Abstract
This is the first prospective clinical trial in which patients with acute bacterial exacerbation of chronic bronchitis have been stratified by degree of underlying illness.
Uncomplicated patients were randomised to levofloxacin 750 mg once daily (q.d.) for 3 days or azithromycin q.d. for 5 days. Complicated patients were randomised to levofloxacin 750 mg q.d. for 5 days or amoxicillin 875 mg/clavulanate 125 mg twice daily for 10 days.
Regardless of therapy, complicated patients demonstrated lower clinical and microbiological success than uncomplicated patients. Clinical success for clinically evaluable patients was similar for levofloxacin and azithromycin (93.0 versus 90.1%, respectively), and levofloxacin and amoxicillin/clavulanate (79.2 versus 81.7%, respectively). For microbiologically evaluable patients, clinical response to levofloxacin for 3 days was superior to azithromycin for 5 days (96.3 versus 87.4%, respectively), and levofloxacin for 5 days was similar to amoxicillin/clavulanate for 10 days (81.4 versus 80.9%, respectively). Microbiological eradication was superior for levofloxacin for 3 days compared with azithromycin for 5 days (93.8 versus 82.8%, respectively), and similar for levofloxacin and amoxicillin/clavulanate for 10 days (81.4 versus 79.8%, respectively).
In conclusion, levofloxacin 750 mg for 3 days was comparable to azithromycin for 5 days for uncomplicated patients with acute bacterial exacerbation of chronic bronchitis, while 5 days of 750 mg levofloxacin was comparable to 10 days of amoxicillin/clavulanate for complicated acute bacterial exacerbation of chronic bronchitis.
- Acute bacterial exacerbation of chronic bronchitis
- amoxicillin/clavulanate
- azithromycin
- chronic bronchitis
- levofloxacin
- risk assessment
Chronic bronchitis is a slowly progressive disease punctuated by acute exacerbations 1. These episodes contribute to morbidity, mortality and diminished quality of life (QoL) 2–5. Failure to treat appropriately with an antibacterial that eradicates the pathogen quickly is associated with persistent airway inflammation, and can increase the requirement for hospitalisation and repeated courses of therapy 6, 7.
Underlying factors have been shown to influence treatment outcome in acute bacterial exacerbation of chronic bronchitis (ABECB). These include baseline lung function, the number of exacerbations from the preceding year and certain comorbidities 8–11. Recent ABECB treatment guidelines 11 and published classification schemes are built upon the increasingly accepted premise 7, 8, 12–16 that patients should be grouped and antibacterial treatment appropriately selected according to these risk factors.
This study is the first large-scale prospective trial to stratify patients with ABECB into uncomplicated or complicated groups, based on the absence or presence of risk factors for treatment failure. Levofloxacin 750 mg was utilised for 3 (uncomplicated group) or 5 (complicated group) days. Comparator therapy was selected to match the anticipated microbiology and was in accordance with current treatment guidelines 11. This study applies classification schemes proposed in ABECB treatment guidelines and helps to define the appropriate use of anti-infectives in ABECB via patient stratification.
MATERIALS AND METHODS
Study population
Outpatient males and females (aged ≥18 yrs) with chronic bronchitis, as defined by the American Thoracic Society 17, and signs and symptoms consistent with type I or type II (increased sputum volume and sputum purulence required) ABECB, as defined by Anthonisen et al. 18, were eligible for enrolment. Patients with calculated creatinine clearance <50 mL·min−1 were excluded.
Study design
This was a randomised, blinded, parallel-group, non-inferiority study conducted at 73 sites in the USA. The protocol was approved by the participating institutions' Institutional Review Boards and informed consent was obtained prior to any study procedures.
Prior to randomisation, patients were assigned to uncomplicated and complicated treatment groups, based on predetermined criteria (table 1⇓). Patients were classified as current smokers, ex-smokers or nonsmokers. Uncomplicated patients were randomised to levofloxacin 750 mg once daily (q.d.) for 3 days or azithromycin 500 mg on day 1, followed by 250 mg q.d. for days 2–5. Complicated patients were randomised to levofloxacin 750 mg q.d. for 5 days or amoxicillin 875 mg/clavulanate 125 mg twice daily (b.i.d.) for 10 days. All treatment was orally administered. For uncomplicated patients, the treatment was double-blinded. For complicated patients, the large size of the amoxicillin/clavulanate tablets precluded over-encapsulation and blinding. A blister card was used to blind study personnel to the treatment, and patients were educated on the importance of maintaining the blind.
Collection of clinical and microbiological data
At study entry, clinical signs and symptoms of ABECB were evaluated. Sputum samples were collected and assessed for adequacy (≥25 polymorphonuclear neutrophils per low-powered field (LPF) and ≤10 squamous epithelial cells per LPF). A chest radiograph (to rule out pneumonia) and blood samples were obtained. Spirometry testing was performed if a historical baseline value was not available.
Clinical signs and symptoms were reassessed during on-therapy (days 3–6), post-therapy (days 10–19 uncomplicated or days 17–26 complicated patients) and post-study visits (days 35–40 uncomplicated/days 40–45 complicated patients).
Clinical and microbiological efficacy analyses
Post-therapy clinical response was classified as follows. Cure: all study entry criteria were met with resolution of acute signs and symptoms present at study entry (return to pre-exacerbation baseline) and no other antimicrobial therapy was administered. Improvement: all study entry criteria were met with clear, appreciable improvement in the acute signs and symptoms present at study entry, but with incomplete resolution of clinical evidence of infection (however, the infectious process was controlled and no further antimicrobials were required or were administered). Failure: persistence or worsening in the signs and symptoms of the ABECB process that indicated an inadequate response to treatment and resulted in additional nonstudy antimicrobial therapy for treatment of the original infection. Unable to evaluate: the clinical judgment of cure, improvement or failure could not be made due to the subject being lost to follow-up before the post-therapy visit.
Microbiological outcome was determined for patients who had a respiratory pathogen identified at admission. At post-therapy, microbiological response was classified as follows. Eradicated: absence of the study entry pathogen in a sputum culture obtained at the post-therapy visit in the absence of potentially effective systemic antimicrobials. Presumed eradicated: presumed absence of the study entry pathogen at the post-therapy visit for subjects deemed clinical successes (clinical cure or improvement) for whom no respiratory material was available for culture at the post-therapy visit. Persisted: continued presence of the original study entry pathogen in the post-therapy sputum culture. Presumed persisted: presumed presence of the study entry pathogen at the post-therapy visit for subjects who clinically failed and either had no post-therapy culture taken or had a negative post-therapy culture in the presence of potentially effective systemic antimicrobials. Unknown: no post-therapy culture obtained because subject was lost to follow-up, withdrew from therapy prematurely or had a sputum culture obtained in the presence of potentially effective systemic antimicrobials (except as noted previously for presumed persisted, which pertains to clinical failures).
Pharmacoeconomics and symptom resolution
A subset of patients volunteered to participate in a pharmacoeconomic sub-study. Patients were contacted monthly for up to 9 months in order to assess whether they used any acute (i.e. hospitalisation or emergency room), nonstudy, respiratory-related healthcare service and whether they used any additional antibacterials (respiratory healthcare utilisation questionnaire). These patients were additionally administered the transition dyspnoea index (TDI) questionnaire at each study visit and maintained a daily symptom diary (cough, breathlessness and sputum production). For the symptom diary, patients were asked to rate the severity of each symptom (cough, breathlessness or sputum production) on a four-point scale, ranging 0–3, as follows. 0: no symptoms; 1: “mild symptoms”; 2: “moderate symptoms”; and 3: “severe symptoms”. The scores for each item were added to create an overall symptom score ranging 0–9. Improvement was predefined as the 1st day on which a one-point reduction in total symptom scores occurred from baseline. Kaplan-Meier estimates were used to analyse the data, giving the percentage of the total patients who experienced a one-point reduction in total symptom scores from baseline by day 3. Additionally, resolution of sputum volume, purulence and cough were rated by the investigator at each patient visit (including the on-therapy visit).
Statistical analyses
Efficacy analyses were performed separately for uncomplicated and complicated patients (smokers, ex-smokers and nonsmokers combined). The primary efficacy variable was clinical success (cured plus improved) in clinically evaluable (CE) patients at the post-therapy (test-of-cure) assessment.
To be CE, a patient had to meet the following requirements: confirmed diagnosis of ABECB; availability of post-therapy data; adequate therapy (≥80% but <120% of protocol-specified doses taken); and no concurrent use of effective antibacterial agents within 72 h of study entry through to the post-therapy visit. To be microbiologically evaluable (ME), a patient had to be CE and have a pathogen identified at study entry. Microbiological response at the post-therapy visit was a secondary efficacy variable.
Comparability of treatment groups for baseline characteristics was evaluated by Fisher's exact test for dichotomous variables, the Wilcoxon Mann-Whitney test for ordinal variables and the unpaired t-test for continuous variables. For efficacy analyses, two-sided 95% confidence intervals (CI) around the difference between treatment groups (comparator minus levofloxacin) were computed. To conclude that the levofloxacin regimen was at least as efficacious as the comparator, a 95% CI upper bound of ≤15% was used for clinical success <90%, and if clinical success rates in one of the treatment groups was ≥90%, the upper bound of the 95% CI had to be ≤10%. The 95% CI were computed with a continuity correction outlined by Hauck and Anderson 19. Superiority statistical assessment was, therefore, not prespecified.
The analyses of the difference in TDI score (levofloxacin minus comparator) were evaluated by the Wilcoxon rank-sum test. The respiratory healthcare utilisation questionnaire and symptom resolution analyses were evaluated using Fisher's exact test. Symptom resolution was defined as a one-point reduction in the total respiratory symptoms score.
RESULTS
Study population
A total of 763 ABECB patients were enrolled in the intent-to-treat (ITT) population (all randomised patients who took at least one dose of study drug), 394 uncomplicated and 369 complicated. In the uncomplicated population, 192 patients received levofloxacin and 202 patients received azithromycin. In the complicated patient population, 187 patients received levofloxacin and 182 received amoxicillin/clavulanate.
The most common reasons for clinical and microbiological nonevaluability were inappropriate post-therapy evaluation dates (i.e. the evaluation occurred outside the time window specified) and unconfirmed diagnosis (table 2⇓).
Demographic and disease characteristics at baseline were different when comparing the uncomplicated and complicated CE populations (table 3⇓). Within the uncomplicated population, patients receiving levofloxacin for 3 days had characteristics that were comparable to patients receiving azithromycin for 5 days. For the complicated population, patients receiving levofloxacin for 5 days had characteristics comparable to patients receiving amoxicillin/clavulanate for 10 days. For all treatment groups, the CE population was comparable to the ITT population. Both the uncomplicated and complicated groups contained some nonsmokers. Information regarding second-hand smoke or other environmental exposure among the nonsmokers was not specifically collected.
Clinical efficacy
For CE uncomplicated patients, clinical success at post-therapy was 93.0% for levofloxacin and 90.1% for azithromycin (fig. 1⇓). When nonsmokers were excluded, the result of the analysis was 93.0% for levofloxacin and 89.4% for azithromycin. However, for ME patients, the clinical response for 3 days of levofloxacin was superior (95% CI upper limit <0) to 5 days of azithromycin (96.3 versus 87.4%). Clinical response rates by pathogen of interest (pathogens isolated from more than five patients) are presented in table 4⇓. Post-study clinical success rates for the CE population of 92.2 and 92.5% were seen with levofloxacin and azithromycin, respectively.
Clinical success at post-therapy for the CE complicated patients was 79.2% for levofloxacin and 81.7% for amoxicillin/clavulanate (fig. 1⇑). Levofloxacin 750 mg q.d. for 5 days was as effective as 10 days of amoxicillin/clavulanate b.i.d. When nonsmokers were excluded, the result of the analysis was 79.8% for levofloxacin and 81.5% for amoxicillin/clavulanate. For ME patients, the clinical response for 5 days of levofloxacin was similar to 10 days of amoxicillin/clavulanate (81.4 versus 80.9%, respectively). Clinical response rates between treatments by pathogen of primary interest are presented in table 4⇑. Post-study clinical success rates of 80.4 and 82.7% were seen in levofloxacin and amoxicillin/clavulanate patients, respectively.
In vitro susceptibility of bacterial isolates at study entry
In the uncomplicated group, 318 study entry pathogens were isolated; 151 were isolated from 110 levofloxacin patients and 167 were isolated from 120 azithromycin patients. Of the pathogens with known susceptibilities, 140 out of 141 (99.3%) isolated from levofloxacin patients and 106 out of 114 (93.0%) isolated from azithromycin patients were susceptible to treatment.
Of the 23 Streptococcus pneumoniae isolates from uncomplicated patients, eight (34.8%) were resistant to azithromycin, whereas none was resistant to levofloxacin.
In the complicated group, 341 study entry pathogens were isolated; 172 were isolated from 125 levofloxacin patients and 169 were isolated from 123 amoxicillin/clavulanate patients. Of the pathogens with known susceptibilities, 149 out of 156 (95.5%) isolated from levofloxacin patients and 114 out of 127 (89.8%) isolated from amoxicillin/clavulanate patients were susceptible to treatment.
Microbiological efficacy
Post-therapy microbiological eradication rates are summarised in figure 1⇑ and table 5⇓. For uncomplicated patients, microbiological eradication was 93.8% for 3 days of levofloxacin and 82.8% for 5 days of azithromycin. The 95% CI was <0, indicating that levofloxacin 750 mg q.d. for 3 days was statistically superior to azithromycin administered for 5 days.
For complicated patients, the microbiological eradication rates were 81.4% for 5 days levofloxacin and 79.8% for 10 days amoxicillin/clavulanate. For Haemophilus influenzae, amoxicillin/clavulanate eradicated 100%, while levofloxacin eradicated 83.3%. All of the H. influenzae isolates were susceptible to levofloxacin with minimal inhibitory concentrations of 0.03 μg·mL−1, and none of the patients had documented persistence of the pathogen at post-therapy.
Concomitant steroid use
To examine the relationship between outcome and concomitant steroid use, a sub-analysis was performed by grouping patients into categories of either acute or chronic concomitant steroid use. For all populations, there was no evidence that acute steroid use impacted upon the clinical success or microbiological eradication rates (table 6⇓).
Pharmacoeconomics and symptom resolution
Results of key assessments are summarised in table 7⇓. In the CE complicated population, levofloxacin demonstrated a greater improvement (p = 0.036) in TDI score by days 3–6 compared with amoxicillin/clavulanate. A similar improvement in TDI score was found between treatment groups among the uncomplicated patients.
Patient-recorded diary entries indicated that respiratory symptoms were significantly improved by day 3 (p = 0.015) in the uncomplicated ME levofloxacin group compared with azithromycin. For CE complicated patients, by the on-therapy physician visit, more levofloxacin than amoxicillin/clavulanate patients experienced resolution of their cough (56.3 and 43.2%, respectively; p = 0.019) as well as reduction of sputum volume (63.0 and 48.3%, respectively; p = 0.01) and sputum purulence (54.3 and 37.7%, respectively; p = 0.003), as assessed by the study investigator. Similar improvement at the on-therapy visit was found between treatment groups among the uncomplicated patients.
Safety
Of the 394 ITT uncomplicated patients, 389 (190 levofloxacin/199 azithromycin) were safety evaluable (all randomised patients who took at least one dose of study drug and had available safety information). Similarly, of 369 ITT complicated patients, 362 (183 levofloxacin/179 amoxicillin) were evaluable for safety. Within-group reporting rates for at least one treatment-emergent adverse event were 34.7% for levofloxacin when compared with 36.7% for azithromycin (95% CI −7.8–11.7) and, in the complicated, 42.1% for levofloxacin when compared with 48.6% for amoxicillin/clavulanate (95% CI −4.0–17.0).
For the uncomplicated group, 12 (6.3%) levofloxacin patients and 16 (8.0%) azithromycin patients reported drug-related adverse events. The most common drug-related adverse events in uncomplicated patients were nausea (four levofloxacin and three azithromycin patients) and diarrhoea (three levofloxacin and 10 azithromycin patients). Adverse events resulting in discontinuation of treatment (four in levofloxacin and one in azithromycin) were considered probably or very likely to be related to the study drug. Four (2.1%) levofloxacin patients experienced treatment-emergent serious adverse events (SAEs) compared to one (0.5%) patient in the azithromycin group. None of the SAEs in either treatment group was assessed by the investigator as being related to the study drug.
For complicated patients, 16 (8.7%) levofloxacin patients and 16 (8.9%) amoxicillin/clavulanate patients reported drug-related adverse events. The most common drug-related adverse event was diarrhoea (four levofloxacin and five amoxicillin/clavulanate patients). Adverse events resulting in discontinuation of treatment (five in levofloxacin and one in amoxicillin/clavulanate) were considered probably or very likely to be related to the study drug. Eleven (6.0%) levofloxacin patients experienced treatment-emergent SAEs compared with 13 (7.3%) amoxicillin/clavulanate patients. The only SAE assessed by the investigator to have a possible relationship to the study drug in either treatment group was aggravated depression in an amoxicillin/clavulanate-treated subject.
The rates of all drug-related adverse events, SAEs, events resulting in discontinuation of treatment and markedly abnormal laboratory findings were not notably different between levofloxacin and comparator. No deaths occurred, and no unusual or unexpected safety problems were reported.
DISCUSSION
The major goals of therapy for ABECB should be to provide rapid clinical resolution, eradicate the causative pathogen and return respiratory function to pre-exacerbation baseline as quickly as possible 16. The selection of antibacterial therapy is generally made empirically, based on probable pathogens and local susceptibility patterns. Recent guidelines now additionally recommend choosing antibacterial therapy based on patient risk factors 7, 8, 11–16, 20. The recommended antibacterial therapy reflects the spectrum of potential pathogens and the contribution of risk factors to outcomes.
This is the first large-scale, prospective trial in ABECB to use risk stratification of patients to guide the selection of antibacterial regimens, in accordance with recently published treatment guidelines. Only patients with well-defined exacerbations were enrolled, and care was taken to exclude patients who were likely to have viral infections. Prior to randomisation, patients were stratified into uncomplicated and complicated groups. Different durations of levofloxacin, either 3 days for uncomplicated patients or 5 days for complicated patients, were studied. Different comparators were employed for uncomplicated (azithromycin) and complicated (amoxicillin/clavulanate) patients, as the guidelines recommend 11, 16, 20.
This study demonstrated, in a prospective fashion, that complicated patients, regardless of the therapeutic agent employed, had poorer clinical and microbiological outcomes than uncomplicated patients. For CE patients, clinical success rates were similar for levofloxacin and azithromycin in the uncomplicated group, and for levofloxacin and amoxicillin/clavulanate in the complicated group. Coupled with the 750-mg dose, a shortened course of levofloxacin therapy (3–5 days) did not limit clinical efficacy. In fact, for microbiologically confirmed cases, clinical success and microbiological eradication were superior with 3 days of levofloxacin 750 mg when compared to 5 days of azithromycin, while only 5 days of levofloxacin 750 mg were similar to 10 days of amoxicillin/clavulanate. Additionally, patients treated with levofloxacin 750 mg were more likely to experience an improvement in TDI score and earlier resolution of respiratory symptoms, such as increased sputum production, sputum purulence and cough, than their amoxicillin/clavulanate-treated counterparts in the complicated group. Similar findings were noted in a recent study of levofloxacin 750 mg for 5 days for community-acquired pneumonia 21.
These data confirm that the selection of antibiotics can influence the clinical and bacteriological outcome in ABECB. The difference in success rates seen among ME patients in the uncomplicated group who received different treatments supports the position that antibacterial therapy can improve outcomes in selected patients with ABECB 4, and confirms that the choice of agent or class of agents is, in fact, clinically meaningful. In this clinical trial, among the uncomplicated patients, no S. pneumoniae isolates were resistant to levofloxacin. However, 34.8% were resistant to azithromycin. This concurs with national surveillance studies, such as Tracking Resistance in the United States Today (TRUST), which continues to demonstrate 99% susceptibility to levofloxacin, but with resistance to macrolides of ∼28% 22.
This study has confirmed the importance of evaluating additional outcome measures in ABECB trials 23, 24. As ABECB is known to diminish QoL 2–5 and increase the likelihood of hospitalisation and repeated antibacterial courses 7, 25, 26, results from a daily diary and a respiratory healthcare utilisation questionnaire were utilised to demonstrate the rapid resolution of symptoms in both uncomplicated and complicated levofloxacin patients. Others have also suggested improved symptom resolution, health status and healthcare utilisation with quinolones in selected ABECB patients 27–29.
The optimal duration of therapy in ABECB remains to be established 11. Recently, investigators have cogently argued for using the shortest duration of antibacterials possible to achieve multiple goals, including improved compliance, lower cost and a reduced frequency of potential adverse events.
As a concentration-dependent agent 28–33, therapeutic outcome with levofloxacin is most closely linked to area under the concentration–time curve/minimum inhibitory concentration (MIC) or maximum concentration/MIC ratios. This pharmacodynamic principle, paired with levofloxacin's favourable safety index (which allows flexible dosing), supports the approach of using higher doses, such as the 750-mg dose, to achieve rapid microbiological eradication, rapid resolution of the clinical signs and symptoms of infection, and to allow for a shorter duration of therapy. Levofloxacin's excellent clinical efficacy in pulmonary infections and its rapidly bactericidal action suggest that a higher daily dose, coupled with a shorter treatment duration, could optimise ABECB therapy.
There were several challenges and limitations of this study. First, therapy in the complicated group could not be blinded due to the large size of the amoxicillin/clavulanate tablets. However, investigators were blinded to treatment, and patients were instructed not to reveal the study medication to blinded study personnel. This study could not definitively exclude patients that may have experienced an acute exacerbation for a reason other than bacterial infection. Additionally, ∼13% (39 patients) of the uncomplicated ABECB group were aged <35 yrs, and 10 of these patients were also nonsmokers, which may indicate a respiratory condition other than chronic bronchitis. Yet, the small percentage of these patients, as well as equal distribution between treatment groups (six in the levofloxacin group and four in the azithromycin group), should limit their effects on the overall conclusions from the current study. A minority of randomised patients did not report a smoking history. The results of analyses, however, were similar when these putative nonsmokers were excluded. Steroid therapy was not standardised 24, although administration was not associated with differential response rates by antibacterial agents. Interestingly, as others have noted 28, response rates were lower when acute steroids were administered. Spirometry was obtained at the time of exacerbation in some of the patients. This should be of limited concern as forced expiratory volume in one second changes little during outpatient exacerbations 34; this approach would also be of greater applicability in clinical practice where baseline spirometry may not be available. Sputum cultures were used to detect the presence of persistent pathogens at post-therapy, although this method may underestimate the presence of colonisation, particularly by H. influenzae 35.
In conclusion, patient stratification may be a worthwhile approach in the management of patients with acute bacterial exacerbation of chronic bronchitis. Additional prospective clinical trials would be valuable to validate antimicrobial selection based on patient stratification by severity of illness. Underlying risk factors that influence the likelihood of treatment failure may guide the choice of an appropriate antibacterial agent. This study demonstrates that short courses of levofloxacin 750 mg are at least as effective as traditional courses of azithromycin and amoxicillin/clavulanate for the spectrum of patients with acute bacterial exacerbation of chronic bronchitis.
- Received September 10, 2004.
- Accepted January 27, 2005.
- © ERS Journals Ltd