Background: Cirrhosis is a risk factor with nonhepatic surgery, but only three series regarding esophagectomy are reported. The Model for End-Stage Liver Disease (MELD) score has shown benefit in risk evaluation, but there is no experience regarding esophagectomy. This study aimed to compare the outcomes of surgery for esophageal cancer between cirrhotic and noncirrhotic patients and to evaluate whether the MELD score has a prognostic value for risk stratification.
Methods: From the authors' esophageal cancer database, they selected all the patients with concomitant cirrhosis who underwent surgery with curative intent and a matched cohort of patients without cirrhosis. The preoperative data included demographics, medical history, blood work, American Society of Anesthesiologists (ASA) score, Child-Turcotte-Pugh (CTP) score, and MELD score. The operative data included type of surgery, radicality, operative time, and blood loss. The postoperative data included hemoderivatives, 90-day morbidity and mortality rates, lab works, and hospital length of stay. The cirrhotic patients were further divided and analyzed according to a MELD score cutoff of 9.
Results: Of 3445 esophageal cancer patients, 73 cirrhotic patients underwent surgery. Their 90-day morbidity and mortality rates were higher than those for 146 noncirrhotic patients. The cirrhotic patients also had more respiratory events (p = 0.013) and infections (p = 0.005). The anastomotic complications among the cirrhotic patients were significantly more severe (p = 0.046). No difference in 5-year survival rates was registered. Stratification according to the MELD score showed that patients with a MELD score higher than 9 had a significantly worse postoperative course (5-year survival: p = 0.004). The patients with a MELD score of 9 or lower showed an outcome similar to that of the noncirrhotic patients.
Conclusions: Liver cirrhosis is not an absolute contraindication to esophagectomy. The MELD score can be applicable for esophagectomy risk assessment for cirrhotic patients.