P262\u2005Can we improve upon clinician prediction of survival in advanced COPD using clinically measurable prognostic factors?

Abstract

Introduction and objectives Although advanced COPD is associated with poor survival, individual trajectories are difficult to predict and clinician ability to recognise those in the last year of life is unknown. Factors which are prognostic for mortality have been identified across the spectrum of COPD severity. Whether they retain prognostication in advanced disease is also unknown. We investigated clinician prognostication and clinically measurable prognostic factors for mortality in patients with advanced COPD. Methods Patients were recruited from an advanced COPD service between October 2013 and July 2018 forming a prospective observational cohort. Measures collected at baseline included spirometry, MRC dyspnoea grade, body mass index (BMI), four metre gait speed (4MGS), exacerbation history, home-oxygen use and presence of comorbidities. Clinicians’ prediction of one year mortality was recorded in response to the ‘Surprise Question’: ‘Would I be surprised if this patient died within the next 12 months?’. Receiver operating characteristic (ROC) analysis was performed to determine the accuracy of clinician assessment of prognosis. Mortality data were censored at February 2019. Survival analysis was performed using multivariate Cox regression. Results 398 were patients recruited, 59% male, 24% current smokers, mean±SD age 66±9 yr, FEV1% predicted 35±13%, BMI 26±7 kg/m2, 88% MRC dyspnoea scale grade ≥4, 29% used home oxygen, 91% had a COPD exacerbation in the past year, and 88% had co-morbidities. Average follow-up time was 888 days, 145 deaths (36%) occurred and one-year mortality rate was 12%. The positive and negative predicted values for clinicians’ prediction of one-year mortality were 24% and 93%, with an area under ROC of 0.65. Adjusted time to event analysis for patients with complete baseline data (n=277) showed older age, lower BMI and slower 4MGS were independently associated with increased risk of mortality (Table 1).Abstract P262 Table 1 Multivariate Cox regression analysis of survival in patients with advanced COPD HR SE CI p-value Age at first CRA (years) 1.05 0.02 1.02 – 1.08 < 0.001 FEV1 (L) 0.54 0.24 0.23 – 1.28 0.16 MRC dyspnoea scale grade 0.94 0.24 0.8 – 1.54 0.81 4m gait speed (m/s) 1.38 0.18 1.07 – 1.79 0.01 BMI (kg/m2) 1.03 0.01 1.01 – 1.06 0.01 Home oxygen use No 1 Yes 1.27 0.35 0.74 – 2.17 0.39 COPD exacerbation(s) in previous 12 months No 1 Yes 3.12 1.87 0.96 – 10.11 0.06 Comorbidity present No 1 Yes 1.59 0.69 0.68 – 3.74 0.28 Covariates: high to low except for 4m gait speed and BMI where lower gait speed and BMI are associated with increased mortality BMI = body mass index; FEV1= Forced expiratory volume in one second; MRC = Medical Research Council dyspnoea scale grade Conclusions In patients with advanced COPD, clinicians do not accurately identify those within the last year of life. Alongside age and BMI, 4MGS is an independent predictor of mortality in advanced COPD. A prognostic scoring system including these indices has the potential to assist clinicians identify patients in the last year of life supporting proactive development of advance care plans.