The value of forced expiratory volume in 1 s in screening subjects with stable COPD for PaO2 < 7.3 kPa qualifying for long-term oxygen therapy

Guidelines on the management of chronic obstructive pulmonary disease (COPD) issued by the European Respiratory Society (ERS), British Thoracic Society (BTS), American Thoracic Society (ATS), and Department of Health for England and Wales (DoH) suggest differing values of forced expiratory volume in 1 s (FEV1) below which arterial blood gas analysis should be performed to determine the presence of severe hypoxaemia and possible long-term oxygen therapy (LTOT) requirement. This study aimed to determine the value of FEV1 at these different levels in screening for LTOT requirement defined as PaO2 < 7.3 kPa in subjects with stable COPD. Comparative measures were taken against other lung function tests of volume and diffusing capacity. A retrospective analysis of paired lung function and arterial oxygen measurements in 491 subjects was made. The positive and negative predictive values, sensitivity and specificity of FEV1 < 70% predicted (ERS), FEV1 < 50% predicted (ATS), FEV1 < 40% predicted (BTS) and FEV1 < 1.51 (DoH) were determined for fulfilling LTOT criteria (PaO2 < 7.3 kPa). The correlation between lung function variables and PaO2 was established. Logistic regression analysis was used to classify subjects with PaO2 < 7.3 kPa and PaO2 > or = 7.3 kPa. Using FEV1 to screen for LTOT requirement produced a high negative predictive value at all four suggested limits (FEV1 < 70% 100%, FEV1 < 50% 96%, FEV1 < 40% 95%, FEV1 < 1.51 97%). However, the positive predictive values were low (FEV1 < 70% 13%, FEV1 < 50% 16%, FEV1 < 40% 19%, FEV1 < 1.51 15%) as were sensitivities. No single lung function variable was a strong determinant of PaO2. FEV1 % pred (r = 0.40), FVC % pred (r = 0.34) and TLCO % pred (r = 0.27) had the strongest relationships. Logistic regression also placed FEV1 % pred and TLCO % pred as the best predictors of PaO2 < 7.3 kPa. We conclude no lung function variable correlates well with PaO2 in subjects with stable COPD. The best predictor of PaO2 < 7.3 kPa was FEV1 % pred. Whilst a low FEV1 is a poor predictor of LTOT requirement in an individual, PaO2 < 7.3 kPa is only found in subjects with a low FEV1. A high FEV1 may be used to exclude subjects from further investigation for LTOT and prevent unnecessary arterial sampling.