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



The trial recruited between May 14th 2015 and June 29th 2016. The CONSORT diagram of the flow of the 90 recruited participants through the trial is shown in Figure 1. One participant withdrew after 18 min of air-driven nebulisation because of feeling flushed, and so complete data was available for PtCO2 for 89 participants. The baseline PtCO2 for this participant was 34.3 mmHg and at the time of withdrawal it was 34.6 mmHg. Oxygen-driven nebulisation was stopped in another participant at 27 min when the PtCO2 rose by > 10 mmHg from baseline, a pre-defined safety criterion. The baseline PtCO2 for this participant was 43.4 mmHg and at the time of withdrawal it was 54.1 mmHg. This participant had study measurements continued after this for the full duration of the study. No clinical adverse events were noted during the intervention periods.

Figure 1.

Participant flow through the study and allocation of interventions

A summary of baseline participant characteristics are shown in Table 1. Participants predominantly had severe airflow obstruction with a mean FEV1 of 34.5% predicted. The mean (range) baseline PtCO2 was 37.6 mmHg (24.3 to 58.5 mmHg), and mean SpO2 was 93%. Patients randomised to the oxygen group were more likely to have required assisted ventilation previously. The mean (SD) time for the nebulised salbutamol to dissipate from the chamber was 5.2 (1.2) minutes.


The mean (SD) change in PtCO2 after 35 min was 3.4 (1.9) mmHg in the oxygen group (n = 45), compared to 0.1 (1.4) mmHg in the air group (n = 44). The difference (95% CI) in PtCO2 for oxygen compared to air-driven nebulisations after 35 min was 3.3 mmHg (2.7 to 3.9), p < 0.001. (Table 2 and Figure 2). After adjustment for baseline PtCO2, a history of assisted ventilation, previous hypercapnia and baseline SpO2, were not associated with the PtCO2 at 35 min in either randomised group.

Figure 2.

PtCO2 change from baseline (T = 0) to T = 35 min. Mean PtCO2 with error bars showing one SD, by time and intervention

In 18/45 (40%) participants receiving oxygen-driven nebulisation, PtCO2 increased from baseline by ≥4 mmHg at some stage during the intervention compared to none of the participants receiving air-driven nebulisation, risk difference (95% CI) 40% (25.7 to 54.3), p < 0.001. The full data description and comparisons at each time point are shown in the OLS. Two participants receiving oxygen-driven nebulisation had a rise in PtCO2 ≥ 8 mmHg, one of whom required intervention termination, exceeding the predefined safety criterion of a rise ≥10 mmHg from baseline.

The estimate (95% CI) of the time-related difference, 15 min minus six minutes, for oxygen compared to air, was 0.73 mmHg (0.11 to 1.35), P = 0.021; and for 35 min minus 26 min, 0.43 mmHg (− 0.19 to 1.06), P = 0.17. In the oxygen treatment arm the proportion of patients in whom the PtCO2 increased from baseline by ≥4 mmHg at 6 min was less than the proportion at 15 min: 6/45 (13.3%) and 13/45 (28.9%) respectively, paired difference in proportions (95% CI) 15.6% (3.3 to 27.8), P = 0.013 (Additional file 1: Table S1). The proportion of patients in whom the PtCO2 increased from baseline by ≥4 mmHg at 26 min (6 min into the second oxygen-driven nebulisation) was also less than the proportion at 35 min (completion of the second oxygen-driven nebulisation), although this difference was not statistically significant: 10/45 (22%) and 14/45 (31%) respectively, paired difference in proportions (95% CI) 8.9% (− 3.3 to 20.9), P = 0.15.

The median (25th to 75th percentile) time taken for PtCO2 to return to baseline after cessation of the second nebulisation was 40 (40 to 45) minutes in the air group compared to 50 (45 to 50) minutes in the oxygen group, hazard ratio (95% CI) 1.59 (1.01 to 2.52), P = 0.047.

PcapCO2 and pH

Data summaries for capillary blood gas sampling are shown in Table 3. The difference (95% CI) between oxygen and air for PcapCO2 after 35 min was 2.0 mmHg (1.1 to 2.8), p < 0.001. Thirteen (31.7%) participants receiving oxygen had a rise in PcapCO2 of ≥4 mmHg compared with three (7.7%) receiving air; risk difference (95% CI) 24% (7.5 to 40.5), p = 0.01. In addition to the two participants in whom the PtCO2 increased by ≥8 mmHg, there were two additional participants with capillary data receiving oxygen who had a rise in PcapCO2 of ≥8 mmHg and none from the air group. The mean (95% CI) difference in pH after 35 min was 0.015 units (0.008 to 0.024, p < 0.001) lower for oxygen nebulisation compared to air. One participant experienced a reduction in pH of 0.06 units (from 7.38 to 7.32) in association with a rise in PcapCO2 of 9 mmHg (55 to 64 mmHg).

SpO2 and Heart Rate

The SpO2 was higher throughout both the nebulisation and initial washout periods in the oxygen compared with the air group (see Additional file 3: Table S2). Figure 3 shows the trend for the SpO2 in the oxygen group to fall below that of the air group after cessation of the second nebulisation. At the end of the observation period (80 min), the SpO2 was lower in the oxygen group (difference − 1.22%, 95% CI -2.04 to − 0.39, p = 0.004). The maximum reduction in SpO2 from baseline was 0·8% (95% CI -0.2 to 1.7, P = 0.10) lower after oxygen compared with air nebulisation. The heart rate was slower in the oxygen group at 35 min by 3.3 bpm (95% CI 0.31 to 6.25), p = 0.031 (see Additional file 1: Table S3).

Figure 3.

Time-course of SpO2 throughout study period (Blue = Oxygen-driven nebuliser group, Red = Air-driven nebuliser group)

Methods of PCO2 Measurement

Due to the requirement to change the primary outcome measure, a post-hoc analysis was undertaken to compare the two methods of measuring PaCO2. Based on data for 80 paired PtCO2 and PcapCO2 measurements at baseline and 35 min, the mean (SD) change in PtCO2 was 1.7 mmHg (2.2) with a range of − 2.5 to 8.0 mmHg, and the mean (SD) change in PcapCO2 was 1.7 mmHg (2.3), with a range − 3.0 to 9.0 mmHg. The estimate of bias for change in PcapCO2 minus PtCO2 was − 0.03 mmHg (95% CI -0.44 to 0.38), P = 0.89. The limits of agreement between PtCO2 and PcapCO2 were +/− 3.8 mmHg for each individual measurement obtained.