RESULTS: We included 151 patients (83 males, 68 females), 117 had COPD, 74 patients were on home O2, 23 on home CPAP or BIPAP. Sixty nine patients required NIV. Mean±SD values were 68±12 for age and 7.31±0.06, 67±22 mmHg, 34±10 mmol/L for venous pH, PCO2 and bicarbonates respectively. As compared to intubated patients (n=24), non-intubated patients (n=127) had a higher Glasgow Coma Score (14.8±0.9 versus 13.6±2.5, p=0.03) but no significant difference in venous pH (7.32±0.06 versus 7.29±0.07, p=0.08), PCO2 (66±22 versus 71±19 mmHg, p=0.33) or bicarbonates (34±10 versus 34±9 mmol/L, p=0.98) was observed. ROC curve showed poor sensitivity and specificity of venous parameters for prediction of intubation with an AUC of 0.37 for pH. Hospital mortality was 7% (n=11), without significant differences in VBG associated with mortality. As compared to patients not requiring NIV (n=82), patients started on NIV (n=69) had lower venous pH (7.30±0.06 versus 7.33±0.06, p=0.005) and higher PCO2 (74±24 versus 61±17 mmHg, p=0.005) and bicarbonates (36±10 versus 32±9 mmol/L, p=0.04), and among this group patients who subsequently required intubation (n=16) had lower venous PCO2 (63±17 versus 77±25 mmHg, p=0.02) and bicarbonates (31±8 versus 37±10 mmol/L, p=0.03) than non-intubated patients, while their pH was not significantly different (7.30±0.08 versus 7.30±0.06, p=0.78).