Lower limb amputations appear to have been performed as early as 3500BC, and possibly even earlier (Vanderwerker, 1976). Primarily, amputation is either traumatic (secondary to trauma or accident) or planned (scheduled secondary to disease or other event). A third type of amputation is non-traumatic, unanticipated (unplanned) amputation. This last type can be especially worrisome if the unanticipated amputation transpires secondary to hospital admission for some other reason.

Using ANOVA and Tukey's post-hoc tests this retrospective cross sectional study examined hospital characteristics and payer source to see if a positive, predictive relationship existed between such factors and in-hospital mortality rate for this specific patient population. In addition, comorbid conditions, age, gender, and race were considered. Multivariate analysis examined the relationship between LOS, chronic conditions, and comorbid conditions. For inclusion in this study, participants had to (a) have been admitted with one of the top 16 diagnoses common to patients who undergo amputation, and (b) have an unanticipated non-traumatic lower limb amputation following admission for one of those 16 diagnoses. The study used data from the 2008 Healthcare Cost and Utilization Project Nationwide Inpatient Sample (HCUP NIS, 2008).

For this patient population, hospitalization in the Northeast carried a higher overall risk of mortality, smaller hospitals appeared to provide more favorable outcomes than medium or large hospitals, and private insurance carried a lower mortality risk, especially when compared to Medicare. Understandably, patients over 85 had the highest mortality risk, while those aged 18-64 had the lowest mortality risk. Gender-wise, males had a higher frequency of unanticipated lower limb amputation but females had higher risk of mortality when they did undergo the procedure. Blacks appeared to have a lower mortality risk despite the apparent increased incidence of many of the comorbid conditions. Some of our variables did not return significant results. In this regard, hospital location (rural/urban) and teaching status both appeared to have insignificant impact on mortality risk, A surprising result from the analysis indicated that the more comorbid conditions a person had the lower their mortality risk. Multivariate analysis revealed an inverse relationship between increasing comorbid conditions and length of hospital stay, while also finding that less comorbid conditions equaled less chronic conditions. Finally, MANOVA revealed no overall significant comorbid condition effect. Thus, results were mixed, indicating that there are certain conditions that may predict higher likelihood of mortality as an outcome for patients receiving an unanticipated amputation. Awareness of these risk factors may help in planning future proactive care options for patients at risk for in-hospital mortality following lower limb amputations.