• COPD exacerbations
• non-invasive ventilation

The national chronic obstructive pulmonary disease (COPD) audit confirms the high mortality associated with acute hypercapnic respiratory failure (AHRF) in COPD, particularly in severely acidotic patients.1 The authors highlight the observations that significant numbers of patients eligible for non-invasive ventilation (NIV) do not receive it and that NIV is almost universally the ceiling of care with only 5% of acidotic patients receiving invasive mechanical ventilation (IMV). Comparisons are made with the outcomes of clinical trials of NIV, and there is an implication that in clinical practice NIV is not being used optimally with patients being denied potentially life-saving treatment. However, patient selection is the likely explanation for the higher mortality rates in the ‘real world’. The greater mortality rates in those receiving NIV at all levels of acidosis, even after allowing for early iatrogenic acidosis due to high flow oxygen, suggests NIV is often used in patients with no chance of survival. The high mortality rate reflects the fact that for many COPD patients AHRF represents the end stage of inexorable decline.

While pH identifies patients in need of ventilatory support, other factors should be considered to determine the appropriate level of intervention. Clinicians use ‘clinical judgement’ and objective evidence to support this may be obtained on routine clinical assessment. Previous national audits identified performance status as an important predictor of survival in patients admitted to hospital with an acute exacerbation of COPD (AECOPD).2 3 We have recently shown that in patients dying from AECOPD a WHO performance score (WHO-PS) ≥3 is a powerful marker of end-stage disease and a better predictor of death than pH.4 In 2009 we prospectively studied COPD patients admitted to hospital with AHRF treated with NIV (n=65). Inpatient mortality was 33.8% and on univariate analysis, factors associated with mortality included poor performance status, long-term oxygen therapy, early warning score, severe acidosis (pH<7.20) and anaemia (table 1). On multivariate analysis only performance status (WHO-PS≥3: OR (95% CI) 39.1 (6.8 to 223.6; p<0.0001) and anaemia (OR (95% CI) 5.87 (1.27 to 26.7; p=0.023) were significant.

We acknowledge that the authors may have highlighted possible deficiencies in delivery of NIV and perhaps more patients should be considered for IMV, but we contend that of equal importance is identification of those patients in whom neither NIV nor IMV is likely to be beneficial so that they may be offered more appropriate end-of-life care.