P220 Transfer factor and arterial oxygen partial pressure are predictors of survival in hospital outpatients with COPD

Abstract

Of potential prognostic variables for COPD to use in a multidimensional staging tool, BMI, severity of airflow obstruction (FEV 1 % predicted) and the MRC Dyspnoea Score (BOD) have been shown to be useful for a more complete assessment of the impact of the disease.1 In 1999e2002 we identified 431 patients (208 women) with COPD in primary care surgeries and recorded their BOD scores (maximum 7). For men and women (mean6SD): age was 66610 and 64610; pack/years 34619 and 34618; BMI 2765 and 2566; FEV 1 % predicted 55615 and 57616; and MRC Dyspnoea Score 2.661.0 and 2.661.0; with BOD scores 1.961.5 and 2.061.6 respectively. The low mean BOD scores indicate the mainly moderate severity for this primary care cohort (only one subject had an MRC score of 5). Because of the similarities in these variables for men and women further analysis is of the whole cohort. Mortality was assessed in October 2010 and a KaplaneMeier analysis for those with BOD scores 0 and 1 shows 80% and 75% probability of survival, whereas for a BOD score of >4 the 10-year survival was <20%. By way of contrast GOLD stages 1 and 2 had a 75% and 60% survival with Stage 4 showing a 60% survival. The Cox regression model (Abstract P219 table 1) demonstrated that BOD was a better predictor of survival than age and that smoking history was a significant covariate. Ten year survival was <10% when smoking history (with age and co-morbidity score) was added to a BOD score of >4 in a KaplaneMeier plot. Figure 1 presents the KaplaneMeier survival curves, adjusted for age and PO 2 , for the two population groups, separated using the TLCOc% median value as a cut-off point (>38.0 and ¼38.0% predicted).Abstract P220 Figure 1 Survival curves adjusted for PO 2 and age for the two patient groups, separated according to TLCOc% predicted value.Conclusions Gas transfer measurement provides additional prognostic information compared to spirometry.