RESUMO
BACKGROUND & AIMS: Progresses in management make a higher proportion of cirrhotic patients with gastrointestinal (GI) cancer candidates to chemotherapy. Data are needed on the safety and liver-related events associated with the use of chemotherapy in these patients. METHODS: Forty-nine patients with cirrhosis receiving chemotherapy against GI cancer from 2013 to 2018 were identified in the French Health Insurance Database using ICD-10 codes K70-K74, and matched 1:2 to non-cirrhotic controls (n = 98) on age, tumour type and type of treatment. Adverse events (AE), dose tapering, discontinuation rate, liver-related events and survival rate were compared. RESULTS: Patients with cirrhosis (Child-Pugh A 91%) more often received lower doses (38.8% vs 7.1%, p < .001), without significant differences in terms of grade 3/4 AE or dose tapering rates (29.6% vs. 36.7%; 22.3% vs 24.4%, respectively). Treatment discontinuation rate was higher in patients with cirrhosis (23.3% vs. 11.3%, p = .005). Child-Pugh (p = .007) and MELD (p = .025) scores increased under chemotherapy. Five patients with cirrhosis (10.2%) had liver decompensation within 12 months, and 17.2% of deaths in the cirrhosis group were liver-related versus 0% in matched controls. WHO-PS stage > 1 (HR 3.74, CI95%: 2.13-6.57, p < .001), TNM-stage M1 (HR 3.61, CI 95%: 1.82-7.16, p < .001), non-colorectal cancer (HR 1.73, CI 95%: 1.05-2.86, p = .032) and bilirubin higher than 5 mg/dL (HR 2.26, CI 95%: 1.39-3.70, p < .001) were independent prognostic factors of 2-year mortality, whereas cirrhosis was not. CONCLUSIONS: Chemotherapy should be proposed only in patients with compensated cirrhosis with close monitoring of liver function. Dose management remains challenging. Multidisciplinary management is warranted to improve these patients' outcomes.