Background and study aim: The introduction of direct acting antiviral agents shifted the management of chronic hepatitis C virus (HCV) infection to a new level. Pretreatment predictors of benefit are needed to help the selection of patients for treatment. The aim of this work is to study if Model for End Stage Liver Disease (MELD) score can be reliably used as a predictor of response to treatment with direct acting antivirals (DAAs) in ‘difficult to treat' chronic HCV patients. Patients and Methods: This is a retrospective study where files of 91 “difficult to treat" patients were randomly selected from the follow up clinic. Patients' data were collected before and after treatment including history taking, clinical examination, laboratory investigations and abdominal ultrasonography. MELD and Child-Turcotte-Pugh (CTP) scores were calculated. Results: After treatment, MELD score was significantly improved in 28.6% of patients, remained stable in 57.1% and worsened in 14.3%. MELD score was significantly higher among patients with complications than those without complications before and after treatment.No significant difference was detected between patients with and without sustained virologic response (SVR) as regard MELD score changes after treatment. Conclusion: Baseline MELD score cannot predict the response to treatment of “difficult to treat" chronic HCV patients but can predict the occurrence of complications.