Abstract Background Using an updated dataset with more patients and extended follow-up, we further established cancer patient characteristics associated with COVID-19 death. Methods Data on all cancer patients with a positive reverse transcription-polymerase chain reaction swab for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) at Guy’s Cancer Centre and King’s College Hospital between 29 February and 31 July 2020 was used. Cox proportional hazards regression was performed to identify which factors were associated with COVID-19 mortality. Results Three hundred and six SARS-CoV-2-positive cancer patients were included. Seventy-one had mild/moderate and 29% had severe COVID-19. Seventy-two patients died of COVID-19 (24%), of whom 35 died <7 days. Male sex [hazard ratio (HR): 1.97 (95% confidence interval (CI): 1.15–3.38)], Asian ethnicity [3.42 (1. 59–7.35)], haematological cancer [2.03 (1.16–3.56)] and a cancer diagnosis for >2–5 years [2.81 (1.41–5.59)] or ≥5 years were associated with an increased mortality. Age >60 years and raised C-reactive protein (CRP) were also associated with COVID-19 death. Haematological cancer, a longer-established cancer diagnosis, dyspnoea at diagnosis and raised CRP were indicative of early COVID-19-related death in cancer patients (<7 days from diagnosis). Conclusions Findings further substantiate evidence for increased risk of COVID-19 mortality for male and Asian cancer patients, and those with haematological malignancies or a cancer diagnosis >2 years. These factors should be accounted for when making clinical decisions for cancer patients.