# Development of a Clinical and Laboratory-Based Predictive Nomogram Model for Unfavorable Functional Outcomes Among Patients Who Undergo Interventions for Aneurysmal Subarachnoid Hemorrhage

## TL;DR

This study creates a predictive model to identify patients at risk of poor outcomes after treatment for aneurysmal subarachnoid hemorrhage using clinical and lab data.

## Contribution

A novel nomogram model combining clinical and perioperative laboratory changes to predict unfavorable outcomes in aneurysmal SAH patients.

## Key findings

- Diabetes mellitus, WFNS grade 3–5, clipping, and changes in WBC, ALT, sodium, and glucose are independent predictors of poor outcomes.
- The nomogram model achieved high predictive accuracy in both derivation and validation cohorts (AUC 0.839 and 0.797, respectively).
- Perioperative lab changes improve prediction performance for unfavorable outcomes in aneurysmal SAH patients.

## Abstract

Objective: This study elucidates the prognostic significance of perioperative changes in laboratory indicators for aneurysmal SAH and develops a nomogram model for outcome prediction. Methods: Aneurysmal SAH patients who received clipping or coiling at our institution between January 2016 and December 2022 were included. All patients were randomly assigned to derivation and validation cohorts. Independent predictors of unfavorable outcomes were identified by multivariate analyses. Three models were conducted to evaluate whether perioperative laboratory changes improve prediction performance. A nomogram including all independent predictors was developed in the derivation cohort and verified in both cohorts. Results: Diabetes mellitus [OR (95% CI) = 2.84 (1.44–5.59)], WFNS grade 3–5 [OR: (95% CI), 9.17 (5.49–15.33)], clipping [OR (95% CI) = 1.71 (1.03–2.85)], perioperative changes in white blood cell count [OR (95% CI) = 2.15 (1.17–3.96)], and concentrations of ALT [OR (95% CI) = 1.41 (1.04–1.91)], sodium [OR (95% CI) = 5.40 (3.01–9.71)], and glucose [OR (95% CI) = 2.18 (1.05–4.53)] were independent predictors of an unfavorable outcome. The predictive nomogram incorporated the aforementioned predictors and performed well in the derivation cohort (AUC, 0.839; 95% CI: 0.810–0.866) and the validation cohort (AUC, 0.797; 95% CI: 0.734–0.850). Conclusions: Perioperative changes in laboratory indicators can be predictors of unfavorable outcomes in aneurysmal SAH patients. The nomogram based on clinical and laboratory risk factors can be used as a convenient tool to facilitate individualized decision making.