Comparative associations of baseline frailty status and age with postoperative mortality and duration of hospital stay following metastatic brain tumor resection

Alis J. Dicpinigaitis,Simon Hanft,Jared B. Cooper,Chirag D. Gandhi,Syed Faraz Kazim,Meic H. Schmidt,Fawaz Al-Mufti,Christian A. Bowers
DOI: https://doi.org/10.1007/s10585-021-10138-3
2022-01-13
Abstract:Metastatic brain tumors are the most common intracranial neoplasms diagnosed in the United States. Although baseline frailty status has been validated as a robust predictor of morbidity and mortality across various surgical disciplines, evidence within cranial neurosurgical oncology is limited. Adult metastatic brain tumor patients treated with resection were identified in the National Inpatient Sample during the period of 2015–2018. Frailty was quantified using the 11-point modified frailty index (mFI-11) and its association with clinical endpoints was evaluated through complex samples multivariable logistic regression and receiver operating characteristic (ROC) curve analyses. Among 13,650 metastatic brain tumor patients identified (mean age 62.8 years), 26.8% (n = 3665) were robust (mFI = 0), 31.4% (n = 4660) were pre-frail (mFI = 1), 23.2% (n = 3165) were frail (mFI = 2), and 15.8% (n = 2160) were severely frail (mFI ≥ 3). On univariable assessment, these cohorts stratified by increasing frailty were significantly associated with postoperative complications (13.6%, 15.9%, 23.9%, 26.4%; p < 0.001), mortality (1.2%, 1.4%, 2.7%, 3.2%; p = 0.028), and extended length of stay (eLOS) (15.7%, 22.5%, 28.9%, 37.7%; p < 0.001). Following multivariable logistic regression analysis, frailty (by mFI-11) was independently associated with postoperative mortality (aOR 1.34, 95% CI 1.08, 1.65) and eLOS (aOR 1.26, 95% CI 1.17, 1.37), while increasing age was not associated with these endpoints. ROC curve analysis demonstrated superior discrimination of frailty (by mFI-11) in comparison with age for both mortality (AUC 0.61 vs. 0.58) and eLOS (AUC 0.61 vs. 0.53). Further statistical assessment through propensity score adjustment and decision tree analysis confirmed and extended the findings of the primary analytical models. Frailty may be a more robust predictor of postoperative outcomes in comparison with age following metastatic brain tumor resection.
oncology
What problem does this paper attempt to address?