The objective of this trial is to compare the total and regional deposition of aerosol in the lungs and oropharynx of patients with COPD and known poor MDI inhalation technique following inhalation of Berodual delivered via the Respimat inhaler and Berodual delivered via an HFA-metered dose inhaler achieved with their "natural" inhalation technique compared with taught "optimal" technique.

This is a single dose, randomised, active-controlled, four period, open-label cross-over trial in adult patients with COPD who have demonstrated a poor MDI technique. Berodual (fenoterol hydrobromide 50 μg + ipratropium bromide 20 μg) will be delivered via the Respimat inhaler on two test days and via the MDI on two test days. Test days with no instruction on correct usage will occur prior to the test days with taught technique, so that the patient's own technique will not be influenced by recent instruction.

Each device will thus first be used with no instructions on correct device use provided. On these no instruction test days each device will be demonstrated and patients will be allowed time to practice on their own with a placebo device. The second time each device is used full instructions will be provided on the correct usage with patients practicing with placebo either from the Respimat inhaler or from the MDI until they are judged competent. On these two test days the Respimat or MDI inhalers will be fired by the investigator one second after the patient has started to inhale. Thus on Test Days 1 and 2 patients will use their own natural inhalation technique without receiving any instruction on correct usage. On Test Days 3 and 4 patients will use a supervised optimal technique having received instruction on correct usage and with the investigator firing the device.

The primary analysis will be carried out using the Sign Test. This is a non-parametric analysis in which no assumptions are made about the shape of the distribution of the responses from the Respimat inhaler and from the MDI under the null hypothesis.

Study Hypothesis:

The null hypothesis is that poor technique has the same effect on the Respimat and MDI devices. The alternative hypothesis is that poor technique has a different effect on the Respimat inhaler than on the MDI. This means that under the null hypothesis the median of the differences between the Respimat inhaler and MDI pairs is zero i.e. the differences are equally likely to be positive or negative. Under the alternative hypothesis the median of the differences between the Respimat inhaler and MDI pairs is not zero i.e. the frequencies of the positive and negative signs are different.

Comparison(s):

Baseline comparability will be achieved by the use of a cross-over trial design with every patient receiving all four treatments and by ensuring at each test day that baseline lung function is within 15% of the value obtained at the first test day, pre-dose FEV1 is < 65% of predicted value and patients have abstained from inhaled bronchodilators for at least 4 hours prior to the visit. Treatment sequence will not be fitted as a term in the analysis of variance models.

Inclusion criteria:

COPD patients: FEV1 less or equal 65% pre FEV1 less or equal 70% of FVC

Exclusion criteria:

Patients with any upper respiratory infection in the past 14 days prior to the screening visit (visit 1) Patients with any unstable or life-threatening cardiac arrhythmia