Development and validation of 'Patient Optimizer' (POP) algorithms for predicting surgical risk with machine learning.

BACKGROUND: Pre-operative risk assessment can help clinicians prepare patients for surgery, reducing the risk of perioperative complications, length of hospital stay, readmission and mortality. Further, it can facilitate collaborative decision-making and operational planning. OBJECTIVE: To develop effective pre-operative risk assessment algorithms (referred to as Patient Optimizer or POP) using Machine Learning (ML) that predict the development of post-operative complications and provide pilot data to inform the design of a larger prospective study. METHODS: After institutional ethics approval, we developed a base model that encapsulates the standard manual approach of combining patient-risk and procedure-risk. In an automated process, additional variables were included and tested with 10-fold cross-validation, and the best performing features were selected. The models were evaluated and confidence intervals calculated using bootstrapping. Clinical expertise was used to restrict the cardinality of categorical variables (e.g. pathology results) by including the most clinically relevant values. The models were created with logistic regression (LR) and extreme gradient-boosted trees using XGBoost (Chen and Guestrin, 2016). We evaluated performance using the area under the receiver operating characteristic curve (AUROC) and the area under the precision-recall curve (AUPRC). Data was obtained from a metropolitan university teaching hospital from January 2015 to July 2020. Data collection was restricted to adult patients undergoing elective surgery. RESULTS: A total of 11,475 adult admissions were included. The performance of XGBoost and LR was very similar across endpoints and metrics. For predicting the risk of any post-operative complication, kidney failure and length-of-stay (LOS), POP with XGBoost achieved an AUROC (95%CI) of 0.755 (0.744, 0.767), 0.869 (0.846, 0.891) and 0.841 (0.833, 0.847) respectively and AUPRC of 0.651 (0.632, 0.669), 0.336 (0.282, 0.390) and 0.741 (0.729, 0.753) respectively. For 30-day readmission and in-patient mortality, POP with XGBoost achieved an AUROC (95%CI) of 0.610 (0.587, 0.635) and 0.866 (0.777, 0.943) respectively and AUPRC of 0.116 (0.104, 0.132) and 0.031 (0.015, 0.072) respectively. CONCLUSION: The POP algorithms effectively predicted any post-operative complication, kidney failure and LOS in the sample population. A larger study is justified to improve the algorithm to better predict complications and length of hospital stay. A larger dataset may also improve the prediction of additional specific complications, readmission and mortality.

RNA Interference With Zilebesiran for Mild to Moderate Hypertension: The KARDIA-1 Randomized Clinical Trial.

Importance: Angiotensinogen is the most upstream precursor of the renin-angiotensin-aldosterone system, a key pathway in blood pressure (BP) regulation. Zilebesiran, an investigational RNA interference therapeutic, targets hepatic angiotensinogen synthesis. Objective: To evaluate antihypertensive efficacy and safety of different zilebesiran dosing regimens. Design, Setting, and Participants: This phase 2, randomized, double-blind, dose-ranging study of zilebesiran vs placebo was performed at 78 sites across 4 countries. Screening initiation occurred in July 2021 and the last patient visit of the 6-month study occurred in June 2023. Adults with mild to moderate hypertension, defined as daytime mean ambulatory systolic BP (SBP) of 135 to 160 mm Hg following antihypertensive washout, were randomized. Interventions: Randomization to 1 of 4 subcutaneous zilebesiran regimens (150, 300, or 600 mg once every 6 months or 300 mg once every 3 months) or placebo (once every 3 months) for 6 months. Main Outcomes and Measures: The primary end point was between-group difference in least-squares mean (LSM) change from baseline to month 3 in 24-hour mean ambulatory SBP. Results: Of 394 randomized patients, 377 (302 receiving zilebesiran and 75 receiving placebo) comprised the full analysis set (93 Black patients [24.7%]; 167 [44.3%] women; mean [SD] age, 57 [11] years). At 3 months, 24-hour mean ambulatory SBP changes from baseline were -7.3 mm Hg (95% CI, -10.3 to -4.4) with zilebesiran, 150 mg, once every 6 months; -10.0 mm Hg (95% CI, -12.0 to -7.9) with zilebesiran, 300 mg, once every 3 months or every 6 months; -8.9 mm Hg (95% CI, -11.9 to -6.0) with zilebesiran, 600 mg, once every 6 months; and 6.8 mm Hg (95% CI, 3.6-9.9) with placebo. LSM differences vs placebo in change from baseline to month 3 were -14.1 mm Hg (95% CI, -19.2 to -9.0; P < .001) with zilebesiran, 150 mg, once every 6 months; -16.7 mm Hg (95% CI, -21.2 to -12.3; P < .001) with zilebesiran, 300 mg, once every 3 months or every 6 months; and -15.7 mm Hg (95% CI, -20.8 to -10.6; P < .001) with zilebesiran, 600 mg, once every 6 months. Over 6 months, 60.9% of patients receiving zilebesiran had adverse events vs 50.7% patients receiving placebo and 3.6% had serious adverse events vs 6.7% receiving placebo. Nonserious drug-related adverse events occurred in 16.9% of zilebesiran-treated patients (principally injection site reactions and mild hyperkalemia) and 8.0% of placebo-treated patients. Conclusions and Relevance: In adults with mild to moderate hypertension, treatment with zilebesiran across a range of doses at 3-month or 6-month intervals significantly reduced 24-hour mean ambulatory SBP at month 3. Trial Registration: ClinicalTrials.gov Identifier: NCT04936035.