INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The medical management of patients with chronic obstructive pulmonary disease (COPD) who receive tracheal intubation and mechanical ventilation for acute respiratory failure (ARF) commonly includes the administration of bronchodilator aerosols in order to reduce the dynamic hyperinflation and the markedly elevated airway resistance. Aerosols can be generated by two distinct devices, namely, metered-dose inhalers (MDI) and small-volume nebulizers (NEB). The optimal deposition of aerosols to the lower respiratory tract in these patients depends on a lot of factors, including the device and its location, the particle size, the drug dose given, factors related to ventilator and circuit (tidal volume, duty cycle, humidification), presence and size of endotracheal tube, and disease severity (1). The choice between MDI and NEB is also based upon the effects of the bronchodilator on the respiratory mechanics and cost-effectiveness considerations. In vitro investigations have been done to study the lung deposition of aerosols during mechanical ventilation with MDI or NEB by varying some of the factors mentioned above (2). From these in vitro studies, recommendations have been proposed to optimally use MDI and NEB in intubated patients (1). Only a few studies assessed the effectiveness of MDI or NEB in mechanically ventilated patients. Noncomparative studies have shown that MDI was effective in improving respiratory mechanics (8). Studies comparing MDI and NEB showed either no difference (12, 13) or a greater effect of NEB over MDI (14). To our knowledge no study compared the cost- effectiveness for MDI and NEB administration. Therefore, some controversy remains regarding the choice of MDI versus NEB administration of bronchodilator aerosols in intubated and mechanically ventilated patients (15). Fenoterol-ipratropium bromide combines a selective ₂-adrenergic agonist, fenoterol, with a cholinergic antagonist, ipratropium bromide, and has been found more effective than ipratropium alone (9). No study in the literature compared the effects on respiratory mechanics of fenoterol-ipratropium bromide administered by MDI or NEB to mechanically ventilated patients with COPD. Therefore, in the present study in mechanically ventilated patients with COPD in ARF, we aimed at studying in detail the respiratory mechanics, and specifically the flow-resistive properties of the respiratory system, after delivery of fenoterol- ipratropium bromide by either MDI or NEB.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

We enrolled 18 patients (13 male) who received tracheal intubation and mechanical ventilation for ARF in our Medical Intensive Care Unit. Their mean ± SEM values of age, height, and weight were 67 ± 3 yr, 166 ± 3 cm, and 76 ± 3 kg, respectively. They all had COPD, which was diagnosed by clinical history, chest radiographs, and pulmonary function tests. The mean values of FEV₁ and FEV₁/VC before ARF were 0.77 ± 0.06 L (31 ± 3% of predicted) and 0.46 ± 0.04 (63 ± 6% of predicted), respectively (16). Acute respiratory failure had been triggered by an acute exacerbation in 10 patients and pneumonia in eight patients. They were investigated 1 to 10 d after the onset of tracheal intubation and mechanical ventilation (mean ± SEM 2.7 ± 0.7 d). The investigation was approved by the institutional ethics committee, and informed consent was obtained from the next of kin for each patient.

The patients were orotracheally intubated (Mallinckrodt cuffed-endotracheal tube 7.5, 8, or 8.5 mm ID; Mallinckrodt Laboratories, Athlone, Ireland) and mechanically ventilated (Cesar; Taema, Anthony, France) under volume-controlled mode with a squared inspiratory flow on zero end-expiratory pressure (ZEEP). During the study all patients were sedated with midazolam (0.2 mg/kg) and paralyzed with atracurium bresylate (0.3 to 0.6 mg/kg). The mean values of the baseline ventilatory settings, which were kept constant throughout the experiment, are listed in Table 1. The inspiratory to total cycle duration ratio (TI/Ttot) was 0.27 ± 0.01. Airflow () was measured with a heated pneumotachograph (Fleisch No.2; Fleisch, Lausanne, Switzerland) inserted between the endotracheal tube and the Y-piece of the ventilator. The pressure drop across the two ports of the pneumotachograph was measured with a differential piezoelectric pressure transducer (163PC01D36, ± 12.7 cm H₂O; Micro Switch, Freeport, IL). The response of the pneumotachograph was linear over the experimental range of . Pressure at the airway opening (Pao) was measured proximal to the endotracheal tube with a piezoelectric pressure transducer (143PC03D, ± 176 cm H₂O; Micro Switch). Tracheal pressure (Ptr) was measured via a polyethylene catheter (1.5 mm ID), with multiple side holes and an occluded end hole, placed 2 cm past the carinal end of the endotracheal tube and connected to a piezoelectric pressure transducer (143PC03D, ± 176 cm H₂O; Micro Switch). With the system used to measure Pao and Ptr, there was no appreciable shift or alteration in amplitude up to 20 Hz. The equipment dead space (not including the endotracheal tube) was 150 ml. All variables were recorded on an IBM compatible computer by a 12-bit analog digital board (DT2801-A) interfaced with data acquisition software (Labdat; RHT-Infodat Inc., Montreal, Canada) at a sample frequency of 100 Hz. Subsequent data analysis was made with Anadat (RHT-Infodat). In this analysis, inflation volume (V) was obtained by digital integration of the signal. Special care was taken to avoid gas leaks in the equipment and around the tracheal cuff. To reduce the effects of the compliance and resistance of the system connecting the subjects to the ventilator on the mechanics measurements, we used a single length of standard adult low compliance tubing supplied with the machine (2 cm ID, 110 cm long) and omitted the humidifier during the assessment of respiratory mechanics.

Arterial blood gases were measured with an ABL 500 blood gas analyzer (Radiometer, Copenhagen, Denmark). Before the experiment, Pa_O2 averaged 84 ± 6 mm Hg, Pa_CO2 averaged 53 ± 3 mm Hg, and pH was 7.41 ± 0.02 under the baseline ventilatory settings listed in Table 1. Heart rate, systemic blood pressure measured noninvasively (HPM 1008B; Hewlett-Packard, Waltham, MA), and transcutaneous O₂ saturation (Sp_O2) (HPM 1020A; Hewlett-Packard) were continuously monitored.

Experimental Design

Basically, we underwent a prospective single-blind, randomized, crossover study. Patients received sequentially in a random order (random order table) fenoterol-ipratropium bromide by MDI and NEB administered by the same respiratory therapist. A period of at least 10 h was allowed between the administration of the bronchodilator with the two modalities. Fenoterol-ipratropium bromide delivered by MDI was given as four puffs (50 µg fenoterol and 20 µg ipratropium bromide per puff) by using a nonvalved MDI adapter (Medispacer; Baxter Laboratories, Maurepas, France). The MDI adapter was positioned in the inspiratory limb of the ventilator circuit 15 to 20 cm from the Y piece. After shaking the MDI canister, MDI was actuated just before the onset of the mechanical breath and a 4-s inflation hold was then applied. We waited 60 s between each actuation and repeated the same maneuver until the total dose was delivered. During this procedure, the ventilatory settings were not changed. Treatment with NEB was given at a dosage of 1.25 mg fenoterol and 500 µg ipratropium bromide in 5 ml of saline. The drug was placed in a nebulizer (Misty-Neb; Baxter Laboratories), which was inserted at the same place in the ventilator circuit as for the MDI inhalation. This nebulizer has a dead volume of 2 to 5 ml of gas and of 0.5 ml of saline. The gas flow from the ventilator was 5 L/min. The device was run until almost dry by visual inspection, which was obtained after 30 min on average. NEB was connected to a specific port at the ventilator allowing us to use the ventilator's flow during nebulization. A correction factor was used to keep the tidal volume delivered by the ventilator constant during nebulization. Fenoterol-ipratropium bromide specimens and MDI adapters and nebulizers were provided by Boehringer-Ingelheim Laboratories (Reims, France) and Baxter Laboratories, respectively.

Respiratory mechanics and vital signs (heart rate, Sp_O2, and systemic blood pressure) were assessed before and 30 min after the end of each modality of administration. Before inhalation of bronchodilator and before respiratory signal sampling, the trachea was gently suctioned to remove secretions. Six patients received inhaled bronchodilators other than fenoterol-ipratropium bromide before entry into the present study. In these patients, this treatment was withheld for at least 4 h before the onset of the investigation.

Physiologic Measurements and Data Analysis

Patients were investigated in a semirecumbent position.

Respiratory Mechanics

Respiratory mechanics were assessed by the constant- rapid airway occlusion method previously described in detail (17). Whereas baseline V was kept constant (Table 1), was changed randomly from 0.2 to 1.2 L/s for single test breaths by regulating TI with the appropriate knob of the ventilator (iso-V experiment). The end-inspiratory occlusions, obtained by pressing the end-inspiratory hold knob on the ventilator, lasted 5 s. Before each test breath, an end-expiratory occlusion was performed by pressing the end-expiratory hold button on the ventilator. This allowed us to quantify PEEPi and to start the test breath from a fixed static elastic equilibrium condition. It should be noted that on ZEEP the ventilator generated a slightly positive end-expiratory pressure, averaging 1.13 ± 0.11 cm H₂O. Therefore, the pressure measured by the end-expiratory occlusion is the sum of PEEPi and the positive pressure generated by the ventilator. We termed this pressure PEEP total (PEEPt). Because PEEPi implies dynamic pulmonary hyperinflation (i.e., that the end-expiratory lung volume during mechanical ventilation exceeds the relaxation volume of the respiratory system [Vr]), we also measured the difference between end-expiratory lung volume and Vr (EELV) by reducing the ventilator frequency to its lowest value during baseline expiration, thus prolonging expiratory duration to allow the patient to exhale to Vr. Vr was achieved when expiratory flow became nil and expiratory occlusion resulted in no change in airway pressure (i.e., no PEEPi). After each test breath, the baseline ventilation was resumed until V, , Pao, and Ptr returned to their baseline values (usually within a few breaths). Each measurement was done twice.

After end-inspiratory airway occlusions, Ptr and Pao exhibited an initial rapid drop (Pmax-P₁) followed by a slow decay to an apparent plateau pressure. During this period, the contribution of reduction in pressure caused by volume loss by continuing gas exchange should be negligible. Ptr measured at 5 s was taken to represent the static end-inspiratory elastic recoil pressure of the respiratory system (Pst,rs). Dividing Pmax,tr-Pst,rs and Pmax,tr-P₁,tr by the immediately preceding the occlusion, total resistance (Rrs) and interrupter resistance (Rint,rs) of the respiratory system were obtained. The additional tissue resistance (Rrs) was calculated as the difference between Rrs and Rint,rs. In computation of Rint,rs the errors caused by the closing time of the ventilator valve were corrected as previously described (18). The static elastance of the respiratory system (Est,rs) was computed by dividing the values of Pst,rs-PEEPt by baseline V.

Model and Curve-fitting

Our data were analyzed in terms of the model of the respiratory system proposed by D'Angelo and colleagues (18) for humans (Figure 1). This model comprises two compartments in parallel. The first is a dashpot representing Rint,rs, which explains the initial fast pressure drop observed in Prs immediately after the end-inspiratory occlusion. Rint,rs is the sum of the lung and chest wall. Bates and colleagues (19) found that in dogs, in the absence of the chest wall, the initial pressure jump reflects only the airway resistance. Contrary to dogs (20), in normal anesthetized paralyzed humans (21) and in patients with COPD (17), the chest wall does not substantially contribute to Rint,rs. The second compartment of the model in Figure 1 is a Kelvin body, which consists of a standard static elastance (Est,rs) in parallel with a Maxwell body, i.e., a spring E₂ and a dashpot R₂ arranged serially. In normal anesthetized paralyzed subjects, E₂ and R₂ probably virtually entirely reflect the viscoelastic properties of the tissues of the lungs and chest wall (21).

View larger version (13K): [in this window] [in a new window]   Figure 1.   Scheme of spring-and-dashpot model for interpretation of respiratory mechanics during constant flow interruption. Respiratory system consists of standard resistance (Rint,rs) in parallel with standard elastance (Est,rs) and a series spring-and-dashpot body (E2 and R2, respectively) that represents stress adaptation units. Distance between the two horizontal bars is analogue of lung volume (V) and tension between these bars is analogue of pressure at airway opening (Pao).

During constant- inflation, the model in Figure 1 predicts that Rrs should increase with TI according to the following function (14):

(1)

where ₂ = R₂/E₂.

Because during constant- inflation TI = V/, Equation 1 can be rewritten (14)

(2)

In the present study, we could fit Equation 1 to our experimental data of Rrs before and after the inhalation of fenoterol-ipratropium bromide by MDI and NEB in all of the patients, as shown in Figure 2 in a representative patient.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Relationships of additional resistance (Rrs) to inspiratory time (TI) in Patient 13 with COPD before and after inhalation of fenoterol-ipratropium bromide by MDI and NEB. Curves are computed according to Equation 1. Solid and broken lines are regression lines before and after treatment, respectively.

In vitro Assessment of Inhaled Mass

Inhaled mass, defined as the quantity of drug actually reaching the end of the endotracheal tube (22), was determined in vitro under conditions that replicated the clinical study. The mean values of the ventilatory settings given in Table 1 were used. The drug was delivered with the same devices and at the same dosages. With NEB, the duration of nebulization was set to 30 min. With MDI, four puffs were given at 1-min intervals and then the mechanical ventilation was prolonged to a 30-min duration. A circular low-resistance absolute filter 70 mm in diameter (Gelman Sciences, Ann Arbor, MI) was interposed between the distal extremity of the endotracheal tube and a low-resistive bag (3). At the end of the 30-min period, the filter was removed and the bronchodilator molecules were extracted. Extraction was achieved by placing the filter in a glass bag beaker containing 10 ml 0.1N NaOH that was vortexed for 1 min. Then the solution was centrifuged between 0.500 and 0.600 × g for 10 min. Six experiments were conducted with NEB and 12 with MDI then averaged. Fenoterol was assayed as described by Diot and colleagues (5). Briefly, 2-ml samples of supernatant were read by spectrophotometer (Beckman DU-6; Beckman Instruments, Irvine, CA) at a wavelength of 240 nm, as determined by a previous spectrometer analysis of fenoterol in a 0.1 N NaOH solution. The blank was prepared by processing a circular filter 70 mm in diameter without fenoterol, as described previously. The standard curve was established by assaying known amounts of fenoterol between 50 and 300 µg pipetted onto the surface of a circular filter 70 mm in diameter. Its reproducibility was established on three different days. Assuming that the maximum mass of ipratropium bromide deposited on the filters would be 35% and 25% of the mass of drug delivered by the MDI or NEB, respectively (23), the maximum estimated concentrations of ipratropium bromide that could be obtained after washing the filter with 10 ml 0.1 N NaOH would be 2.8 and 12.5 µg/ml, respectively. To our knowledge, there is no method that would allow us to assay ipratropium bromide at such low concentrations, and we focused the analysis on fenoterol.

Particle Size Distribution

The particle size distribution of fenoterol-ipratropium bromide obtained with each aerosol generator was determined in vitro by using a laser velocimeter (APS 33 T. S. I., St. Paul, MN). The bronchodilator was delivered under the same conditions as in the clinical study (same dosages, same devices, same ventilator with respiratory parameters producing the same flow rate by using the mean values of the ventilatory settings, as given in Table 1).

Data Analysis

The investigators who performed the postsampling analysis of the respiratory signals were blinded to the treatment modality. The comparisons of the values of respiratory mechanics and vital signs before and after inhalation were made within and between delivery modalities by using Student's paired t tests. We applied Bonferroni's correction to obtain the significance level that an individual p value must satisfy to achieve significance at a p value < 0.05. Regression analysis was made by the least-squares method. Comparisons of the values of respiratory resistances obtained at different during the isoV experiments were made by repeated measures analysis of variance (ANOVA). The values were expressed as their mean ± SEM. For our statistics, we used SPSS software (SPSS for Microsoft Windows V 6.0.1; SPSS Inc., Chicago, IL).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All patients investigated in the present protocol exhibited PEEPi that averaged 6.52 ± 0.89 cm H₂O and was associated with expiratory flow limitation. In all patients indeed, the passive expiratory -V curve exhibited a concavity toward the volume axis, a finding that has been found associated with expiratory flow limitation (24). Consistent with PEEPi, all patients exhibited dynamic hyperinflation, as reflected by the values of EELV (Table 2).

Respiratory Mechanics at Baseline Ventilatory Settings

As shown in Table 2, at the baseline ventilatory settings, the mean values of inspiratory resistances, PEEPt, EELV, and Est,rs before inhalation of fenoterol-ipratropium bromide were not different between MDI and NEB. Both modalities of administration reduced to the same extent the mean values of Rrs, EELV, and PEEPt. However, the two components of Rrs were not similarly affected. Indeed, with MDI, Rint,rs decreased significantly, with no change in Rrs. By contrast, with NEB, an opposite finding was observed, namely, a significant reduction of Rrs without any significant change in Rint,rs. The values of Est,rs were not different before MDI and NEB and did not change significantly after inhalation of fenoterol-ipratropium bromide with either mode of delivery (Table 2).

Respiratory Resistances During iso-V Experiment

In all patients, Rint,rs increased linearly with increasing , at constant V (Figure 3) according to the following function:

(3)

View larger version (10K): [in this window] [in a new window]   Figure 3.   Mean relationships of interrupter resistance (Rint,rs) to inflation flow () at constant inflation volume before and after inhalation of fenoterol-ipratropium bromide by MDI and NEB. Bars are SEM.

where K₁ and K₂ are Rohrer's constants whose mean ± SEM values are given in Table 3. The mean values of K₁ and K₂ were not different before inhalation between MDI and NEB and did not change significantly after inhalation of fenoterol-ipratropium bromide from MDI. The same was true for the values of K₁ after inhalation from NEB. By contrast, the values of K₂ decreased significantly after administration by NEB and therefore were significantly lower than the values of K₂ obtained after inhalation by MDI (Table 3). However, by applying repeated-measures ANOVA to the data of Figure 3, we found that Rint,rs was significantly influenced by (p < 0.0001), but not by the modality of delivery.

The average values of the "viscoelastic" constants of the respiratory system of our patients with COPD are shown in Table 4. The values of R₂, ₂, and E₂ were not different before administration between the two modalities and did not change significantly after inhalation. The average iso-V relationships of Rrs with are illustrated in Figure 4. The values of Rrs decreased progressively and significantly (p < 0.0001) with increasing and were significantly reduced with NEB as compared with MDI over the whole range of (p < 0.05 by repeated-measures ANOVA).

View larger version (10K): [in this window] [in a new window]   Figure 4.   Mean relationships of additional resistance (Rrs) to inflation flow () at constant inflation volume before and after inhalation of fenoterol-ipratropium bromide by MDI (left panel ) and NEB (right panel ). Bars are SEM.

The average iso-V relationship of Rrs with is displayed in Figure 5. Before inhalation, the values of Rrs initially decreased until a minimum value was reached at a of 0.6 L/s; thereafter, Rrs tended to increase slightly with further increase in . The values of Rrs were significantly reduced by NEB administration (p < 0.01 by repeated-measures ANOVA).

View larger version (10K): [in this window] [in a new window]   Figure 5.   Mean relationships of total respiratory resistance (Rrs) to inflation flow () at constant inflation volume, before and after inhalation of fenoterol-ipratropium bromide by MDI ( left panel ) and NEB (right panel ). Bars are SEM.

Vital Signs

As shown in Table 5, heart rate, systolic and diastolic systemic blood pressure, and Sp_O2 did not change significantly with either modality of inhalation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study of intubated and mechanically ventilated patients with COPD in ARF, we have found that: (1) 30 min after inhalation of fenoterol-ipratropium bromide, Rrs, EELV, and PEEPt significantly decreased; (2) at baseline Rrs was reduced through a decrease in Rint,rs with MDI but through a decrease in Rrs with NEB; and (3) over a range of the Rohrer's constant K₂ and the iso-V changes of Rrs were significantly influenced by NEB but not by MDI administration.

To our knowledge, only one study reported the effects of inhaled fenoterol-ipratropium bromide in intubated and mechanically ventilated patients with COPD (9). In this study, Fernandez and coworkers (9) administered by MDI two puffs of fenoterol-ipratropium bromide (half the dose we presently used) in 12 patients with COPD intubated for an acute exacerbation. They observed that 60 min later PEEPi and Rrs were reduced on average by 24% and 16%, respectively. These changes were greater than those obtained in the present study, suggesting that the maximal effect of fenoterol-ipratropium bromide could be obtained later than the 30-min period we have chosen; however, in nonintubated asthmatics, the maximal effects on FEV₁ of nebulized fenoterol and of nebulized ipratropium bromide were observed 30 min after drug administration (25). However, the study of Fernandez and coworkers (9) and our own are difficult to compare for the following reasons. First of all, they administered the drug in a different way than we did in that they used a catheter connecting the MDI device to the endotracheal tube and delivered the bronchodilator through a manual inflation. Second, to our knowledge, the dose-response relationship of inhaled fenoterol-ipratropium bromide is unknown in intubated and mechanically ventilated patients. Third, in their study, the measurements of Rrs included the resistance caused by the endotracheal tube, resulting in an overestimation of the intrinsic total resistance of the respiratory system. Finally, the duration of the end-inspiratory pause in their study was much shorter than what we used, leading to an overestimation of the values of Pst,rs and an underestimation of those pertaining to Rrs.

Surprisingly, in our study, we found that apparently the device used to deliver fenoterol-ipratropium bromide to the respiratory system could have influenced the components of the inspiratory resistances differently. Intuitively, the main therapeutic effect of a bronchodilator we were expecting was a reduction in airway resistance. Only a few studies have partitioned Rrs into its airway and its tissue components. Dhand and colleagues (10, 26), in mechanically ventilated patients with COPD, found that the significant decrease of Rrs after inhalation of albuterol was essentially due to a reduction of Rint,rs. In our study, at baseline applied during mechanical ventilation, the values of Rint,rs were significantly reduced after MDI but not after NEB administration. The variability between patients might have blunted the significance of the difference in Rint,rs observed with NEB (Table 2). The baseline values of Rint,rs in our study are lower than those previously reported in mechanically ventilated patients with COPD in ARF (17, 27). This difference can be partly explained by the relatively low applied to our patients with COPD. We tried to extend the results of Dhand and colleauges (10, 26) obtained at a single by varying at a constant V. The average value of the constant K₂, which was thought to reflect the additional resistance caused by the turbulent flow in the central airways, was significantly reduced by NEB without any significant effect of MDI.

Rrs was not significantly changed with albuterol administered by MDI in the patients studied by Dhand and colleagues (10, 26). We have observed the same finding by using MDI to deliver fenoterol-ipratropium bromide. In marked contrast, in our study, the baseline values of Rrs decreased significantly after inhalation of fenoterol-ipratropium by using the NEB device. In our patients, Rrs fitted Equation 1 in the four experimental conditions. After inhalation of fenoterol-ipratropium bromide, with both devices, the average values of the "viscoelastic constants" did not change significantly (Table 4). However, the isoV relationships of Rrs to were not similar regarding the modality of inhalation (Figure 5). At any given , the mean values of Rrs were significantly different before and after NEB inhalation, which was not the case with MDI administration. Some variability in our results may have blunted the significance of these changes. There is no previous data, to our knowledge, upon the effects of bronchodilator administration on the "viscoelastic constants" of the respiratory system to compare with. The structural meaning of the viscoelastic constants to date remains unknown. Assuming that the effects of NEB administration on Rrs are true, how could a bronchodilator affect tissue resistance and why such a difference between the two modalities of administration? In animals, the contribution of tissue resistance to the increase in lung resistance observed after induced bronchoconstriction has been shown to be dramatic. Indeed, in rats challenged with nebulized metacholine, Nagase and colleagues (30) found that tissue resistance increased by 9-fold and represented 81.8% of total lung resistance after metacholine inhalation. Because no bronchodilator was further administered we cannot argue from this study that bronchodilator agent may reverse bronchoconstriction by acting predominantly at the tissue level. In addition, in our patients, the low level of applied during baseline ventilation may have made of Rrs the predominant component of Rrs relative to Rint,rs. Why did the two modalities of administration behave differently in this respect? This might be due to differences in drug deposition into the respiratory tract. In vitro, the inhaled mass of fenoterol averaged 225 ± 11.6 µg with NEB and 60.3 ± 4.2 µg with MDI (p < 0.001). These values corresponded to 18.1% of the mass of fenoterol placed in the nebulizer and to 30.1% of that delivered by the MDI device. As already mentioned, for technical reasons, we did not measure the inhaled mass of ipratropium bromide. The results of the particle size distribution of fenoterol-ipratropium bromide obtained with both aerosol devices in vitro are shown in Table 6. The values of geometric standard deviation (GSD) were those of polydisperse aerosols with each generator. However, the mean value of mass median aerodynamic diameter (MMAD) obtained with MDI was significantly higher than that obtained with NEB, whereas the percentage of particles with MMAD < 5 µm was significantly lower with MDI than with NEB. The probability of deposition in thoracic airways can be estimated from the model (31). Accordingly, an overall deposition of 34.7% and of 28.5% could be computed for MDI and NEB, respectively. With MDI and NEB, the probability of deposition can be estimated to 24% and 21%, and to 10.7% and 7.5% at the alveolar and airway levels, respectively. Hence, by merging the results of these two in vitro experiments together, we could try to estimate, at least for fenoterol, the intrathoracic quantitative mass deposition. The total mass of fenoterol deposited could amount to 60 µg with MDI and to 225 µg with NEB. The mass of fenoterol deposited in the airways could be around 6 µg with MDI and 17 µg with NEB, whereas the corresponding values for the alveolar deposition could be 14 and 47 µg, respectively. These in vitro studies showed that the deposition of fenoterol could have been different with the two modalities of inhalation, the NEB modality delivering to the lungs four times more fenoterol than the MDI modality. The quantity of drug deposited at specific levels of the respiratory tract could have also been different. However, we did not obtain estimates of lung deposition for ipratropium bromide. So, it is speculative to infer from these results the difference of the values of the components of Rrs we observed in the clinical study. In addition, the agreement between in vitro and in vivo studies is inconstant. For example, in the study of Fuller and colleagues (12) in mechanically ventilated patients, the lung deposition of radiolabeled fenoterol was higher when the drug was administered by MDI than by NEB, whereas the changes in peak inspiratory pressure were not different between the two modalities. It must be stressed, however, that recent studies, where the inhaled mass was accurately assessed (4, 5, 32), in fact showed that large amounts of drug can be delivered to the patients with nebulizers. Hence, our present data are consistent with these studies, whereas the initial study of Fuller and colleagues (12) is not.

In conclusion, our study has provided a systematic description of the effects on respiratory resistances of the administration of inhaled fenoterol-ipratropium bromide given by MDI and NEB in patients with COPD intubated and mechanically ventilated for acute respiratory failure. We have found that both devices equivalently reduced Rrs without side effects, but they apparently affected differently the components of Rrs. Finally, the in vitro assessment of the inhaled mass of fenoterol showed that the nebulizer delivered a larger amount of drug than the metered-dose inhaler we tested.