Abstract
BACKGROUND: Frailty in patients undergoing craniotomy may affect perioperative outcomes. There have been a number of studies published in this field; however, evidence is yet to be summarized in a quantitative review format. We conducted a systematic review and meta-analysis to examine the effects of frailty on perioperative outcomes in patients undergoing craniotomy surgery. METHODS: Our eligibility criteria included adult patients undergoing open cranial surgery. We searched MEDLINE via Ovid SP, EMBASE via Ovid SP, Cochrane Library, and grey literature. We included retrospective and prospective observational studies. Our primary outcome was a composite of complications as per the Clavien-Dindo classification system. We utilized a random-effects model of meta-analysis. We conducted three preplanned subgroup analyses: patients undergoing cranial surgery for tumor surgery only, patients undergoing non-tumor surgery, and patients older than 65 undergoing cranial surgery. We explored sources of heterogeneity through a sensitivity analysis and post hoc analysis. RESULTS: In this review of 63,159 patients, the pooled prevalence of frailty was 46%. The odds ratio of any Clavien-Dindo grade 1-4 complication developing in frail patients compared to non-frail patients was 2.01 [1.90-2.14], with no identifiable heterogeneity and a moderate level of evidence. As per GradePro evidence grading methods, there was low-quality evidence for patients being discharged to a location other than home, length of stay, and increased mortality in frail patients. CONCLUSION: Increased frailty was associated with increased odds of any Clavien-Dindo 1-4 complication. Frailty measurements may be used as an integral component of risk-assessment strategies to improve the quality and value of neurosurgical care for patients undergoing craniotomy surgery. ETHICS AND DISSEMINATION: Formal ethical approval is not needed, as primary data were not collected. SYSTEMATIC REVIEW REGISTRATION: PROSPERO identification number: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=405240.