OBJECTIVES: To compare how different European HTA agencies assess surrogate endpoints to demonstrate efficacy. METHODS: We identified 8 therapies with surrogate endpoints that were evaluated in the last 6 years by NICE and/or SMC (UK), HAS (France) and G-BA (Germany).  The acceptability of the use of surrogate endpoints and any specific comments made by these agencies were analysed. RESULTS:  Commonly used surrogate endpoints such as glycated haemoglobin (HbA1c) in diabetes, progression free survival (PFS) in oncology and forced expiratory volume (FEV1) for respiratory diseases have been generally accepted as sufficient evidence to gain reimbursement by HTA agencies. Especially when a surrogate endpoint has been accepted by EMA, it is usually considered a valid outcome measure. Less well-accepted were several surrogate cardiovascular endpoints such as 6 minute walk test, blood pressure and LDL cholesterol. For G-BA it is important that surrogate endpoints have been properly validated and are patient relevant but they did accept endpoints such as sustained virological response (SVR) for hepatitis treatments, FEV1 and body mass index (BMI) based on minor evidence. NICE and SMC also strongly value evidence to demonstrate the correlation between surrogate endpoints and clinical outcomes. Interestingly SMC has recently become more cautious in accepting widely established endpoints such as HbA1c. With regards to the HAS, they often did not comment on the use of surrogate endpoints at all in their published reports. CONCLUSIONS: The use of surrogate endpoints in the assessment of clinical benefit is still controversial; however, attempts are made to establish clearer regulations such as the recently published EUnetHTA guidelines regarding surrogate endpoints. In the absence of evidence on final patient-relevant clinical endpoints, several commonly used biomarkers and intermediate endpoints will be considered as valid surrogate endpoints by HTA agencies. Newer, less established surrogate endpoints will be more subject to strict validation requirements.