Malignant esophagorespiratory fistulas: a comparative effectiveness survival analysis

To analyse the survival of patients with malignant esophagorespiratory fistulas (ERF) and perform a comparative effectiveness analysis of ERF-directed interventions.

Fifty-five patients met our inclusion criteria, and data on ERF-directed interventions (stent placement, surgical repair and best supportive care) and their outcomes, along with clinical and treatment details, were recorded. The primary endpoints were overall survival (OS) and survival after developing ERF (F-OS).

The median OS and F-OS for the entire cohort was 299 days (SE=23.2) and 123 days (SE=11.63), respectively. On univariable analysis, the American Joint Committee on Cancer (AJCC) stage (I-III vs IV) influenced both OS (403 vs 171 days; p=0.006) and F-OS (129 vs 67 days; p=0.034). Proximal location of ERF influenced OS favourably (494 vs 285 days; p=0.021), whereas patients developing ERF late in their disease course experienced inferior F-OS (96 vs 232 days; p=0.03). On multivariable analysis, the AJCC stage (IV vs I-III, HR=3.03 (1.41-6.50)), time to developing ERF from diagnosis (greater than vs within 3 months, HR=5.82 (1.84-18.36)) and location of ERF (distal vs proximal, HR=2.47 (1.14-5.34)) had a significant impact on F-OS. The comparative efficacy (OS and F-OS) of best supportive care was statistically equivalent to any intervention irrespective of AJCC stage and success/failure of initial or subsequent intervention(s).

The survival of patients with ERF is dismal, and our analysis suggests that best supportive care results in equivalent OS and F-OS when compared with any intervention.