with decreased adverse outcomes. Randomized trials are required to determine whether the outcomes in critically ill patients improve with early maintenance of a SBP level at 120 mm Hg to 140 mm Hg.