CLINICAL PREDICTORS OF 30-DAY MORTALITY IN HOSPITALIZED PATIENTS WITH LUNG CANCER: A RETROSPECTIVE SINGLE-CENTER OBSERVATIONAL STUDY

Background: Lung cancer patients experience high rates of hospitalization, mainly due to the high risk of complica- tions that emerge during the natural history of disease. We designed a retrospective single-center observational study aimed at defining clinical predictors of 30-day mortality in hospitalized lung cancer patients. Methods: Clinical records from the first admission of lung cancer patients to the Oncology ward of University Hospital of Parma from January 1 st, 2017 to January 1 st, 2022 were collected. Results: 251 consecutive patients were enrolled at the time of data cut-off. Median age was 66 years (Range, 29-85). 161 (64%) patients were men and 210 (84%) cur- rent/former smokers. ECOG-PS was 0/1 in 175 (69%) cases. The most frequent histology was adenocarcinoma (n=120, 48%), followed by small-cell lung cancer (n=73, 29%). The majority of patients had stage IV disease (n=209, 83%) and half of them presented bone metastasis. Main reasons for hospital admission were anticancer treat- ment start (n=146, 58%) and cancer-related complications (n=75, 30%). The most frequently reported symptoms at admission were pain (n=131, 52%) and dyspnea (n=92, 37%). Median Blaylock Risk Assessment Screening Score (BRASS) was 7 (Range, 1-28), with 153 (61%) patients having at least one previous hospital admission within the last 3 months. Clinical deterioration (n=102, 41%) and infections (n=71, 28%) were the main complications dur- ing the hospital stay. Baseline clinical predictors of 30-day mortality were high BRASS score (>20 vs ⩽ 20: 34.3% vs 11.9%, p<0.001), previous hospital admission (Yes vs No: 22.2% vs 11.2%, p=0.038), number of metastatic sites (⩾ 3 vs < 3: 26.5% vs 13.4%, p=0.017), bone metastasis (Yes vs No: 29.0% vs 10.8%, p=0.001) and presence of pain (Yes vs No: 24.4% vs 11.7%, p=0.009). Among com- plications, clinical deterioration (Yes vs No: 37.3% vs 5.4%, p<0.001) and blood transfusion (Yes vs No: 32.6% vs 15.1%, p=0.010) relevantly increased 30-day mortality. At multivariate analysis, high BRASS score (OR 2.89, 95% CI 1.17-7.12, p=0.021), bone metastasis (OR 2.99, 95% CI 1.07-8.37, p=0.037) and clinical deterioration (OR 9.58, 95% CI 4.18-21.96, p<0.001) were significantly associated with increased risk of 30-day mortality. Conclusions: Our results suggest that basal high BRASS score, bone metastasis and clinical deterioration during hospital stay could be used as clinical predictors of 30-day mortality in hospitalized lung cancer patients.