INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The administration of -adrenergic agents to patients with airways obstruction often results in a transient decrease in Pa_O2 despite concomitant bronchodilation (1). This has been attributed to the pulmonary vasodilator action of these agents, increasing blood flow to poorly ventilated lung regions and thus increasing ventilation-perfusion inequality, and to a shuntlike effect (4, 5). In contrast, anticholinergic bronchodilators have been shown to have relatively small effects on arterial blood gases in either stable chronic obstructive pulmonary disease (COPD) (6) or in acute exacerbations of COPD (7). The recent introduction of the long-acting -adrenergic agent salmeterol for use in patients with COPD raises the question of whether salmeterol has effects on gas exchange that are similar to those of other, more traditional, -agonists in these patients.

We report here the results of a three-way crossover study in which the acute effects of salmeterol inhalation on the arterial blood gas tensions of 20 patients with COPD and arterial hypoxemia at rest were compared with those of the conventional -agonist albuterol and the anticholinergic agent ipratropium, both of which are widely used in treating COPD.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Subjects were recruited from the Hines Veterans Affairs Hospital outpatient population on the basis of having an age of more than 50 yr, a cigarette smoking history of more than 10-pack-yr, a clinical history of COPD (8), an FEV₁ of less than 70% predicted, an FEV₁/FVC ratio of less than 70%, and regular use of at least one bronchodilator. We excluded subjects with unstable comorbidities, any clinical features of asthma (9), an exacerbation of respiratory disease within 1 mo of study, a resting Pa_O2 of less than 55 mm Hg, or use of long-term oxygen therapy.

Study Agents

The study agents were salmeterol xinafoate (2 puffs, 42 µg), albuterol sulfate (2 puffs, 180 µg), and ipratropium bromide (2 puffs, 36 µg), each given under double-blind conditions from metered-dose inhalers (MDIs) without spacers.

Experimental Protocol

Each subject was studied on 3 d, separated from one another by at least 2 d. On each study day, subjects were required to withhold conventional inhaled bronchodilators for at least 12 h and salmeterol or long-acting theophyllines for at least 24 h before study. Subjects receiving alternate-day oral corticosteroids were studied on nonsteroid days, whereas those receiving an inhaled corticosteroid were instructed to take this as usual, together with all their usual nonpulmonary medications. Subjects reported to the laboratory in the morning, after a light breakfast without caffeinated beverages and without having smoked. After a rest of 15 min, an arterial catheter was placed under local anesthesia in the radial or brachial artery. Samples of arterial blood (5 ml) were removed at 5-min intervals for measurement of Pa_O2, Pa_CO2, and pH with a blood gas analyzer (Model 1620; Instrumentation Laboratories, Lexington, MA) that was calibrated daily. The machine output was checked daily with a standard test sample. During the study period, the SD values were ± 0.5 mm Hg for PO₂, ± 0.4 mm Hg for PCO₂, and ± 0.009 for pH). When consecutive Pa_O2 levels fell within 3 mm Hg of each other, the mean of the two values was accepted as the baseline Pa_O2 and the patient received one of the three study treatments under supervision, with the order of treatment being randomized. Blood gas analysis was repeated at 10, 20, 30, 60, 90, and 120 min, previous studies having shown that the changes at these intervals are relatively short-lived (4). Spirometry was performed at baseline, and was repeated 60 min after administration of each study drug to ensure that each agent had been administered and had had its intended effect on airflow.

The study was approved by the Human Studies Subcommittee of the Hines Veterans Affairs Hospital, and each subject gave prior written informed consent.

Analysis of Data

The change in Pa_O2 after each treatment, from the baseline obtained on that day, was the primary outcome variable. The magnitude of this change at each analysis time was compared among treatments. In addition, the area between the Pa_O2 curve and the baseline value (area under the curve [AUC]) was calculated with the trapezoidal method. The AUC calculation included only data from 0 to 90 min after drug administration, since all variables had reverted to baseline by 90 min. The paired t test and analysis of variance (ANOVA) were used to determine the significance of differences among agents. Regression analysis was used to determine whether the change in Pa_O2 was related to the baseline Pa_O2 on that day. Secondary outcomes were changes in Pa_CO2, pH, and the calculated (A-a)DO₂. Statistical significance was accepted at p < 0.05, without correction for multiple tests.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Twenty of the 24 study subjects completed all three study days (Table 1). All had a clinical diagnosis of COPD, with a substantial cigarette smoking history, airways obstruction, and arterial hypoxemia.

The changes in Pa_O2 values following drug administration (Figure 1) show small but statistically significant decreases from baseline after each adrenergic agent. Albuterol resulted in the earliest and greatest absolute decline in Pa_O2, but this was short-lived, being no longer significantly below baseline at 30 min. In comparison, the decline in Pa_O2 after salmeterol was slightly delayed and of slightly lesser magnitude than that of albuterol at its nadir, but more prolonged. Ipratropium resulted in the smallest declines in Pa_O2, which did not bring Pa_O2 significantly below baseline at any time.

View larger version (25K): [in this window] [in a new window]   Figure 1.   Mean changes in PaO2 with time after administration of salmeterol (closed squares), albuterol (closed circles), and ipratropium (closed triangles). Error bars are SEM. (*p < 0.05, +p < 0.005: statistically significant differences between drug and baseline.) Differences were not statistically significant if not shown. Differences among drugs were not statistically significant at any time point.

The area between the Pa_O2 curve and baseline from 0 to 90 min (AUC_0-90min) was greatest for salmeterol (2.46 ± 1.10 mm Hg/h [mean ± SEM]), next greatest for albuterol (1.73 ± 1.33 mm Hg/h), and least for ipratropium (0.48 ± 1.30 mm Hg/h). There were no statistically significant differences among drugs either at individual time points or with respect to AUC_{0-90 min} values.

We sought the largest individual declines in Pa_O2 in order to determine whether administration of any of the study agents might pose an occasional hazard. After administration of salmeterol, the largest individual decline in Pa_O2 was from 76 mm Hg at baseline to 63 mm Hg at 60 min after drug administration, the lowest Pa_O2 in any subject being 59 mm Hg at 60 min. Following albuterol the largest individual decline was from 71.5 to 59 mm Hg at 20 min, this being the lowest Pa_O2 observed with albuterol. Following ipratropium, the largest individual decline was from a baseline value of 70 mm Hg to 59 mm Hg at 30 min. Thus, none of the treatments resulted in a decline in Pa_O2 below 59 mm Hg.

Data for Pa_CO2 (not shown) revealed mean declines of between 0.1 and 1.1 mm Hg for each agent, changes that were not statistically significantly different from baseline or among agents. The changes in arterial pH reflected the Pa_CO2 changes, and were also small and not statistically significant. Data for changes in (A-a)DO₂ (not shown) were thus qualitatively and quantitatively reciprocal to those for changes in Pa_O2.

To further explore the relationship between changes in gas exchange and baseline Pa_O2, we performed regression analysis of the greatest decline in Pa_O2 against the baseline Pa_O2 on that day for each drug. We found for each drug a trend toward a greater decrement in Pa_O2 in subjects with higher baseline Pa_O2 values, but none of these changes was statistically significant.

The FEV₁ increased from baseline by 0.23 ± 0.05 L (mean ± SEM) at 1 h after salmeterol inhalation, by 0.19 ± 0.04 L after albuterol, and by 0.23 ± 0.05 L after ipratropium (p < 0.005 for each agent), the increases not being significantly different among agents. No adverse effects occurred during the course of the study.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The principal finding of this study was that both salmeterol and albuterol resulted in statistically significant but small and transient decreases in Pa_O2 below baseline values. These effects tended to be greater than those seen after administration of ipratropium. The maximal change after albuterol occurred earlier and was greater than the maximal change after salmeterol. However, the AUC _0-90min was slightly but not significantly greater after salmeterol than after albuterol, owing to a more prolonged effect of salmeterol than of albuterol, in accord with the known chronology of effects of these two drugs on airflow.

The results of the study also indicate that the average decreases in Pa_O2 and increases in (A-a)D_O2 were quite small and of questionable clinical significance. Moreover, there was a trend (not statistically significant) for greater declines in Pa_O2 to occur in subjects with higher baseline Pa_O2 values, as has been previously suggested (4). Thus, no subject experienced a decline in Pa_O2 to below 59 mm Hga level that is not unsafefollowing any of the three agents studied.

The observed effects on gas exchange, which are consistent with previous reports for conventional -agonists, have been explained as being due to these drugs' pulmonary vasodilator effects, abrogating the pulmonary vascular reflexes that act to minimize ventilation-perfusion mismatching in COPD, and to possible increases in cardiac output (2, 3). The increase in (A-a)DO₂ after each adrenergic agent used in the present study accords with this explanation. Additionally, because salmeterol is considerably more specific for ₂-receptors than is albuterol (12), its gas-exchange effects are most likely to be mediated via ₂-receptors on vascular smooth muscle. However, it must be noted that all of the study agents' observed effects on gas exchange were of much shorter duration, by about an order of magnitude, than their effects on airflow, which typically persist for 4 to 6 h after albuterol and for 12 h after salmeterol. One could therefore postulate either that the time-constants for agonist-receptor interaction on vascular smooth muscle are very much shorter than those for airway smooth muscle, or that alternative, unknown adaptive mechanisms come into play when vascular reflexes that correct for ventilation-perfusion inequality are overridden by a -adrenergic agonist.

As previously reported in patients with COPD, the anticholinergic agent examined in our study resulted in only small, statistically insignificant changes in arterial blood gases (6). This has been explained as being due to the very slight effects of anticholinergic agents on the pulmonary vasculature (6).

In conclusion, both salmeterol and albuterol, when taken in recommended dosage, resulted in significant but small and transient declines in Pa_O2 and increases in (A-a)DO₂ that could be attributed to their pulmonary vasodilator effects. The effects of salmeterol on gas exchange tended to be slower in onset and more prolonged than those of albuterol, as is consistent with the chronology of the effects of these agents on airflow. Ipratropium tended to have smaller effects than either of the adrenergic agents, but there were no statistically significant differences among the three agents' effects on gas exchange, and none had effects that could be considered to pose a clinical risk either collectively or in individual patients.