Empty pelvis syndrome as a cause of major morbidity after pelvic exenteration: validation of a core data set.

Abstract

<h4>Background</h4>Pelvic exenteration (PE) is a potentially curative treatment for advanced pelvic cancers. However, PE procedures are associated with empty pelvis syndrome (EPS), a spectrum of complications including pelvic sepsis, sinus formation, fistulae, and bowel obstruction. Inconsistent reporting has impeded progress in understanding EPS. The PelvEx Collaborative introduced a core data set of descriptors and outcomes to address these issues and the aim of this study was to validate this data set.<h4>Methods</h4>An observational cohort study applied the EPS core data set to a prospectively maintained PE database. Patterns of major and minor manifestations were evaluated; logistic regression was used to explore relationships between descriptors and outcomes, and inter-descriptor correlation was assessed using Cramer's V.<h4>Results</h4>EPS occurred in 32.1% of patients (105 of 327) and was the leading cause of major morbidity. Infected pelvic collections (occurring in 23.5%) were associated with subsequent chronic sinus formation (OR 3.08, P = 0.01) and fistulae (P = 0.05). The risk of EPS increased with external beam radiotherapy (OR 1.01 per 1 Gy, P = 0.01), sacrectomy (OR 3.78, P < 0.001), total cystectomy (OR 2.46, P = 0.001), internal iliac vessel ligation (unilateral OR 1.94, P = 0.045; bilateral OR 3.65, P < 0.001), and infralevator exenteration (OR 3.69, P < 0.001). Omentoplasty reduced pelvic bowel obstruction (OR 0.27, P = 0.004) and perineal flaps were linked to a higher rate of reconstruction-related major morbidity compared with biological mesh alone (20.8% versus 1.2% respectively, P = 0.002).<h4>Conclusion</h4>The PelvEx Collaborative core data set standardizes reporting of EPS, with this study detailing the acute and chronic complications arising as a consequence. Biological mesh was associated with reduced reconstruction-related morbidity compared with perineal flaps. Further validation in additional cohorts is required to address potential confounding factors.