Study design
This was a 26-week, two arm, open label, parallel groups study. Subjects were randomized to the Lung Flute or usual care. The clinical trial registration number of this trial is NCT01186822. The study was approved by the Human Studies subcommittee for the Department of Veterans Affairs (VA) Western New York Healthcare system. All participants provided written informed consent prior to any study procedures.
Subjects
We had planned to enroll 80 subjects (see power analysis below) with COPD with chronic bronchitis at a single center (Buffalo VA Medical Center). Inclusion criteria were: a) between 30-80 yrs of age, b) presence of airflow obstruction by spirometry post-bronchodilator forced expiratory volume in 1 second (FEV1) to forced vital capacity (FVC) ratio <70% and FEV1 < 80% predicted, GOLD Stage 2-4), c) presence of chronic bronchitis, i.e. cough productive of sputum on most days of the week, d) current smoker or ex-smoker with at least 10 pack yrs of smoking history, e) able to vibrate the reed of the Lung Flute® using the standard therapeutic maneuver. Exclusion criteria were a) exacerbation of COPD within 4 weeks prior to enrollment, b) predominant asthma and bronchiectasis by clinical assessment, c) pregnant or nursing women d) chronic use of a mucolytic medication.
Procedures
The study consisted of a screening and randomization visit and then on treatment clinic visits at 2, 14 and 26 weeks. All visits were performed on an outpatient basis. In addition, standardized telephonic assessments were made at 8 and 20 weeks. Participants who met the inclusion/exclusion criteria were randomized to either the Lung Flute or usual care on a 1:1 allocation basis. Randomization was by predetermined random sequence generated independently and kept in sealed envelopes until the time of randomization.
At baseline prior to randomization, symptoms were measured with the CCQ and health status with the St. George's respiratory questionnaire (SGRQ) [9]-[11]. Lung function was assessed by post-bronchodilator spirometry, exercise capacity by the six minute walk test, dyspnea by the measurement of the modified Medical Research Council (mMRC) score and body mass by calculating the body mass index (BMI). These measurements were used to calculate the BODE index as has been described earlier [12].
At each subsequent clinic visit, all the above evaluations were repeated and review of concomitant medication, compliance assessment, adverse event surveillance and exacerbation history were performed. The telephone calls consisted of a standardized assessment of concomitant medication, compliance assessment, adverse event surveillance and exacerbation history.
Intervention
The Lung Flute arm participants were instructed to blow twice in to the Lung Flute device vigorously enough to make the reed oscillate, followed by 5 normal breaths. This was repeated 10 times, followed by 3 huff coughs to complete 1 cycle. Two such cycles were recommended twice a day. One of these cycles was performed under supervision of the study personnel at the time of enrollment and at each subsequent study visit. Baseline COPD medication regimen was continued in all participants, although the primary physicians of the participants could make medically necessary changes. Chest physical therapy, additional breathing exercises and formal pulmonary rehabilitation programs were not prescribed to any of the participants during the study.
Endpoints
The primary endpoint of this study was comparison of the change in COPD symptoms assessed by the CCQ questionnaire between the intervention and control group at 26 weeks. Secondary endpoints included comparison of the changes in SGRQ score, spirometric lung function, BODE index, and exacerbation frequency at 26 weeks between the 2 arms.
Data analysis
All analyses were performed in an intention to treat (ITT) manner. Subject demographics were compared with t tests and chi square analyses as appropriate. For normally distributed outcomes (e.g. CCQ and SGRQ scores), paired t-test was applied to compare week 26 (week 14/ week 2) to baseline for the two arms, respectively. Also 2-sample independent t-test is used to compare the two arms at baseline, and to compare the changes of week 26 (week 14/ week 2) from baseline in the two arms. For outcomes which failed the normality test (e.g. BODE score), Wilcoxon signed ranks test is applied to compare week 26 (week 14/ week 2) to baseline for the two arms, respectively. Also Wilcoxon-Mann Whitney test is used to compare the two arms at baseline, and to compare the changes of week 26 (week 14/week 2) from baseline in the two arms. Fisher's exact test is applied to compare the exacerbation frequencies of the two arms. A p < 0.05 was considered significant. Missing data were imputed by carrying forward the last observation.
Power analysis
In order to determine sample size, we used the primary outcome of the CCQ score. In previous 8 week studies with Lung Flute, an average improvement of 0.4 (S.D. of 0.64) in the CCQ score has been seen with this device. We assumed that the control group will have a change of .05 in the CCQ score. With an alpha = 0.1 (one sided t test) and power = 0.8 and a 1:1 allocation, a total sample size of 76 was required. Allowing for a 5% drop-out rate, a total of 80 patients were to be included in this study.