L. AUegra', N. Konietzko*, P. Leophonte% J. Hosie', R. Pauwels', J. N. Guyei/ and P. Petitpretz'

Journal of Antimicrobial Chemotherapy (1996) 37, Suppl. A, 93-104 Comparative safety and efficacy of sparfloxacin in the treatment of acute exacerbations of chronic obstructive pulmonary disease: a double-blind, randomised, parallel, multicentre study L. AUegra', N. Konietzko*, P. Leophonte% J. Hosie', R. Pauwels', J. N. Guyei/ and P. Petitpretz' "I.R.C.C.S. Ospedale Maggioze, via Francesco Sforza, 35, 20122, Milan, Italy; b Arztf. Inn. Med. und Pneumologie-Allergologie, Chefarzt der Ruhrlandklinik Tiischener Weg 40, 4300, Essen 16, Germany; 'Service de Pneumologie, Hopital Rangeuil, 31052, Toulouse, France; 'Great Western Medical, 1980, Great Western Road, Glasgow, G13 2SW, UK; 'Department Respiratory Diseases, University of Gent Hospital, De Pintelaan 185, 9000 Gent, Belgium; f Hopital de la Salpetriere, Laboratoire de Bacteriologie, 47 boulevard de VHopital, 75013, Paris, France; 'Service de Pneumologie, Hopital Andre Mignot, 177 avenue de Versailles, 78157, Le Chesnay Cedex 11, France In a double-blind, placebo-controlled trial, patients with acute exacerbations of chronic obstructive pulmonary disease (COPD) were randomly allocated to oral treatment with sparfloxacin (200 mg loading dose followed by 100 mg once daily) or amoxycillin/clavulanic acid (500 mg/125 mg tds) for a total treatment duration of 7 to 14 days. Patients were evaluable if they had a FEV,/FVC ratio of less than 70% at stable state and presented with a suspected infectious exacerbation defined as increases in dyspnoea, sputum volume and sputum purulence. The primary efficacy variable was the overall success (defined as disappearance or improvement of dyspnoea and reductions in sputum volume and purulence) at end of treatment and follow-up. Overall efficacy was assessed in both the intent-to-treat (728 patients) and the evaluable (351 patients) populations. At the end of treatment and follow-up, success rates were identical for the sparfloxacin (87.3% and 78.7%) and amoxycillin-clavulanic acid (88.8% and 79.8%) treatment groups. Similar figures were found for the intent-to-treat population. The analysis of drug safety was similar for both treatment groups. The most frequently encountered pathogens were Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis. Sparfloxacin appeared superior for bacteriological eradication of Haemophilus influenzae, and Moraxella catarrhalis. Sparfloxacin in a single daily dose appears at least as effective as amoxycillin/clavulanic acid in the treatment of patients with acute exacerbations of COPD. Introduction Patients with chronic obstructive pulmonary disease (COPD) regularly develop acute exacerbations (Isada & Stoller, 1994). These exacerbations may be infectious, inflammatory or both, and the respective roles are difficult to distinguish in individual patients. Whatever the mechanism, most patients are symptomatic and demonstrate 93 0305-7453 96'37AO93 + 12 $12.00 0 f 1996 The British Society for Antimicrobial Chemotherapy

94 L. Allegra et at. increased dyspnoea as well as sputum volume and purulence (Chodosh, 1992). The simultaneous presence of these three symptoms in patients with obstructive disease was defined by Anthonisen et al. (1987) as a Type 1 exacerbation of COPD for which antibiotic therapy was shown to be useful. Although precise microbiological diagnosis may be difficult due to the normal bacterial colonization of the lower airways in COPD patients, Haemophilus influenzae, Streptococcus pneumoniae and Moraxella catarrhalis have been the most frequently implicated pathogens (Isada & Stoller, 1994). Many antibiotic regimens have been suggested for the treatment of acute exacerbations of COPD (Isada & Stoller, 1994), including the co-amoxiclav (amoxycillin/clavulanic acid) combination which is, at least in Europe, considered as a standard regimen. Fluoroquinolone antibiotics such as ofloxacin and ciprofloxacin have also been used, but their activity against many strains of streptococci, including S. pneumoniae is limited (Canton et al., 1992). In contrast with older fluoroquinolones, sparfloxacin is active against S. pneumoniae demonstrating an MICo of 0.25 mg/l (Cooper et al., 1990; Chin et al., 1991; Canton et al., 1992). It is also active against M. catarrhalis and H. influenzae (MIG^s < 0.1 mg/l) and other pathogens that may be involved, although less frequently, during infectious exacerbations of COPD, such as staphylococci and Enterobacteriaceae, and intracellular pathogens such as Mycoplasma pneumoniae and Chlamydia pneumoniae (Cooper et al., 1990; Chin et al., J991). The objective of this controlled trial was to compare the safety and efficacy of sparfloxacin with that of co-amoxiclav in the treatment of patients with a suspected acute bacterial exacerbation of COPD. Study design Materials and methods This was a multicentre, randomised, double-blind, comparative study of patients with an acute exacerbation of COPD which was conducted between October 1990 and November 1991. Study population Patients of both sex (excluding female patients who were pregnant or lactating) aged at least 18 years old could be enrolled in the study if they had: a history of chronic bronchitis defined as sputum-productive cough for at least 3 months of the year for at least the previous 2 years, emphysema or asthma; symptoms of acute exacerbation defined as an increase in dyspnoea, sputum purulence (with greater than 25 PMN per low power field (or five per high power field) with less than ten squamous epithelial cells per low power field in the examined samples) and sputum volume (placing the patient in Anthonisen's Type 1 exacerbation (Anthonisen et al., 1987); an FEV,/FVC ratio of less than 70%; and a chest X-ray showing no evidence of pneumonia on admission. All patients were to have pre-existing stable signs and symptoms of chronic pulmonary disease. If a stable baseline FEV,/FVC ratio was not available for a patient before the entry exacerbation, this parameter was to be measured 2 months after completion of the study and when the patient had been free of exacerbation for 1 month. Cultures of expectorated sputum, or samples obtained by protected brush fibreoptic

Sparfloxacin in exacerbations of COPD 95 bronchoscopy or transtracheal aspiration were obtained before beginning treatment. Culture results were not available before study inclusion. To be evaluable, pathogens from sputum were to have a colony count of at least 10 7 cfu/ml (10*cfu/mL for 5. pneumoniae) and those from protected brushing or transtracheal aspiration a colony count of at least 10 3 cfu/ml. Identities of isolates were verified in most cases by reference laboratories (Central Microbiological laboratory, Pitie-Salpetriere Hospital, Paris, France or Central Microbiological laboratory, University of Milan, Italy). The reference laboratories also performed the MIC determinations on the pathogens isolated. Exclusion criteria were as follow: presence of bronchiectasis; severe concomitant disorders likely to interfere with clinical course (such as cystic fibrosis, uncontrolled cardiovascular disorders, cancer or AIDS); hepatic disease (AST or ALT greater than twice or total bilirubin greater than 1.5-times the upper limit of normal); renal failure (serum creatinine greater than 170 //mol/l); anaemia (haemoglobin less than 10 g/dl); history of photosensitivity or allergy to quinolones or /J-lactams; serious psychiatric disease; and concomitant administration of iron salts, allopurinol, or administration of any investigational drug within 1 month of study entry. Patients were advised to avoid exposure to sunlight and tanning salons for the duration of treatment. Study medications Patients were randomly assigned to receive sparfloxacin plus a matching placebo or co-amoxiclav plus another matching placebo. Sparfloxacin was administered as a loading dose of 200 mg the first day, followed by 100 mg once daily. Co-amoxiclav 5OOmg/125mg was administered thrice daily. Study medications, whether containing active drug or placebo, were taken thrice daily (morning, mid-day and evening) with a glass of water. Treatment duration was 7-14 days at the discretion of the investigator, with 10 days being the protocol-recommended duration of treatment. Medications were dispensed according to a randomized code and balanced for each study centre in blocks of four. This study was conducted in accordance with the declaration of Helsinki (Hong Kong amendment) and the protocol was approved by the Ethics Committees of all participating institutions. Written or witnessed informed consent was obtained from all patients. Study conduct and assessment of outcome An external Scientific Review Committee provided advice on conduct of the study, reviewed blinded data for individual patients, and made decisions (also blinded) concerning the evaluability and efficacy classifications of these patients. Patients were assessed by the investigators on admission (visit 1), after 4 days of treatment (visit 2), at the completion of 7-14 days of treatment (visit 3) and at a follow-up visit 10 days after the completion of treatment (visit 4). A complete medical history and physical examination (including peak flow pulmonary function testing) were performed at visit 1 and the physical examination repeated at visits 3 and 4. Sputum cultures (by the same method as for inclusion, if possible) were to be repeated at visits 2, 3 and 4. Scales were used for the recording of sputum volume (none, less than 30 ml, 30-100 ml and greater than 100 ml), sputum purulence (none; mucoid; mucopurulent and purulent) and dyspnoea (none; only on exertion, during normal activity and present at rest).

96 L. Allegra et al. Additionally, patients were asked to complete a diary card recording sputum volume and appearance, dyspnoea symptoms, results of personal peak flow testing and morning body temperature. Efficacy analyses were performed for two distinct populations: evaluable patients for primary efficacy analysis (Anthonisen Type 1 patients), and the intent-to-treat efficacy population. To be evaluable for the primary efficacy analyses, it was necessary that the patient fulfil all inclusion criteria (FEV,/FEVC ratio less than 70% at stable state and an increase in dyspnoea, sputum volume and purulence at entry), have no exclusion criteria, and have received at least 4 days of treatment with study medication except in the case of failure. The intent-to-treat study population included all patients who received at least one dose of treatment and had a history of chronic bronchitis. Efficacy analyses compared two intervals: from study entry (visit 1) to end of treatment (visit 3) and from study entry (visit 1) to follow-up (visit 4). Analyses for these two intervals were separate except that any patient considered a treatment failure at visit 3 was automatically considered a treatment failure at visit 4. Three basic efficacy analyses were performed: an evaluation of overall efficacy according to clinical efficacy, irrespective of microbiological results (except in the case of persistence of Pseudomonas aeruginosa); microbiological response only as assessed by bacterial culture; and clinical response only as assessed by the symptoms of sputum volume and purulence and dyspnoea. Thus, any clinical success was considered an overall success irrespective of microbiological results (except for persistence of P. aeruginosa which was considered as indicative of failure) and any clinical failure was considered an overall failure, irrespective of microbiological results. A clinical success was defined as cure (return to baseline stable state in all three clinical symptoms) or improvement (decrease in all three symptoms or return of both sputum volume and purulence to baseline stable state with dyspnoea unchanged). Clinical failure was defined as persistence, worsening or relapse of clinical symptoms. Bacteriological cure required eradication or presumed eradication of pathogens. Microbiological eradication was defined as disappearance of all pathogens from protected brushing or transtracheal aspiration, or a decrease in pathogen counts isolated from sputum to less than 10 3 cfu/ml. Presumed eradication was defined as the absence of source for repeat culture because patient was no longer producing sputum or because repeat bronchoscopy or transtracheal aspiration was not justified in a clinically cured or improved patient. Safety analysis All patients who entered the study and who received at least one dose of study medication were included in the analysis of safety. Adverse experiences were recorded at each visit after admission and were classified by the investigator according to the severity and relation to study medication. Changes in the results of haematology, biochemistry or urinalysis testing were monitored during the study. Statistical analysis All values are expressed as the mean ± S.E.M. Analyses were performed for the intent-to-treat population and for evaluable patients. According to the guidelines for

Sparfloxacin in exacerbations of COPD 97 the evaluation of anti-infective drugs (Chow et al., 1992), an equivalence approach was used based on a two-sided 90% confidence interval (CI) of the difference between success rates of the co-amoxiclav group and the sparfloxacin group. The objective was to demonstrate that the co-amoxiclav success rate was not greater than the sparfloxacin success rate by more than 10%. Therefore, only the upper limit of the 90% CI of the difference was considered for the statistical interpretation of the results. The 90% CI indicated that there was a 90% probability that the true difference between treatments lies within the interval, and more particularly that the probability of a true difference greater than the upper limit in favour of amoxycillin/clavulanic acid was only 5%. This approach was used for analysis of overall success rate and bacteriological cure rates at both visits 3 and 4. All statistical analyses were carried out with SAS software package (SAS Institute, Cary, NC, USA). Population description Results Of 734 enrolled patients, 733 received the study medication. Three hundred and seventy patients were assigned to sparfloxacin and 363 to co-amoxiclav. Five patients were excluded from intent-to-treat analysis because they had no diagnosis of chronic bronchitis. Three hundred and sixty-six patients were non-evaluable for overall primary efficacy (FEV,/FVC greater than 70%, lack of increase of all three clinical symptoms, or presence of bronchiectasis were the commonest reasons for exclusion). Thus, 362 patients with Anthonisen's Type 1 exacerbation were initially evaluable for the overall efficacy analysis. Eleven and 35 patients were non-evaluable at visits 3 and 4, respectively. Thus a total of 351 patients (173 in the sparfloxacin group and 178 in the co-amoxiclav group) were evaluable at visit 3 (end of treatment). A total of 327 patients (164 in the sparfloxacin group and 163 in the co-amoxiclav group) were evaluable at visit 4 (follow-up). Demographic and main clinical characteristics of the overall study population at inclusion appear in Table I. No significant differences were observed between the treatment groups. The average frequency of acute exacerbations of COPD over the previous 2 y was also similar in both study groups; most patients (64.4%) had experienced more than three such exacerbations. The number of days since the last episode of exacerbation varied from less than 30 days to greater than 150 days (less than 4 months in 59% of patients), but was similar in both groups. The proportions of never, previous and current smokers were also similar in the two groups (approximately 33% in each category). For the majority of patients, the current exacerbation had been present for 1 to 6 days, with no difference existing between study groups. The FEV,/FVC ratio was similar in both study groups at stable state and during the current exacerbation. A FEV,/FVC less than 70% was an inclusion criterion for COPD and therefore 100% of evaluable patients with available data (n = 350) had a ratio less than 70% at steady state. Twelve additional patients with no stable state FEV,/FVC had determinations performed after cure and were included in this evaluable population with Anthonisen's Type 1 exacerbation (leading to a total number of 362 patients). In this population, the mean FEV,/FVC was 54% ± 1. Clinical symptoms at baseline were similar in both treatment groups except for dyspnoea which was more frequent at rest

98 L. Allegra et al. or during normal daily activity in the sparfloxacin group (19.7%) than in the co-amoxiclav group (14.9%). Previous or concomitant therapy was common (90% of patients) in both study groups. The most frequently used medications were 0-2 agonists, corticosteroids, theophylline and ipratropium bromide. Antibiotics had been used to treat the current exacerbation in 31 patients in the overall study population (20 sparfloxacin and 11 co-amoxiclav) before inclusion in the study. Sixteen evaluable patients (11 sparfloxacin and 5 co-amoxiclav) received antibiotics for the current exacerbation before study inclusion. The assigned duration of treatment was most frequently 10 days in both study groups (42% of sparfloxacin patients and 39% of co-amoxiclav patients) and was similar in both treatment groups for all intervals from 7 to 14 days. The mean (and range) actual treatment durations were 9.99 (1-16) for sparfloxacin and 10.34 (1-15) for co-amoxiclav. Microbiology Where the method of culture was specified (462 of 734 cases), cultures were obtained from expectorated sputum in 96% of cases, by bronchoscopy in 1% and by bronchial aspiration in 3%. Less than half the patients in both groups (39% sparfloxacin and 42% co-amoxiclav) had a pathogen present at baseline (Table II). The percentage of patients with a pathogen on admission was somewhat higher in evaluable patients Table I. Clinical findings on admission (overall study population) Age (y) Sex (M/F, %) Respiratory rate (per min) Temperature ( C C) Underlying respiratory disorder asthma emphysema chronic bronchitis bronchiectasis FEV,/FVC at stable state (n = 654) < 70% > 70% FEV,/FVC at inclusion (n = 519) < 70% > 70% Peak flow (L/min) at stable state (n = 527) at inclusion (n = 662) Chest X-ray (normal) Chest X-ray (abnormal) radiodensity of the bronchial tree other abnormalities Mean values ± S.E.M. Sparfloxacin (n = 370) 60 ±0.7 64.2/35.7 20.2 ± 0.3 37.2 ± 0.04 128 (34.6%) 119 (32.2%) 324 (87.6%) 18 (4.9%) 218 (68.1%) 102 (31.9%) 176 (67.7%) 84 (32.3%) 322 ±9 256 ±6 98 (26.5%) 200 (54.1%) 192 (45.9%) Co-amoxiclav (n = 363) 60.8 ± 0.7 69.1/30.9 20.2 ± 0.3 37.2 ± 0.04 127 (35.0%) 119 (32.8%) 316 (87.1%) 13 (3.6%) 232 (69.5%) 102 (30.5%) 175 (67.3%) 85 (32.7%) 323 + 9 260 ±6 92 (25.3%) 202 (55.6%) 190 (44.4%)

Sparfloxacin in exacerbations of COPD 99 Table II. Microbiological findings at inclusion (overall study population) Sparfloxacin Co-amoxiclav Number of samples received (% of patients) 370 (100%) 363 (100%) Number of pathogens isolated 0 225 (60.8%) 210 (57.9%) 1 120(32.4%) 112(30.9%) 2 2 (5.9%) 33 (9.1%) 3 3 (0.8%) 8 (2.2%) Number of pathogens present (% of strains) 173 (100%) 202 (100%) H. influenzae 53 (30.6%) 51 (25.2%) M. catarrhalis 23 (13.3%) 20 (9.9%) S. pneumoniae 49 (28.3%) 49 (24.3%) Staphvlococcus aureus 10(5.8%) 11(5.4%) Enterobacteriaceae 17 (9.8%) 38 (18.8%) P. aeruginosa 19(11.0%) 23(11.4%) 0-Haemolytic streptococci (Groups A, C, G) 1 (0.6%) 4 (2.0%) Other 1 (0.6%) 6 (0.3%) (sparfloxacin 50% and co-amoxiclav 53%). The commonest pathogens were H. influenzae, S. pneumoniae and Enterobacteriaceae, followed by M. catarrhalis and P. aeruginosa (Table II). At study inclusion, 82 of the 104 isolates of H. influenzae had a MIC determination. The MIC90 values for sparfloxacin and co-amoxiclav were 0.03 mg/l (range 0.0075 to 0.25 mg/l) and 1.0 mg/l (range ^0.06 to 64 mg/l), respectively. Only 84% of the strains (69/82) were sensitive to co-amoxiclav at 0.5 mg/l or less; five strains were sensitive at 1 mg/l; and eight at 2 mg/l or more. MICs were determined in 83 and 81 isolates of S. pneumoniae for sparfloxacin and co-amoxiclav, respectively. MIG» for sparfloxacin was 0.5 mg/l (range 0.0075 to 4.0 mg/l), and three strains had an MIC of 2.0 mg/l or higher. The MIGw for co-amoxiclav was 4.0 mg/l (range ^0.06 to 32 mg/l), and only 84% of strains (68/81) were inhibited by a co-amoxiclav concentration of 0.5 mg/l or less. The MICs were determined in 37 strains of M. catarrhalis. The MIC90 was 0.03 mg/l for sparfloxacin (range 0.0075 to 2.0 mg/l) with two strains which demonstrated an MIC of 2.0 mg/l. The MIC W was 1.0 mg/l for co-amoxiclav (range <0.06 to 32 mg/l) with 86% of strains (32/37) sensitive at concentrations of 0.5 mg/l or less, and five strains at concentrations of 1 mg/l or more. Overall, 407 of 463 (87%) strains were sensitive to sparfloxacin concentrations of 0.25 mg/l or less, and 280 of 461 (61%) were sensitive to co-amoxiclav concentrations of 0.5 mg/l or less. Overall efficacy For the intent-to-treat study population, the rate of overall success was almost identical in both the sparfloxacin and co-amoxiclav groups at the end of treatment and had declined to a similar extent in both groups at the follow-up visit (Table III). For evaluable patients the rates of success at the end of treatment and relapse at follow-up were 7.5 and 7.3% in the sparfloxacin and co-amoxiclav groups, respectively. The effect of various pathogens on overall success rates appears in Table IV. Sparfloxacin appeared to have superior overall success rates compared with

31/34 (91%) 21/27 (78%) 12/18 (67%) 6/10 8/8 4/4 1/1 83/102 (81.4%) 18/30 (60%) 26/33 (79%) 10/14 (71%) 8/16 17/19 (89%) 1/3 2/2 2/2 84/123 (68.3%) Table 111. Overall and clinical efficacy in the intent-to-treat and evaluable patient populations Overall efficacy intenlion-to-treat end of treatment follow-up Overall efficacy evaluable patients end of treatment follow-up Clinical efficacy intention-to-treat at follow-up cured improved Clinical efficacy evaluable patients at follow-up cured improved Sparfloxacin 309/369 (83.7%) 273/369 (74.0%) 151/173 (87.3%) 129/164 (78.7%) 234/302 (77.5%) 46/302 (15.2%) 111/154 (72.1%) 30/154 (19.5%) Co-amoxiclav 310/359 (86.4%) 276/359 (76.9%) 158/178 (88.8%) 130/163 (79.8%) 238/286 (83.2%) 34/286 (11.9%) 119/149 (79.9%) 20/149 (13.4%) Statistical parameter (90% Cl) (-1.73; 6.95) (-2.3; 8.1%) (-4.2; 7.2%) (-6.3; 8.5%) Table IV. Overall clinical efficacy at the end of treatment and at follow-up according to isolated pathogen in evaluable patients sparfloxacin End of treatment co-amoxiclav sparfloxacin Follow-up co-amoxiclav s H. influenzae S. pneunwniae Si. calarrhalis P. aeruginosa Enterobacteriaceae S. aureus /J-Haemolytic streptococci (Group A. C, G) Other Total 33/35 (94%) 24/27 (89%) 15/18 (83%) 7/10 8/8 5/5 1/1 93/104 (89.4%) 25/32 (78%) 32/35 (91%) 12/14 (86%) 8/16 20/21 (95%) 2/3 2/2 2/2 103/125 (82.4%)

Sparfloxacin in exacerbations of COPD 101 co-amoxiclav in patients with H. influenzae-related infections. Overall success rates were comparable in patients from whom M. catarrhalis, S. pneumoniae and Enterobacteriaceae were isolated. Both drugs demonstrated low overall success rates in patients with P. aeruginosa-rdated infections. This could be explained by the decision of the Scientific Review Committee to consider all patients with persistent P. aeruginosa colonisation as treatment failures. In the evaluable study population, overall success rates by bacteriological status at admission (positive vs negative) were identical (78.6% vs 78.8%) for sparfloxacin-treated patients at follow-up. In contrast, overall success rates for the co-amoxiclav group were markedly lower in patients who were bacteriologically positive at admission (69.2% vs 93.1%). Clinical efficacy The three primary symptoms (dyspnoea, sputum purulence and volume) improved markedly over the course of treatment in both groups. Sputum purulence improved in 85.5% and 92.9% of evaluable patients treated with sparfloxacin at the end of treatment and at follow-up, respectively. The corresponding values were 79.2 and 87.2% in the co-amoxiclav group. Similar results were observed with sputum volume which decreased in 75.7 and 87.7% of patients treated with sparfloxacin at the end of treatment and at follow-up, respectively, compared with 69.1 and 83.9% of patients treated with co-amoxiclav. However, persistent dyspnoea was observed more frequently in the sparfloxacin group (9.2 and 3.9% of patients) than in the co-amoxiclav group (7.9 and 0.7%) at the end of treatment and at follow-up, respectively. Overall and clinical efficacy results appear in Table III. Clinical success (cure and improvement) was identical in both treatment groups at follow-up for both the intent-to-treat and evaluable populations. Overall clinical efficacy at the end of treatment and at follow-up solely according to isolated pathogens in evaluable patients is reported in Table IV. Microbiological efficacy Eradication/presumed eradication rates for all bacteriologically evaluable patients and for clinically evaluable patients who were also bacteriologically evaluable were somewhat higher for sparfloxacin (108/125 = 86.4% amd 75/85 = 88.2%) than for co-amoxiclav (101/123 = 82.1% and 62/79 = 78.5%), respectively. Sparfloxacin eradicated 100% (35/35), 89% (16/18) and 82% (22/27) of//, influenzae, M. catarrhalis and 5. pneumoniae isolates, respectively. The corresponding values for co-amoxiclav were 73% (19/26) 80% (12/15) and 86% (24/28). Eradication rates for Enterobacteriaceae at the end of treatment were low for both drugs (7/12 and 14/26 for sparfloxacin and co-amoxiclav, respectively) but were higher at follow-up (11/12 and 16/20, respectively). In contrast, eradication rates for P. aeruginosa were low in the sparfloxacin treatment group (8/11 and 6/16) and were 6/9 and 11/13 in the co-amoxiclav treatment group, at the end of treatment and at follow-up, respectively. Analysis of failures At follow-up, P. aeruginosa and S. pneumoniae accounted for almost all (7/8) sparfloxacin bacteriological failures, but their respective MIC«s were indicative of susceptibility to sparfloxacin in all of these cases. A wider variety of organisms,

102 L. AUegra et al. primarily Gram-negative, accounted for co-amoxiclav failures (17/24), and MIC90S for these organisms were indicative of resistance to this compound in four cases. Persistence of P. aeruginosa, denned as treatment failure, was observed in four patients in the sparfloxacin group and two patients in the co-amoxiclav group. Treatment failures generally responded well to a different treatment (usually another antibiotic with or without other medications). Safety Overall, 129 adverse events were reported by 81 (21.9%) sparfloxacin-treated patients compared with 149 adverse experiences reported by 93 (25.6%) patients in the co-amoxiclav group. The most frequently reported specific adverse events (1 to 3%) in the sparfloxacin versus co-amoxiclav groups were diarrhoea (1.1% vs 3.6%), nausea (1.1% vs 2.2%), vomiting (1.1% vs 1.7%) and rash (1.6 vs 0.8%). Headache occurred in 1.1% of patients receiving sparfloxacin and 0.3% of those treated with co-amoxiclav. A photosensitivity reaction occurred in one sparfloxacin patient. Fifteen per cent of adverse events were classified as severe in each treatment group. The rate of adverse events considered to be probably related to the study drug was higher in the co-amoxiclav group (22%) than in the sparfloxacin group (14%). Treatment was discontinued because of adverse events in 2.7% of patients treated with sparfloxacin and in 2.5% of those receiving co-amoxiclav. Three patients in the sparfloxacin group and one in the co-amoxiclav group died during the study. No death was considered related to the study drugs. Laboratory abnormalities occurred with the same frequency in both groups and all were classified as both mild and non-serious. Discussion This study compared the efficacy of sparfloxacin, a new amino fluoroquinolone with an extended antibacterial spectrum against Gram-positive cocci and S 1. pneumoniae, with that of co-amoxiclav in the treatment of acute exacerbations of COPD patients. To be evaluable, patients were to have chronic airway obstruction (defined as a FEVi/FVC ratio less than 70%) and to present with Anthonisen Type 1 exacerbation (Anthonisen et al., 1987), that is to have worsening of all three clinical symptoms of dyspnoea, sputum volume and sputum pumlence. This identifies a population of COPD patients most likely to derive benefit from antibiotic therapy. Patients with worsening of only two symptoms (Anthonisen Type 2 patients) or patients with FEV,/FVC ratios greater than 70% were excluded from the primary group of evaluable patients although many of these excluded patients would in fact have had a true exacerbation of chronic obstructive pulmonary disease. Thus, 351 patients were evaluable for overall efficacy. This proportion of evaluable patients is quite high considering the very rigid inclusion/exclusion criteria. Notwithstanding, the intention-to-treat analysis revealed no important difference from the results observed in the evaluable population. Overall success rates in evaluable patients of 87.3% for sparfloxacin and 88.8% for co-amoxiclav at the end of treatment, falling to 78.7% and 79.8%, respectively, at follow-up, show an equivalent efficacy of both compounds in this study. The overall classification of success was based primarily on clinical cure or improvement. Indeed, as pointed out by Anthonisen et al. (1987), symptoms are the criteria applied in medical practice to define onset of an acute exacerbation and that resolution should be defined

Sparfloxacin in exacerbations of COPD 103 in similar terms. The mean age of the population (about 60 y), its male predominance, and the frequency of past or present tobacco use resemble the patient group described by Anthonisen et al. (1987). Fifty-nine per cent of the intent-to-treat study population and 48% of the evaluable population had no pathogen isolated at entry. The types of isolated pathogens were comparable to those reported in other studies (Anthonisen et al., 1987; Isada & Stoller, 1994). The long elimination half-life of sparfloxacin (approaching 20 h) allows the maintenance of high plasma concentrations for prolonged periods, even with once-daily administration (Shimada, Nogita & Ishibashi, 1993). Moreover, sparfloxacin concentrations in bronchial mucosa biopsies, epithelial lining fluid and alveolar macrophages are several to many fold higher than serum concentrations, and these concentrations persist for at least 24 h (Honeybourne et al., 1994). It is interesting to note that the overall efficacy in this study (79.2% at follow-up for both groups combined) is higher than the 62.9% efficacy rate reported by Anthonisen et al. (1987). This difference may be the result of improved antibiotic efficacy. In considering clinical efficacy, sparfloxacin was efficient in improving both sputum volume and purulence. However, dyspnoea persisted more frequently in these patients than in those who received co-amoxiclav. This observation may have been due to the slightly more severe dyspnoea score at baseline in patients randomised to receive sparfloxacin. Nevertheless, sparfloxacin appeared superior to co-amoxiclav in terms of microbiological eradication rates, a finding which is consistent with sensitivity results at study inclusion. As explained in the Methods section, the statistical analysis that was applied emphasised the demonstration of similarity rather than proof of difference (Chow et al., 1992). Nevertheless, the superior eradication rates with sparfloxacin for both H. influenzae and M. catarrhalis seem real. The sparfloxacin eradication rate for S. pneumoniae was similar to that of co-amoxiclav. Persistence of P. aeruginosa was considered as indicative of overall failure by the Scientific Review Committee. This was, however, a rare occurrence (four with sparfloxacin and two with co-amoxiclav). The tolerability profile of sparfloxacin was satisfactory and similar to that of co-amoxiclav. However, sparfloxacin was better tolerated than co-amoxiclav in terms of gastrointestinal adverse events. The observed rates of adverse events appeared high because all adverse experiences (including failures or events related to the underlying disease) were recorded. In addition, the number of discontinuations due to study medication were low in both treatment groups. The adverse experiences leading to discontinuation were predominantly gastro-intestinal disorders and skin manifestations. In conclusion, sparfloxacin 100 mg once daily following a loading dose of 200 mg appears as effective and as well tolerated as co-amoxiclav in the treatment of accurately defined populations of patients with acute exacerbations of COPD, expected to benefit from antibiotic therapy. In addition, sparfloxacin requires only once-daily rather than thrice-daily administration. This may improve treatment compliance which is a major issue for success. References Anthonisen, N. R., Manfreda, J., Warren, C. P. W., Hershfield, E. S., Harding, G. K. M. & Nelson, N. A. (1987). Antibiotic therapy in exacerbations of chronic obstructive pulmonary disease. Annals of Internal Medicine 106, 196-204. Canton, R., Peman, J., Jimenez, M. T., Ramon, M. S. & Gobernado, M. (1992). In vitro activity of sparfloxacin compared with those of five other quinolones. Antimicrobial Agents and Chemotherapy 36, 558-65.

104 L. Allegra tt al. Chin, N. X., Gu, J. W., Yu, K. W., Zhang, Y. X. & Neu, H. C. (1991). In vitro activity of sparfloxacin. Antimicrobial Agents and Chemotherapy 35, 567-71. Chodosh, S. (1992). Bronchitis and asthma. In Infectious Diseases (Gorbach, S. L., Bartlett, J. G., Blacklow, N. R., Eds), pp. 476-85. W. B. Saunders, Philadelphia. Chow, A. W., Hall, C. B., Klein, J. O., Kammer, R. B., Meyer, R. D. & Remington, J. S. (1992). Evaluation of new anti-infective drugs for the treatment of respiratory tract infections. Clinical Infectious Diseases 15, Suppl. 1, S62-88. Cooper, M. A., Andrews, J. M., Ashby, J. P., Matthews, R. S. & Wise, R. (1990). In vitro activity of sparfloxacin, a new quinolone antimicrobial agent. Journal of Antimicrobial Chemotherapy 26, 667-76. Honeybourne, D., Greaves, I., Baldwin, D. R., Andrews, J. M., Harris, M. & Wise, R. (1994). The concentration of sparfloxacin in lung tissues after single and multiple oral doses. International Journal of Antimicrobial Agents 4, 151-5. Isada, C. M. & Stoller, J. K. (1994). Chronic bronchitis: role of antibiotics. In Respiratory Infections (Niederman, M. S., Sarosi, G. A. & Glassroth, J., Eds), pp. 621-34. W. B. Saunders, Philadelphia. Shimada, J., Nogita, T. & Ishibashi, Y. (1993). Clinical pharmacokinetics of sparfloxacin. Clinical Pharmacokinetics 25, 358-69.