Oxygen Versus Air-Driven Nebulisers for Exacerbations of Chronic Obstructive Pulmonary Disease

A Randomised Controlled Trial

George Bardsley; Janine Pilcher; Steven McKinstry; Philippa Shirtcliffe; James Berry; James Fingleton; Mark Weatherall; Richard Beasley


BMC Pulm Med. 2018;18(157) 

In This Article


In this study, oxygen-driven nebulisation increased the PtCO2 in hospital in-patients with an AECOPD compared with air-driven nebulisation. Despite the small mean increase in PtCO2 of 3.4 mmHg, the physiological relevance of this response is suggested by the increase in PtCO2 of at least 4 mmHg in 18/45 (40%) of participants receiving oxygen-driven nebulisation, whereas no patient had an increase of 4 mmHg or more following air-driven nebulisation. The clinical relevance of this physiological response is suggested by the requirement to withdraw one participant during the second oxygen-driven nebulisation due to the PtCO2 increasing by > 10 mmHg, and the increase of PtCO2 or PcapCO2 of at least 8 mmHg in 4/45 (9%) patients receiving oxygen-driven nebulisation, one of whom had a fall in pH of 0.06 into the acidotic range (7.32). These findings suggest that air-driven nebulised bronchodilator therapy represents an important component of the conservative titrated oxygen regimen which has been shown to reduce the risk of hypercapnia, acidosis and mortality in AECOPD.[1]

There are a number of methodological issues relevant to the interpretation of the study findings. Both the randomised controlled design and double-blinding of this study allow for robust and reliable data capture. The length of the nebuliser regimen was chosen to ensure adequate time for complete nebulisation to occur, and to replicate 'real-world' back to back treatments in the acute setting, by using two nebulisations separated by five minutes. It is possible that the magnitude of the differences in PCO2 and pH may be even larger with continuous nebulisation which may occur in patients with severe exacerbations not responding to initial treatment or if the nebuliser is inadvertently left in place. The safety-based exclusion criteria of a baseline PtCO2 > 60 mmHg and an oxygen requirement of ≥4 L/minute (to maintain target SpO2 of 88 to 92%), effectively excluded patients with the most severe exacerbations of COPD.

Whilst respiratory rate and neurological symptoms were not formally assessed as outcome measures, no adverse events were identified during the interventions. However, we acknowledge that if changes in PCO2 and pH of this magnitude occurred in more severe patients at the time of their presentation, they would have been at risk of symptoms of hypercapnia and respiratory acidosis, and the requirement to escalate treatment.

The original primary outcome measure and time of measurement was PcapCO2 after 35 min. Following the first 14 study participants, it was evident that obtaining adequate amounts of blood to fill the capillary tubes from some participants was difficult or impossible to the extent that 4 out of 14 participants had one or more missed samples. For this reason, the primary outcome was changed to PtCO2 after 35 min. In other words, the method of capturing the change in PCO2 was revised, rather than the outcome itself. PtCO2 monitoring enabled continuous assessment to be undertaken, and is accurate in AECOPD,[9] and other acute settings.[10–12] The validity of this method was confirmed by the post hoc analysis of 80-paired samples, where each capillary blood gas sample obtained had a corresponding PtCO2 measurement at the same time-point. This showed that the difference between the PcapCO2 and PtCO2 in the mean change from baseline was − 0·03 mmHg with 95% confidence intervals of − 0.44 to 0.38 mmHg. This data suggests that the use of PtCO2 measurements did not adversely affect our ability to determine change in PcapCO2 from baseline.

We did not investigate the potential mechanisms by which oxygen driven nebulisation increases PtCO2. However as demonstrated in mechanistic studies of oxygen therapy in COPD, it is likely to be due to the combination of a reduction in respiratory drive, release of hypoxic pulmonary vasoconstriction, absorption atelectasis, and the Haldane effect.[13,14] Furthermore, the study was not designed to assess costs related to each regimen, however it is reasonable to assume that improved clinical outcomes seen by avoiding a rise in PtCO2 and associated acidosis, would lead to a reduction in healthcare costs.

The findings from our study complement those of our previous randomised controlled trial of a similar design in stable COPD patients in the clinic setting, in which there was a mean PtCO2 difference between the oxygen- and air-driven nebulisation treatment arms of 3.1 mmHg (95% CI 1·6 to 4·5), p < 0·001, after 35 min.[3] In that study one of the 24 subjects was withdrawn due to an increase in PtCO2 of 10 mmHg after 15 min of the first oxygen-driven nebulisation. As with the previous study, an increase in PtCO2 occurred within 5 min, indicating the rapid time course of this physiological response. We had anticipated a greater effect in this current study as the patients had acute rather than stable COPD however the magnitude of the effect was similar, probably reflecting the similar severity of airflow obstruction, with a mean predicted FEV1 of 35% and 27% in this and the previous study respectively.

The two previous open crossover studies of inpatients with AECOPD both showed oxygen-driven nebulisation worsened hypercapnia in patients with Type 2 respiratory failure.[4,5] Gunawardena et al.[4] studied 16 patients with COPD and reported that only those with carbon dioxide retention at baseline (n = 9) demonstrated a rise in PaCO2 after 15 min (mean of 7·7 mmHg), and one patient had a rise of 22 mmHg. Similarly, O'Donnell et al[5] reported that 6/10 patients, all with carbon dioxide retention at baseline, showed a rise in PaCO2 after 10 min (mean of 12.5 mmHg).

The current BTS guidelines recommend air-driven nebulisation and, if this is not available in the ambulance service, the maximum use of 6 min for an oxygen-driven nebuliser. This is based on the rationale that most of the nebulised medication will have been delivered, and is categorised as grade D evidence.[6] We observed the mean time for dissipation of salbutamol solution from the nebuliser chamber of 5.2 min confirming that 6 min is adequate for salbutamol delivery. The proportion of participants with a PtCO2 increase ≥4 mmHg was lower after 6 min than 15 min, suggesting some amelioration of risk with the shorter nebulisation treatment. Alternative methods of bronchodilator delivery include air-driven nebulisers or multiple metered dose inhaler actuations via a spacer.[15]

The potential for rebound hypoxia after abrupt cessation of oxygen therapy has been observed both in the treatment of asthma and COPD.[9,16,17] We identified some evidence consistent with this phenomenon which is a potentially important yet poorly recognised clinical issue.