A Controlled Trial of Ambroxol in Chronic Bronchitis

A Controlled Trial of Ambroxol in Chronic Bronchitis· Gordon H. Guyatt, M.D., F.C.C.P.;t Marie Townsend, B.A;* Farouk Kazim, M.A; and Michael T. Newhouse, M.D.§ Ambroxol is a mucolytic agent which is widely used in chronic bronchitis in Europe. We conducted a double-blind randomized controlled trial of ambroxol vs matched placebo

kbroxol (M ucosolvan) is widely used as a mucolytic agent in Europe. Experimental data suggest biologic reasons why the drug might be effective, and a single randomized controlled trial (ReT) has suggested that the drug may be of benefit in helping patients to For editorial comment see page 580

clear secretions. 1 However, the previous ReT used multiple outcome measures, only one of which was positive; no correction for the use of multiple outcomes was made. We therefore conducted a randomized trial of ambroxol in patients with chronic bronchitis who had difficulty raising secretions. METHODS

Sample Specification The target population for the study was patients with chronic cough and difficulty raising sputum. We identified two groups, or strata, of such patients. Inclusion criteria for both strata included the following: (1) Daily cough productive of sputum for at least the previous three months. (2) Difficulty expectorating sputum. (3) Compliance in the use of a diary card and peak flow meter during the initial run-in period. (4) A smoking history of at least ten pack-years. In addition, patients in the first stratum met the following criteria: (1) Forced expiratory volume in 1 second (FEV J greater than 70 percent of predicted. Ratio of FEV l to vital capacity (VC) greater than 0.7. (2) Currently smoking at least two packs of cigarettes per week. Patients were excluded from the first stratum if they were currently taking, or had previously taken on a regular basis, oral or inhaled bronchodilators. Patients in the second stratum had impaired respiratory function, defined as follows: (1) FEVlless than 70 percent of predicted and FEV l to VC ratio less than 0.7. ·From the Department of Clinical Epidemiology and Biostatistics and Department of Medicine, McMaster University, Hamilton, Ontario, Canada. This study was supported by Boehringer lngelheim. t Assistant Professor of Epidemiology and Biostatistics, and of Medicine. ;Research Associate in Epidemiology and Biostatistics. §Clinical Professor of Medicine. Manuscript received December 11; revision accepted February 13 Reprint requests: Dr. Guyatt, Department of Epidemiology, Room 3H7, 1200 Main Street West, Hamilton, Ontario, Canada LBN 3Z5

818

in 90 patients with chronic bronchitis and difficulty clearing secretions. It was concluded that there was no advantage to taking ambroxol.

Exclusion criteria for both groups included the following:

(1) Patients identified by their clinicians as having asthma or

atopy.

(2) Presence of life-threatening disease.

(3) Previous or present bronchial carcinoma or tuberculosis.

(4) Ongoing acute upper respiratory tract infection.

Stratum 1 patients were recruited primarily through media advertising. Stratum 2 patients were recruited primarily through a respiratory registry of all patients with FE VIless than 70 percent of predicted and FEV l to VC ratio less than 0.7 seen at three respirology group practices over the preceding three years.

Study Design Patients who accurately completed a symptom diary during a one week run-in period were randomized to receive ambroxol, 60 mg bid, or matched placebo during the subsequent four weeks. Investigators, patients, and research staff were blind to allocation. Measures of outcome were obtained at baseline, two and four weeks as follows: (1) Peak Bow measurements were obtained using a peak Bow meter each day in the early morning and evening. (2) The FEV l and forced vital capacity (FVC) and midexpiratory Bow rates (FEF 25-75) were obtained from the best of three expirations into a Collins water spirometer with a 420 microprocessor. (3) Using established principles, 2-4 a questionnaire measuring key symptoms was constructed. Questions focused on symptoms which ambroxol was thought to ameliorate: cough, sputum production, and difficulty expectorating sputum. Response options for each question were framed as a seven point Likert scale (for example, patients were asked to characterize their difficulty expectorating sputum as extremely difficult; very difficult; quite difficult. . . not at all difficult). Following the study, the subjects were asked to make a global rating of whether the medication had helped their cough, sputum, and difficulty expectorating. (4) A diary, a simplified version of the symptom questionnaire, was completed daily by each patient. The four-week treatment period was chosen on the basis of the prompt response to ambroxol seen in animal models and the prompt response reported in uncontrolled trials in human beings.

Statistical Methods At the start of the study, we did not know the variability associated with questionnaire scores, nor the minimal clinically important difference. Therefore, to determine sample size, we specified that if the treated group were as much as one standard deviation better off than the control group, we wished to be able to detect this difference. Setting alpha equal to 0.05 and beta equal to 0.10 would require 20 subjects per group. Since we wished to detect this difference if it existed in either patients with our without airflow obstruction, 20 patients per group were required for both sets of patients. The first three primary outcome variables were analyzed using a Ambroxol in ChronicBronchitis (Guyatt st aJ)

Table I-Baseline Characteristics Ambroxol Sex Age (yr) No. of cigarretts/day Duration of smoking (yr) FEV 1 ·Mean±SEM.

<20

20-40 >40

Placebo

Male 23/45

Male 24/45 56.6±2.1* 10 27

56.0± 1.7 11 31

34.7±2.2 2.14±0.15 L

35.6± 1.9 2.00±0.12 L

8

2

repeated measures analysis of variance of the difference scores between baseline and two and four week follow-up visits. The diary card was analyzed using a repeated measures analysis of variance. The proportion of patients reporting overall improvement was compared using a chi-square test. Confidence intervals (95 percent) around the differences between the two groups were calculated. For the proportion of patients responding, the number of subjects required to exclude a 25 percent risk reduction for ambroxol was calculated. 5 RESULTS

Because results were virtually identical in the two strata, they will be presented together. For some of the baseline characteristics and outcomes, data were missing from one or two subjects. The baseline characteristics of the 90 subjects who began the run-in period are presented in Table 1. The groups are comparable, and none of the small differences observed reached conven tional levels of statistical significance. Of the 45 subjects randomized to ambroxol, eight dropped out prior to completing the study. Four patients reported lack of time for participation; one lost interest; one patient experienced chest pain which necessitated breaking the code, one died suddenly of a myocardial infarction, and one developed a severe respiratory tract infection. Of the 45 subjects randomized to placebo, four dropped out prior to completing the study. One was noncompliant with both medication and follow-up

visits; two reported lack of time for participation; and one patient started a new medication and feared a drug interaction. The primary outcomes are presented in Table 2. Despite a small deterioration in mean FEVl in both groups (0.004 L in the ambroxol group, 0.021 L in the placebo group), and essentially stable fEf 25-75 (increase of 16 mIls in the ambroxol group, deterioration of 123 ml/s in the placebo group), questionnaire scores improved in the major areas in which benefit with ambroxol was anticipated. This improvement was maintained over the four weeks of the trial, but was closely comparable in the two groups. There were no statistically significant differences between active and placebo, nor were there substantial trends. Results of individual questions favored placebo as often as they favored ambroxol. Confidence intervals around the differences in the two groups always overlap zero (reflecting that none of the differences are statistically significant) and are narrow, effectively excluding a large benefit for ambroxol. Results were similar for outcomes not reported in Table 2, including other symptom questions and the daily diary. Eighteen of 36 (50 percent) patients receiving ambroxol reported overall improvement, as did 24 of 40 (60 percent) receiving placebo (p=0.89). The 95 percent confidence interval around the difference in the proportion improved is - 33 to 13 percent. Given these results, we would have required 25 subjects per group to detect a relative risk reduction of 25 percent. Thus, our results indicate that if ambroxol does yield a relative risk reduction in the proportion of patients not improved, that reduction is less than 25 percent. Five subjects receiving ambroxol reported experiencing a bad taste in their mouth, and one each had headaches, chest tightness, and a rash. Five patients receiving placebo reported side effects. One had

Table 2-Major Outcomes Ambroxol

Amount of cough Inability to cough up sputum Frustration over coughing Cough interfering with activities

Placebo

Baseline

Four Weeks

Net Effect

3.62*

4.11

3.80

Baseline

Four Weeks

Net Effect

0.59

3.42

3.97

0.55

4.23

0.43

4.04

4.67

0.63

4.51

4.29

0.78

4.51

4.91

0.40

4.67

4.18

0.51

4.76

5.37

0.60

Net Benefit (Confidence Interval) O.04t (-0.36 to -0.20 ( - 0.76 to +0.38 (- 0.27 to -0.09 (-0.67 to

0.45)* 0.37) 1.02) 0.50)

·Each symptom is rated on a scale of 1 to 7, where 1 represents worst function, and 7 represents optimal function. tThe net benefit is calculated by subtracting the difference between follow-up and baseline in the placebo group from the difference between follow-up and baseline in the ambroxol group-the number is positive when there is a trend in favor of active drug, negative when there is a trend in favor of placebo. *The 95 percent confidence interval around the net benefit indicates the possible range of "true" benefit. CHEST I 92 I 4 I OClOBER. 1987

819

muscle spasms, one sore legs, two swelling of the legs, and one a sensation of sputum in the throat. DISCUSSION

Ambroxol is a trans-4[2-amino-3,5-dibromo-benzyl, amino] cyclohexanol-hydrochloride which has been shown to increase the number and activity of type 2 pneumocytes.r" and thus, to increase surfactant levels and lecithinlsphyngomyelin ratio 1O-13 and mucociliary clearance'v" in a number of animal models. Mucociliary transport and sputum viscosity have improved with ambroxol in human studies,":" but these effects have not been consistently replicated. 18.19 In the present study, all measures of both symptoms and pulmonary function failed to show any trend in favor of ambroxol in subjects with bronchitis with, or without, chronic airflow limitation. Given the relatively small sample size, could the study result have been false negative? The proportion of patients reporting benefit from the medication was actually 10percent greater in the placebo group than in actively treated patients (60 and 50 percent, respectively). Confidence intervals around the difference in the proportion of patients improved is - 33 to 13 percent. That is, the results are consistent with a true added benefit of 33 percent with placebo over ambroxol or a true added benefit of 13 percent with ambroxol over placebo. These results, as well as the narrow confidence intervals around symptom scores, suggest that the results exclude all but a clinically trivial benefit for ambroxol. Ericsson et all have recently reported a double-blind RCT in which 11 of32 placebo treated patients, 18of31 patients treated with 60 mg of ambroxol daily, and 20 of 32 patients treated with 120 mg of ambroxol daily, reported improved chest symptoms. The difference between the placebo and high-dose ambroxol group was statistically significant (p = 0.046). However, Ericsson et al,' like our own study, showed no difference in any of the following: pulmonary function; daily diary monitoring amount, apparent viscosity, color, and difficulty expectorating sputum; end of period questionnaires concerning breathlessness and added lung sounds; and willingness to continue taking the medication. Because there were multiple measures of outcome, no statistical correction for the multiple measures was made, and the one positive finding showed a borderline level of statistical significance, it is plausible that the single positive result represents a chance finding. The present study provides strong evidence that ambroxol is of little or no benefit in patients with chronic bronchitis who have difficulty raising sputum, irrespective of the presence or absence of airflow limitation. The placebo response rate of 60 percent in this study reinforces the need for carefully controlled trials with rigorous double-blinding in the assessment 620

of mucolytic agents. ACKNOWLEDGMENTS: We would like to thank Drs. Geoff Worsely and Alan Davis for their role in the initial planning of this project; Drs. Stewart Pugsley, Freddy Hargreave, John Morse, Helen Ramsdale, Serge Puksa, Allan McLellan, and Roger Haddon for help with recruiting; and Sue Halcrow, Jana Keller, H. Denise Walker, Debbie Maddock, and Christine Meiklejohn for help with collection of the data. REFERENCES

1 Ericsson CH, Juhasz J, Jonsson E, Mossberg B. Ambroxol therapy in simple chronic bronchitis: effects on subjective symptoms and ventilatory function. Eur J Respir Dis 1986; 69: 248-55 2 Kirshner B, Guyatt GH. A methodological framework for assessing health indices. J Chron Dis 1985; 38:27-36 3 Guyatt GH, Bombardier C, Tugwell PX. Measuring diseasespecific quality of life in clinical trials. CMAJ 1986; 134: 889-95 4 Guyatt GH, Berman LB, Townsend M, Taylor DW Should study subjects see their previous responses? J Chron Dis 1985; 38: 1003-07 5 Detsky AS, Sackett DL. When was a 'negative' clinical trial big enough? How many patients you needed depends on what you found. Arch Intern Med 1985; 145:709-12 6 Cerutti ~ Kapanci Y. Effects of metabolite VIII ofbromexine (Na 872) on type II epithelium of the lung. Respiration 1979; 37: 241-51 7 Robertson B. Pharmacological stimulation of surfactant secretion and surfactant replacement. E ur J Respir Dis 1985; 67:63-70 8 Seefeld H~ Weiss JM, Eberhardt H. Stimulation of lung maturity: investigation of ambroxol in various animal models. Acta Physiol Hungarica 1985; 65:305-12 9 Sepulveda J, Valasquez BJ. Study on the influence of NA-872 (ambroxol) and dexamethasone on the differentiation of Clara cells in albino mice. Respiration 1982; 43:363-68 10 Elemer G, Kapanci Y. Morphological approach to surfactant secretion in the lungs with particular reference to ambroxol. Prog Respir Res 1981; 15:234-39 11 Egberts J, Fontijne ~ Wamsteker K. Indication of increase of the lecithin/sphingomyelin (US) ratio in lung fluid in lambs maternally treated with metabolite VIII of Bisolvan. Biol Neonate 1976; 29:315-22 12 Velasquez BM, Sepulveda J. Influence of transplacental NA-872 and dexamethasone therapy on fetal and newborn mouse lung. Respiration 1982; 43:389-400 13 Lachmann B, TIscher AB, Grossmann G, Robertson B. Lung compliance and alveolar expansion in the artificially ventilated premature newborn rabbit after maternal treatment with ambroxol. Respiration 1981; 42:209-16 14 Iravani VJ, Melville GN. Wirkung von Bromhexin-Metabolit VIII and einem neuen adrenergen Stoff auf die mukoziliare funktion des respirationstraktes, Drug Res 1974; 24:849-55 15 Allegra L, Bossi R, Braga P Influence of surfactant on mucociliary transport. Eur J Respir Dis 1985; 142:71-76 16 Olivieri D, Marsico SA, Del Donno M. Improvement of mucociliary transport in smokers by mucolytics. Eur J Respir Dis 1985; 66:142-45 17 Grassi C, Luisetti M, De Rose FiettaA, Sacchi F, Genghini M, et al. Biomedical and functional changes in bronchoalveolar parameters induced by ambroxol treatment of chronic bronchitis. In: Cosmi E~ Scarpelli EM, eds. Pulmonary surfactant system. Elsevier Science Publishers, 1983 18 Weiss 1: Dorow ~ Felix B. Effects of a beta adrenergic drug and a secretolytic agent on regional mucociliary clearance in patients with COLD. Chest 1981; 80:881-85 19 Shimura S, Okubo 1: Maeda S, Aoki 1: Tomioka M, Shindo Y, et al. Effect of expectorants on relaxation behavior of sputum viscoelasticity in vivo. Biorheology 1983; 20:677-83

v

Ambroxol in ChronicBronchitis (Guyatt st 8/)

I