A practical index to predict 30-day mortality after major amputation.

BACKGROUND: Patients at risk of mortality after amputation have not been well identified. We sought to devise a clinical index predicting 30-day mortality after amputation that would allow stratification of intensity of postoperative care.
METHODS: The National Surgical Quality Improvement Program (NSQIP) database (2005-2009) was analyzed for patients who had above- or below-knee amputations. An additive risk index was created based on logistic regression that examined patient demographics, comorbidities, and operative characteristics. A threshold score for clinical action was identified as the score at which the gain in certainty was maximized. The primary outcome measure was 30-day mortality.
RESULTS: Among 9244 patients analyzed, there were 744 deaths (8.1%) at 30 days, with 280 occurring after hospital discharge (37.9%). The final index includes 11 components with a total score range of 0-13: age (60-79 or ≥80 years), history of congestive heart failure, chronic obstructive pulmonary disease, or major cardiac surgery, using steroid medications, having dependent functional status, dyspnea, being on dialysis, having impaired sensorium, or preoperative sepsis. This index has a c-statistic of 0.7391, and the score at which clinical action should be taken is ≥5. The observed probability of 30-day mortality increased from 1.06% at a score of 1 to 10% at 5 and 38.5% at a score of 10.
CONCLUSIONS: More than one-third of deaths within 30 days of major amputation occur after discharge from acute care. A novel index to predict 30-day mortality after major amputation is described. Patients receiving a score ≥5 face a substantial risk of mortality and should be held in the hospital longer or, if discharged, receive closer postoperative follow-up.