Background: Frailty, a state of decreased physiological reserve, is strongly associated with perioperative mortality in older adults. However, the mechanism by which frailty is associated with mortality is not yet understood. Autonomic dysfunction in the form of decreased intraoperative hemodynamic variability has been shown to be associated with increased mortality. We aimed to see whether frail patients have less hemodynamic variability under anesthesia and whether variability mediates the relationship between frailty and 30-day mortality.
Methods: We performed a single-center retrospective study of 1223 patients ≥65 years of age undergoing surgery between July 2008 and December 2012. We used markers of frailty: age >70, preoperative body mass index <18.5, hematocrit <35%, albumin <3.4 g/dL, and creatinine >2.0 mg/dL. We modeled the outcome of 30-day mortality with number of frailty conditions adjusting for gender, length of surgery, American Society of Anesthesiologists class, and need for transfusion. Intraoperative hemodynamic variability was defined as the count of episodes of absolute change >15% in fractional mean arterial pressure (MAP) between consecutive 5-minute intervals. We evaluated the role of intraoperative hemodynamic variability as a mediator (modifier) of the relationship between frailty and mortality, checking for 3 conditions: (1) frailty must affect episodes of absolute change >15% in fractional MAP; (2) episodes of absolute change >15% in fractional MAP must affect 30-day mortality; and (3) mediation effect is significant. We used the product method, in which the mediation effect was estimated as the product of the first 2 relationships. Then we applied the percentile bootstrap method to obtain the 95% CI for the estimate of mediation effect.
Results: Number of frailty conditions and episodes of absolute change >15% in fractional MAP were inversely proportional. Presence of ≥4 frailty conditions was associated with >40% reduction of the number of episodes of absolute change >15% in fractional MAP. Regarding mortality, episodes of absolute change >15% in fractional MAP were protective. The addition of absolute change >15% in fractional MAP in the mortality model resulted in a decrease in the frailty odds ratio from 10.6 to 9.1 (4+ conditions), suggesting that episodes of absolute change >15% in fractional MAP are indeed a mediator. The mediation effect was modest; 5 episodes of absolute change >15% in fractional MAP was 5.2%, 6.4%, 6.9%, and 9.0% for frailty conditions from 1 to 4+, respectively.
Conclusions: Frailty is associated with less intraoperative blood pressure variation, and the relationship of frailty with 30-day mortality is partially mediated by episodes of absolute change >15% in fractional MAP. This suggests that autonomic dysregulation may be a modest part of the mechanism behind the association between frailty and perioperative mortality. Our finding is consistent with recent literature, suggesting that an intact autonomic nervous system confers lower perioperative mortality.