Etiology and Anatomical Level of Lower-Limb Amputations in Two Iranian Referral Hospitals: A Retrospective Record-Based Study

Background: Lower-limb amputation arises from heterogeneous pathways ranging from severe trauma to dysvascular disease and infection, yet comparative data across Iranian hospitals remain limited. We aimed to characterize demographic, etiologic, and anatomical patterns of lower-limb amputation and to evaluate between-hospital heterogeneity. Methods: This retrospective, two-hospital, record-based study included 229 eligible hospitalization records with lower-limb amputation at Al-Zahra and Kashani hospitals in Isfahan, Iran, from 13 April 2014 through 18 March 2020. Primary etiology was reconstructed from ordered diagnostic codes using a predefined hierarchical algorithm; external-cause codes were reviewed as supporting contextual information and to characterize documented traumatic mechanisms. Amputation level was derived from ICD-9-CM procedure codes. Analyses used Python 3.13.5 (NumPy, pandas, SciPy, and statsmodels). Results: Trauma was the leading etiology (49.3%), followed by dysvascular disease (28.8%) and infection (10.5%). Age differed strongly across etiologies (H = 93.71; Holm-adjusted p < 0.001; epsilon-squared = 0.398, 95% CI 0.307–0.509), with trauma-related hospitalizations occurring at younger ages. Toe (35.4%) and below-knee (34.5%) amputations were most frequent. Hospital was associated with both etiology and amputation level (both Holm-adjusted p < 0.001; bias-corrected Cramer’s V = 0.374 and 0.471, respectively). In an adjusted major-versus-minor model, older age and treatment at Al-Zahra Hospital were associated with major amputation, whereas female sex was not statistically significant. Conclusions: The hospitalizations showed marked descriptive differences by age, etiology, and hospital. Trauma-related hospitalizations occurred at younger ages, whereas dysvascular and infection-related hospitalizations occurred later in life. The substantial between-hospital differences indicate that pooled hospital-based estimates may obscure institutional case mix and should not be interpreted as population-level incidence.