Little is known about the reliability of G8 screening tool and the prognostic value of clinical parameters within the Comprehensive Geriatric Assessment (CGA) in clinically fit older patients with haematological malignancies. This study was performed to assess the reliability of G8 as a screening tool and to determine the predictive value of CGA items in terms of one-year overall survival (OS). G8 and CGA were proposed to 107 consecutive patients (65-89yrs) with haematological malignancies that were judged clinically fit to receive chemotherapy. Ninety patients were evaluable with both scales; 72% and 80% were defined as "vulnerable" when evaluated with G8 (≤14.5) or CGA (≥2 impairments) respectively. The area under ROC-curve of G8 compared to CGA was 0.749 ± 0.051. Neither G8 nor CGA total score were predictive of one-year OS. However, age (HR=1.09, 95%CI: 1.009-1.185; p=0.029), diagnosis (HR=5.99, 95%CI: 2.322-15.428; p<0.001) and cognitive status (HR=3.50, 95%CI: 1.132-10.849; p=0.030) were predictive of OS. We conclude that in our selected haematological patients 1) the G8 score does not help selecting patients for CGA 2) the G8 and CGA total score do not predict OS 3) in addition to the age and disease itself, cognitive impairment appears to be a powerful prognostic factor.