Objectives Thirty-day mortality (30 DM) is a measure of quality of cancer treatment and the predictors for 30 DM are important to identify vulnerable patients who are least likely to benefit from chemotherapy. We assessed the incidence and potential predictors of 30 DM in patients receiving chemotherapy.

Methods All patients who received chemotherapy within an 8-month period in our hospital were assessed. Baseline prechemotherapy clinical features, vital signs, Modified Early Warning Scores (MEWS) and laboratory tests were recorded. Potential predictors of 30 DM were evaluated using multivariate logistic regression analysis.

Results Among 4560 patients included, 77 patients (1.7%) died within 30 days of chemotherapy. Patients who died were older (62 vs 58, p=0.002), mostly males (61% vs 43%, p=0.006), had worse Eastern Cooperative Oncology Group performance scores (ECOG PS), and higher MEWS scores compared with those who survived. Multivariate analysis identified age ≥60 years (OR 2.2, 95% CI 1.2 to 4.1, p=0.01), male gender (OR 2.1, 95% CI 1.1 to 3.9, p=0.02), ECOG PS≥3 (OR 3.2, 95% CI 1.1 to 8.8, p=0.03), pulse rate ≥90 bpm (OR 3.8, 95% CI 2.0 to 7.0, p<0.01), systolic blood pressure <110 mm Hg (OR 2.1, 95% CI 1.1 to 4.1, p=0.02), body mass index <25 kg/m² (OR 2.1, 95% CI 1.1 to 3.8, p=0.02) and haemoglobin< 90 g/L (OR 14.2, 95% CI 4.3 to 46.6, p<0.01) to be associated with increased risk of 30 DM.

Conclusions Along with well-known prognostic factors such as ECOG PS and disease stage, other simple and readily available parameters may predict early mortality after chemotherapy and produce a signal for the physicians to carefully reevaluate vulnerable patients before chemotherapy administration.