Randomized Controlled Trial of Sequential Intravenous (i.v.) and Oral Moxifloxacin Compared with Sequential i.v. and Oral Co-Amoxiclav with or without Clarithromycin in Patients with Community-Acquired Pneumonia Requiring Initial Parenteral Treatment

The objective of the present trial was to compare the efficacy,safety, and tolerability of moxifloxacin (400 mg) given intravenously(i.v.) once daily followed by oral moxifloxacin (400 mg) for7 to 14 days with the efficacy, safety, and tolerability ofco-amoxiclav (1.2 g) administered by i.v. infusion three timesa day followed by oral co-amoxiclav (625 mg) three times a day,with or without clarithromycin (500 mg) twice daily (i.v. ororally), for 7 to 14 days in adult patients with community-acquiredpneumonia requiring initial parenteral therapy. A total of 628patients were enrolled and assessed by evaluation of their clinicaland bacteriological responses 5 to 7 days and 21 to 28 daysafter administration of the last dose of study medication. Althoughthe trial was designed, on the basis of predefined outcomes,to demonstrate the equivalence of the two regimens, the resultsshowed statistically significant higher clinical success rates(for moxifloxacin, 93.4%, and for comparator regimen, 85.4%;difference [ ${Delta}$ ], 8.05%; 95% confidence interval [CI], 2.91 to13.19%; P = 0.004) and bacteriological success rates (for moxifloxacin,93.7%, and for comparator regimen, 81.7%; ${Delta}$ , 12.06%; 95% CI,1.21 to 22.91%) for patients treated with moxifloxacin. Thissuperiority was seen irrespective of the severity of the pneumoniaand whether or not the combination therapy included a macrolide.The time to resolution of fever was also statistically significantlyfaster for patients who received moxifloxacin (median time,2 versus 3 days), and the duration of hospital admission wasapproximately 1 day less for patients who received moxifloxacin.The treatment was converted to oral therapy immediately afterthe initial mandatory 3-day period of i.v. administration fora larger proportion of patients in the moxifloxacin group thanpatients in the comparator group (151 [50.2%] versus 57 [17.8%]patients). There were fewer deaths (9 [3.0%] versus 17 [5.3%])and fewer serious adverse events (38 [12.6%] versus 53 [16.5%])in the moxifloxacin group than in the comparator group. Therates of drug-related adverse events were comparable in bothgroups (38.9% in each treatment group). The overall incidenceof laboratory abnormalities was similar in both groups. Thus,it is concluded that monotherapy with moxifloxacin is superiorto that with a standard combination regimen of a ß-lactamand a ß-lactamase inhibitor, co-amoxiclav, with orwithout a macrolide, clarithromycin, in the treatment of patientswith community-acquired pneumonia admitted to a hospital.

INTRODUCTION

Top

Introduction

Community-acquired pneumonia (CAP) is a common condition andcontinues to be an important cause of mortality. Each year inthe United Kingdom approximately 50,000 people with CAP (3)are admitted to a hospital. Some 80% of cases are managed inthe community, where the mortality rate is 1 to 2%. Among thosepatients admitted to a hospital, 5 to 10% will die, with themortality rate rising to 50% among those requiring intensivecare. In the United States, CAP is the sixth leading cause ofdeath (13). The cost of treatment for CAP in the United Kingdomalone is estimated to be £440.7 million per year, withhospitalized patients accounting for as much as 96% of thiscost (11). Consequently, there is growing pressure upon cliniciansto reduce the number and duration of hospital admissions.

The importance of CAP as a major cause of morbidity and mortalitytogether with its impact on health care resources is reflectedin the plethora of published studies. Furthermore, there remaininconsistencies between national clinical guidelines; thesereflect variations in the prevalence of target pathogens, drugresistance patterns, and clinical practice. Clinical trialsdesigned to support the licensing of new agents generally limitany progress in our knowledge of disease management since theyare designed to determine equivalence between the test and controlregimens.

Moxifloxacin is a new fluoroquinolone antibiotic with increasedactivity against gram-positive organisms, notably Streptococcuspneumoniae, and acts by inhibiting bacterial topoisomerasesII and IV. The objective of the trial reported here was to examinethe efficacy, safety, and tolerability of moxifloxacin (400mg) given intravenously (i.v.) once daily followed by oral moxifloxacin(400 mg) for 7 to 14 days for the treatment of CAP. This regimenwas compared with co-amoxiclav (1.2 g) administered by i.v.infusion three times a day followed by oral co-amoxiclav (625mg) three times a day, with or without clarithromycin 500 mgtwice daily (i.v. or orally), for 7 to 14 days in adult patientsrequiring initial parenteral therapy.

MATERIALS AND METHODS

Top

Materials and Methods

Study population. Male and female patients aged 18 years or over with radiologicalevidence of CAP (and who had been in a hospital for less than48 h) were entered into the trial. To be included the patientswere required to have a temperature $>=$ 38.5°C or leukocytosisand at least one of the following clinical symptoms of pneumonia:cough, purulent sputum, dyspnea, rigors, pleuritic chest pain,or auscultatory findings. All patients required initial parenteraltherapy, and approximately half had severe pneumonia, as definedby the criteria of the American Thoracic Society (16). To meetthe definition of severe CAP the patients had to have at leastone of the following: a respiratory rate of $>=$ 30 breaths/min,hypoxemia with a partial oxygen pressure of $<=$ 8 kPa (60 mm Hg),a need for mechanical ventilation, diastolic blood pressure $<=$ 60 mm Hg, a chest X ray showing bilateral or multilobar involvement,or a requirement for treatment with vasopressors for more than4 h.

Exclusion criteria included the presence of a coexisting diseaseconsidered likely to affect the outcome of the study (e.g.,lung cancer, empyema, or severe cardiac failure) or a rapidlyfatal underlying disease; known prolongation of the QT intervalor the use of class IA or class III antiarrhythmics; known hypersensitivityto fluoroquinolones, ß-lactams, or macrolides; aspirationpneumonia; and pretreatment with systemic antibacterial agentsfor more than 24 h prior to enrollment in the study. Patientswho had clearly failed previous antibacterial therapy, whichthey had received for at least 72 h for the current pneumoniaepisode, could be enrolled unless the antibacterial regimencontained a fluoroquinolone or a ß-lactam-ß-lactamaseinhibitor combination.

Patients were recruited by 65 centers in 10 countries: Belgium,France, Germany, Greece, Israel, South Africa, Spain, Switzerland,Russia, and the United Kingdom. The study was conducted in accordancewith the Internal Conference on Harmonization Harmonized TripartiteGuideline for Good Clinical Practice (1997).

Study design. This multinational, multicenter study used a randomized, open,parallel group design to compare sequential i.v. and oral moxifloxacintreatment with that of a standard antibiotic regimen, sequentiali.v. and oral co-amoxiclav with or without clarithromycin, inpatients with CAP who, in the judgment of the investigator,required initial parenteral therapy.

Randomization was performed locally on the basis of a centrallyestablished code. Prior to randomization to one of the two paralleltreatment groups, the patients were stratified into two groupsaccording to disease severity. Patients were to be treated fora period of no less than 7 days and for a maximum of 14 days.After a minimum of 3 days patients could be converted from i.v.to oral therapy if they had demonstrated clinical improvement,were able to tolerate the oral intake of medication, and hadno gastrointestinal condition likely to impair absorption.

The patients in the moxifloxacin group were initially administered400 mg of moxifloxacin in 250 ml of 0.8% saline as an i.v. infusionover 60 min once daily. Thereafter, moxifloxacin was given asa 400-mg tablet once a day.

Patients receiving the comparator regimen were given 1.2 g ofco-amoxiclav in 20 ml of water as an i.v. injection over 3 to4 min three times a day. After conversion to oral therapy, co-amoxiclavwas administered three times a day as 625-mg tablets. If theinvestigator considered that the addition of clarithromycinwas indicated, in order to cover the possibility that the pathogensmight be atypical organisms, it could be given initially aseither an i.v. or an oral dose. Clarithromycin was suppliedas 500 mg of sterile powder for reconstitution in 10 ml of sterilewater and was infused in 250 ml of a suitable diluent over 60min twice daily. Oral clarithromycin was administered as a 500-mgtablet twice a day.

Clinical, bacteriological, and laboratory examinations wereperformed pretreatment, 5 to 7 days after the end of therapy(the test-of-cure [TOC] visit), and 21 to 28 days after theend of treatment. These included a Gram stain and culture ofsputum, blood culture, blood for hematology and chemistry safetyprofiles, and serologic testing for Legionella pneumophila,Chlamydia pneumoniae, and Mycoplasma pneumoniae (testing forthe last organism was done only pretherapy and 21 to 28 daysposttherapy). According to the study protocol, invasive techniqueswere not required to collect specimens for culture. Specificculture or antigen testing was not done to detect atypical pathogens.If a patient failed to respond to the study medication, theassessments were to be completed before and 5 to 7 days afterthe end of an alternative antibiotic therapy.

Vital signs, i.e., temperature, heart rate, respiratory rate,and blood pressure, were monitored daily up to the time of thepatient's discharge from the hospital or the end of treatment,whichever came first. Adverse events (including laboratory abnormalities)were assessed up to the TOC visit. Any adverse events presentat that time were monitored until they resolved or stabilized.On day 3 of i.v. therapy, a standard 12-lead electrocardiogramwas to be performed for all patients within 30 min after theend of the first infusion of moxifloxacin or co-amoxiclav withor without clarithromycin.

The primary efficacy variable was the clinical response 5 to7 days after the end of the drug treatment period (determinedat the TOC visit) in the valid-per-protocol (PP) population.The clinical responses were defined as "clinical cure," whichwas the disappearance of acute signs and symptoms related toinfection with no requirement for further antibiotic therapy;"clinical failure," which was the failure to respond or andan insufficient response to the study drug that required a modificationin antibacterial therapy or that resulted in death from theprimary diagnosis; and "indeterminate," which was a clinicalresponse that could not be determined because of early discontinuationfrom the study or protocol violation. Secondary efficacy parametersincluded the time to resolution of fever, the bacteriologicalresponse 5 to 7 days after treatment, bacteriological and clinicalresponses 21 to 28 days posttreatment, the duration of i.v.therapy, and the duration of hospital admission.

Statistical analysis. In accordance with the notes for guidance on the use of newantibacterial medicinal products of the European Agency forthe Evaluation of Medicinal Products (April 1997), the specifiedequivalence delta between the clinical success rates for thetwo treatment groups was 10%. With an estimated failure rateof 15%, ${alpha}$ equal to 5% (two sided), ß equal to 15%,and a 5% adjustment for the multicenter design, the FarringtonManning sample size calculation called for 312 patients to beenrolled in each treatment group.

Two types of analysis were performed for each efficacy variable:a PP analysis and an intention-to-treat (ITT) analysis.

In order to be included in the primary efficacy analysis, patientshad to have a diagnosis of CAP and a documented rate of compliancewith receipt of the study medication of $>=$ 80%. Study drugs wereto have been given for a minimum of 48 h in the case of clinicalfailure or a full 5 days in the case of clinical cure. The patientcould not have received any concomitant systemic antimicrobialagents except in cases of treatment failure. There were to beno protocol violations that might have influenced efficacy andno missing essential data such as the clinical assessment atthe TOC visit.

The ITT analysis included those patients who had been randomizedand who received at least one dose of study drug.

Demographic and baseline characteristics were summarized bytreatment group by using the mean, standard deviation, median,quartiles, and minima and maxima for quantitative data or frequencycounts for categorical data. Treatment group comparability wasestablished for both the PP and the ITT analyses. The valid-for-safetypopulation was the same as that for the ITT analysis.

A two-sided 95% confidence interval (CI) for the differencebetween success rates (the rate for the moxifloxacin group minusthe rate for the comparator group) was calculated by using Mantel-Haenszelweights. Moxifloxacin was deemed to be not less effective thanthe comparator regimen if the lower limit of this CI was greaterthan -10%. In this analysis any clinical failures occurringbefore the TOC visit were counted as clinical failures at theTOC visit. Similarly, failures at the TOC visit were carriedforward for the late follow-up analysis. Clinical success wasdefined as a clinical cure, and bacteriological success wasdefined as eradication or presumed eradication of the pathogen(s).The findings of the study not only supported the noninferiorityhypothesis but also suggested superiority. Therefore, in accordancewith the Committee for Proprietary Medicinal Products pointsto consider, the exact probability associated with a test ofsuperiority was calculated by using the continuity-adjustedchi-square test.

The time to resolution of fever was analyzed by using both thelog rank test and the Wilcoxon test (since there was evidenceof nonproportional hazards between the two treatment groups).

The safety of each treatment was described by summary statistics;no formal statistical analyses were performed. Treatment groupswere compared with respect to the incidence rate of adverseevents, laboratory data, and electrocardiographic findings.

RESULTS

Top

Results

Efficacy. A total of 628 patients were enrolled in the trial, and 622were valid for the ITT and safety analyses. More than half (n= 321) of the patients in the ITT and safety analyses had adiagnosis of severe CAP. There were 538 patients in the PP analysis.The trial population profile is depicted in Fig. 1.

View larger version (43K):
[in this window]
[in a new window]
FIG. 1. Trial population profile. Percentages are expressed as a fraction of the number of patients per treatment group. incl/exclu, inclusion/exclusion.

All study patients had signs and symptoms of CAP which was confirmedby chest radiography. The baseline characteristics of the patientsincluded in the ITT analysis are summarized in Table 1, andtesting for homogeneity showed that there was no imbalance betweenthe two treatment groups. Similarly, the numbers of patientsreceiving antibiotics prior to entry into the trial and thenumbers of patients with $>=$ 72 h of prior antibiotic therapy werebalanced between the two treatment groups.

View this table:
[in this window]
[in a new window]
TABLE 1. Baseline characteristics of patients in ITT analysis

Bacterial cultures or serology were performed for a total of326 (52%) patients. Not all patients were able to provide materialfor culture. There were 159 patients in the moxifloxacin groupand 167 patients in the comparator group. Causative organismswere identified in 194 patients, giving an incidence of 59.5%(57.2% in patients in the moxifloxacin group and 61.7% in patientsin the comparator group).

S. pneumoniae and Haemophilus influenzae were the most commonorganisms identified, accounting for 55.4 and 19.6% of all pathogensidentified, respectively. Multiple pathogens were identifiedin 6.4% of patients. Atypical organisms (M. pneumoniae, C. pneumoniae,and L. pneumophila) were the causes of infections in 13.8% (45of 326) of patients (infection was defined as a fourfold increasein the antibody titer between acute- and convalescent-phasesera). M. pneumoniae was the most frequent atypical organismidentified. For five patients with positive serologies for atypicalorganisms, a causative bacterial organism(s) was also isolatedfrom cultures of respiratory and/or blood specimens.

The microbiological susceptibilities of the most frequentlyisolated pathogens from pretreatment samples are presented inTable 2. The susceptibilities are presented as the MICs andthe MICs at which 90% of the isolates tested are inhibited,as determined by the E-test methodology.

View this table:
[in this window]
[in a new window]
TABLE 2. Pretherapy bacterial susceptibility, by source, for the most frequently isolated organisms (PP analysis)

Table 3 presents the results of the analysis of efficacy atthe TOC visit and late follow-up visits for the population includedin the PP analysis.

View this table:
[in this window]
[in a new window]
TABLE 3. Analysis of efficacy for population in PP analysis

Moxifloxacin-treated patients had a significantly higher clinicalcure rate, irrespective of the assessment time. The clinicalcure rate at the TOC visit for the moxifloxacin group was 8.05%greater than that for the patients receiving the comparatorregimen (95% CI, 2.91 to 13.19%). The clinical cure rate atthe TOC visit for patients treated with moxifloxacin was alsosignificantly superior to that for patients receiving the comparatorregimen in the ITT analysis (difference [ ${Delta}$ ] = 8.25%; 95% CI,3.13 to 13.6%) and in the subgroup of patients valid for microbiologicalanalysis. The higher rate of clinical cure seen among the patientstreated with moxifloxacin in the PP analysis was irrespectiveof whether the patients on the comparator regimen were treatedwith co-amoxiclav alone (85%) or co-amoxiclav together withclarithromycin (85.6%). The continuity-adjusted P values (asdetermined by the chi-square test) for clinical response atthe TOC assessment were 0.004 for the population in the PP analysisand 0.003 for the population in the ITT analysis (missing andindeterminate responses were dropped from the analyses).

The wider CIs for the point estimates of bacteriological successreflect the smaller patient numbers in the analyses. Nevertheless,the results are consistent with those for clinical cure, althoughstatistical significance can be claimed only for the PP analysisat the TOC visit. The difference in bacterial success ratesbetween moxifloxacin and the comparator groups at the TOC visit(5 to 7 days posttreatment) was 12% in favor of the moxifloxacingroup. For the group of patients valid for microbiological analysis,the superiority was largely related to documented S. pneumoniaeinfections. The rate of bacteriological success for patientsinfected with penicillin-intermediate strains (MICs, $>=$ 1.0 to<2.0 mg/liter) and penicillin-resistant strains (MICs, $>=$ 2.0mg/liter) of S. pneumoniae at days 5 to 7 posttreatment wassignificantly higher for the moxifloxacin group and was supportedby the bacteriological evaluation at the late follow-up on days21 to 28. Clinical cures were achieved in all moxifloxacin-treatedpatients with infections caused by penicillin-resistant (oneof one) or penicillin-intermediate (four of four) strains, whereasclinical cures were achieved in zero of one and one of threecomparator regimen-treated patients with infections caused bypenicillin-resistant and penicillin-intermediate strains, respectively.There were fewer instances of persisting pathogens (all of whichwere presumed cases of persisting infection) in respiratorysamples from patients treated with moxifloxacin than in samplesfrom patients in the comparator group, which included some patientswith documented cases of persisting infections. The few casesof superinfection seen in this study all occurred in patientstreated with the comparator regimen.

Table 4 presents the bacteriological responses by organism andsource.

View this table:
[in this window]
[in a new window]
TABLE 4. Bacteriological responses, by organism

For the subgroup of patients valid for microbiological analysis,the clinical cure rates on days 5 to 7 were 93.8 and 81.7% forthe moxifloxacin (n = 60) and comparator (n = 58) groups, respectively.In order to be valid for microbiological analysis, patientshad to have been eligible for the PP analysis, a causative organism(s)had to have been identified from a pretreatment sputum or bloodspecimen or paired serology samples, and the patient had tohave had an appropriate bacteriological evaluation at the TOCvisit on days 5 to 7. Overall, 17 of 17 moxifloxacin-treatedpatients and 22 of 24 comparator regimen-treated patients withatypical infections had a clinical cure at the TOC visit. Themajority of comparator regimen-treated patients with atypicalinfections received clarithromycin: 10 of 17 patients infectedwith M. pneumoniae, 4 of 5 patients infected with Chlamydiaspp., and 2 of 4 patients infected with L. pneumophila.

The rate of clinical cure was higher in the moxifloxacin groupin both patients with nonsevere pneumonia and patients withsevere pneumonia. Among the moxifloxacin-treated patients, theclinical cure rates were 94.6% for those with nonsevere pneumoniaand 92.2% for those with severe pneumonia. Among the patientstreated with the comparator regimen, the clinical cure ratefor patients with nonsevere pneumonia was 86% and that for patientswith severe pneumonia was 84.7%.

The differences in clinical cure and bacteriological successrates between the two treatment groups, together with the 95%CI for the populations that were included in both the PP andthe ITT analyses, are depicted in Fig. 2.

View larger version (17K):
[in this window]
[in a new window]
FIG. 2. Clinical and bacteriological outcomes. ${blacksquare}$ , point estimate of the difference between groups; horizontal bars, 95% CI.

Before the start of treatment the mean ± standard deviation(SD) temperatures for the populations included in the ITT analysiswere 38.7 ± 0.9°C for the moxifloxacin group and38.8 ± 1.0°C for the comparator group. The time toresolution of temperature (defined as the first day at whichthe patient's peak temperature was less than or equal to 37.5°C)was analyzed by both the log rank test and the Wilcoxon test.The results are illustrated in Fig. 3.

View larger version (14K):
[in this window]
[in a new window]
FIG. 3. Cumulative percentage of patients in whom fever had resolved. , moxifloxacin group; ${blacksquare}$ , comparator group.

The median time to temperature resolution was 2 days for patientstreated with moxifloxacin and 3 days for those treated withthe comparator regimen. Among the population included in theITT analysis, 140 (58.6%) of moxifloxacin-treated patients wereapyrexic by day 2, whereas 119 (46.7%) of the patients in thecomparator group were apyrexic by day 2 (P = 0.025 by the logrank test and P = 0.009 by the Wilcoxon test). The mean timeto resolution of temperature was shorter in the moxifloxacingroup than the comparator group, irrespective of disease severity.When the analysis was repeated by using a definition of apyrexiaas a temperature of $<=$ 38°C, the result was still statisticallysignificant (P = 0.018 by the log rank test and P = 0.002 bythe Wilcoxon test). The faster return to apyrexia in the moxifloxacin-treatedpatients could not be attributed to a greater use of drugs thatmay have contributed to a lowering of their temperatures (Table5). Similarly, when only patients whose fevers resolved by day2 and who used an antipyretic at some point during the studyare considered, fewer moxifloxacin-treated patients (57.8%)than comparator regimen-treated patients (66.1%) were usingantipyretic drugs at day 2.

View this table:
[in this window]
[in a new window]
TABLE 5. Number of patients receiving medications that may have contributed to lowering of temperature (ITT analysis)

Among the population in the ITT analysis, the mean ±SD duration of i.v. treatment for moxifloxacin-treated patientswas 4.02 ± 1.78 days, whereas it was 4.81± 2.07days for the comparator regimen-treated patients, suggestingthat moxifloxacin-treated patients were able to switch to oraltherapy at an earlier stage than those who were receiving thecomparator regimen. Among the patients in the ITT analysis,approximately half (n = 151 [50.2%]) of the patients in themoxifloxacin group received therapy for only 3 days or less,whereas less than a fifth (n = 57 [17.8%]) of the patients treatedwith the comparator drug received therapy for only 3 days orless.

The mean ± SD duration of the hospital stay was 9.49± 7.29 days for the moxifloxacin group (n = 301) and10.41 ±7.49 days for the comparator group (n = 321).The duration of the hospital stay was not the subject of formalstatistical testing, as in the study protocol this parameterwas planned for descriptive analysis only.

Tolerability and safety. Data for 301 moxifloxacin-treated patients and 321 comparatorregimen-treated patients were valid for the safety analysis.Within the latter group, 192 patients (60%) were treated withboth co-amoxiclav and clarithromycin and 129 patients receivedco-amoxiclav alone.

Fifteen patients (5.0%) in the moxifloxacin group and 13 patients(4.0%) in the comparator group were prematurely discontinuedfrom the study due to adverse events. The causes for discontinuationpossibly or probably related to treatment that occurred in morethan one patient included abnormal liver function tests (LFTs)or hepatic disturbance (one in the moxifloxacin group, fivein the comparator group); rash (two in the moxifloxacin group,one in the comparator group); allergic reaction (two in themoxifloxacin group); diarrhea (two in the moxifloxacin group);and pancreatitis and elevation of pancreatic enzyme levels (onein the moxifloxacin group, two in the comparator group). Noother single event caused a high number of discontinuationsin either group. Single events possibly or probably relatedto a study drug that led to discontinuation included dizziness,nausea, acute gastroenteritis, fever, and confusion for themoxifloxacin group and pruritus and gastralgia for the comparatorgroup. All drug-related adverse events that led to discontinuationimproved or resolved without serious sequelae.

During the course of the study there were 26 deaths: 9 in themoxifloxacin group (2 of these occurred after the 30-day follow-up)and 17 in the comparator group. This was not a statisticallysignificant difference. Most deaths were due to pneumonia orunderlying lung disease (6 in the moxifloxacin group and 10in the comparator group). The average age at death was 76 yearsin the moxifloxacin group and 71 years in the comparator group.Apart from death, other serious adverse events were definedas a life-threatening drug experience, hospitalization or prolongationof existing hospitalization, a persistent or significant disabilityor incapacity, and important medical events that required medicalor surgical intervention. Serious adverse events were reportedin 38 patients (12.6%) treated with moxifloxacin and 53 patients(16.5%) treated with the comparator regimen. There were moreserious adverse events within the respiratory system for bothtreatment groups (18 [6.0%] in the moxifloxacin group, 25 [7.8%]in the comparator group). Twelve patients (4.0%) in the moxifloxacingroup and 20 patients (6.2%) in the comparator group were reportedto have serious cardiovascular adverse events.

When all body systems were taken into account, the most common(incidence, $>=$ 3.0%) drug-related adverse events were abnormalLFTs, diarrhea, and nausea for the moxifloxacin group and abnormalLFTs, diarrhea, nausea, and phlebitis for the comparator group.There were no cases of phototoxicity. The incidence of drug-relatedadverse events was highest for those events related to the gastrointestinaltract and was similar in both groups (20.9% for the moxifloxacingroup and 22.1% for the comparator group). The most frequentgastrointestinal events were abnormal LFTs, diarrhea, and nausea.The types and severities of events were similar between thetwo treatment groups. The distributions of mild, moderate, andsevere abnormal LFTs were comparable between the two treatmentgroups (15 mild, 6 moderate, and 1 severe LFTs for the moxifloxacingroup; 9 mild, 9 moderate, and 1 severe LFTs for the comparatorgroup). There were fewer drug-related cardiovascular eventsin the moxifloxacin group (20 [6.6%]) than in the comparatorgroup (32 [10%]). There was only one report of a treatment-emergentclinically significant QT_c prolongation, and this was in a patienttreated with co-amoxiclav. There were no cases of torsade depointes.

The vast majority of drug-related adverse events in both treatmentgroups were mild or moderate, irrespective of whether the studymedication was administered by the i.v. or oral route, and theirincidences were as follows: for moxifloxacin administered i.v.,70 of 79 (89%) patients; for moxifloxacin administered orally,43 of 50 (86%) patients; for the comparator regimen administeredi.v., 77 of 88 (88%) patients; for the comparator regimen administeredorally, 38 of 38 (100%) patients. The incidence of drug-relatedadverse events was 38.9% in both treatment groups, and theiroutcomes were similar. Thus, the proportions of drug-relatedadverse events described as resolved or improved were 84% (98of 117) for the moxifloxacin group and 85% (106 of 125) forthe comparator group.

Most adverse events in both treatment groups were experiencedduring the period of i.v. administration. During i.v. therapy,the incidence rates for all drug-related adverse events were26.2% for the moxifloxacin group and 27.4% for the comparatorgroup. The incidences of drug-related adverse events duringoral therapy were 16.6% for the moxifloxacin group and 11.8%for the comparator group.

High and low laboratory values were defined as those which occurredin $>=$ 2.5% of the study population in at least one of the two treatmentgroups. In addition, there were predefined criteria for clinicallysignificant changes of selected laboratory values (decreaseof the hematocrit level of at least 20%, decrease of the hemoglobinconcentration of at least 2 g/liter, decrease of the plateletcount of at least 25%, increase of the serum creatinine levelof at least 0.6 mg/dl, increase of the blood urea nitrogen levelof at least 75%, increase of the aspartate aminotransferaselevel of at least 100% and at least 10 U/liter, increase ofthe alanine aminotransferase level of at least 100% and at least10 U/liter, increase of the alkaline phosphatase level of atleast 50% and at least 50 U/liter, and increase of the totalbilirubin level of at least 200%). For the moxifloxacin andthe comparator groups, the incidences of high abnormal valueswere 17.5 and 20.3%, respectively; the incidences of low abnormalvalues were 16.9 and 17.6%, respectively; and the incidencesof selected clinically significant values were 9.8 and 10.9%,respectively. The types of abnormalities observed were not differentbetween the two treatment groups, and no unexpected abnormalitieswere observed.

DISCUSSION

Top

Discussion

Because the rate of mortality is significantly increased inpatients with severe CAP and because it may be caused by a numberof different pathogens unknown at the time of presentation,there are compelling reasons for empirical treatment. It isoften difficult to identify a specific etiology for CAP sincethere is no distinctive pattern of signs and symptoms for anindividual group of pathogens or an individual species of pathogenand since appropriate diagnostic tests fail to identify a pathogenin approximately 50% of cases. Because of these diagnostic difficulties,combination therapy is often used to cover all common pathogensthat cause CAP. Moxifloxacin is a new fluoroquinolone with activityagainst the common gram-positive, gram-negative, and atypicalpathogens that cause CAP. Thus, it is an ideal monotherapeuticregimen for the empirical treatment of CAP. The purpose of thisstudy was to compare sequential i.v. and oral moxifloxacin therapyto a standard regimen of co-amoxiclav with or without clarithromycin.

Due to its complexity, an open-label study design was used.Although an open-label design introduces the possibility ofbias, this was minimized by using a standardized methodologyto determine the clinical response to treatment. The objectivityof these clinical judgments was supported by consistency betweenthe clinical and bacteriological outcomes, including the outcomesfor objective criteria such as the resolution of fever.

The microbial epidemiology of the study reflects those describedin previous reports. Causative organisms were identified in59.5% of patients for whom cultures or serology was performed.It is known that no specific cause is found in as many as 30to 60% of patients even when extensive diagnostic testing isperformed (15). In patients with documented infections, themost frequently isolated organism was S. pneumoniae (55.4% ofidentified pathogens), followed by H. influenzae (19.6%). Atypicalorganisms accounted for 13.8% of infections, a result that isconsistent with those from a series of three well-defined studies(12, 13, 14) in which the incidence ranged from 2.7 to 10% inpatients requiring admission to a hospital.

The rate of clinical cure was significantly higher in patientstreated with moxifloxacin than in patients treated with thecomparator regimen. Our results are consistent with those ofa previous study (8) of levofloxacin compared with ceftriaxoneand/or cefuroxime axetil with or without erythromycin or doxycycline.However, in the levofloxacin study, most patients (84%) hadnonsevere CAP and the difference between treatment groups wasless marked (clinical success rate for the levofloxacin group,96%; clinical success rate for the group receiving the comparatorregimen, 90%). Unlike the present study, the observed differencewas not subject to a formal test of superiority. It is of notethat a retrospective examination (10) of the hospital recordsfor 12,945 patients, aged 65 years and older, hospitalized forpneumonia showed that initial treatment with an expanded-spectrumcephalosporin or a nonpseudomonal broad-spectrum cephalosporintogether with a macrolide or fluoroquinolone monotherapy wasindependently associated with lower 30-day mortality rates.The use of a ß-lactam-ß-lactamase inhibitorwith a macrolide or an aminoglycoside with a macrolide was associatedwith an increased 30-day mortality rate.

In an equivalence or noninferiority trial, similar conclusionsfrom both the ITT and PP analyses are required for a robustinterpretation (7). Since the results presented here for theITT analyses were the same as those for the primary outcomemeasure and other secondary clinical outcomes in the PP analyses,it is reasonable to conclude that the differences between treatmentgroups are likely to be medically significant. This interpretationis also supported by consistent differences in favor of moxifloxacinfor those outcomes for which, due to a lack of statistical power,formal analysis was inappropriate.

The time to resolution of fever was significantly shorter formoxifloxacin-treated patients. Although the proportions of patientswith a fever, as measured by various routes (oral, tympanic,axillary, or rectal), were similar in both treatment groups,in order to take account of any differences, a conservativedefinition of apyrexia, namely, a body temperature of $<=$ 37.5°C,was used in the main analysis. However, it must be emphasizedthat even when the definition of resolution of fever was considereda body temperature of $<=$ 38°C, the result was still statisticallysignificant. Furthermore, the rate of use of antipyretics wasless among patients treated with moxifloxacin than among thosegiven the comparator regimen and suggests a direct treatmenteffect.

The observation of a shorter time to resolution of fever isconsistent with a more rapid onset of treatment effect and issupported by the observation that approximately half (50.2%)of the moxifloxacin-treated patients received i.v. therapy foronly 3 days or less, whereas less than a fifth (17.8%) of thepatients treated with the comparator regimen received i.v. therapyfor only 3 days or less. After 3 days, the time at which patientswere converted to oral therapy was entirely at the discretionof the investigator. The mean duration of hospital admissionfor the moxifloxacin-treated patients was approximately 1 dayless than that for those treated with the comparator regimen.

Clinical trials of earlier conversion from i.v. to oral therapy(17, 18) have demonstrated that the duration of hospital admissioncan be reduced without compromising patient outcomes. The PneumoniaPatient Outcomes Research Team cohort study (9) concluded that,despite differences between hospitals in the total costs ofcare, the patterns of daily resource use were similar and thata 1-day reduction in the length of stay might yield substantialcost savings. A preliminary analysis of data from that studyfound health economic benefits of moxifloxacin therapy overcomparator therapy (M. Drummond, R. Finch, I. Duprat-Lomon,P. P. Sagnier, O. Collins, R. Kubin, and M. Hux, Abstr. 41stIntersci. Conf. Antimicrob. Agents Chemother., abstr. 864, 2001).The clinical response at late follow-up, that is, 21 to 28 daysafter the end of treatment, demonstrates that the shorter periodof hospitalization does not affect the longer-term outcomesfor moxifloxacin-treated patients.

The potential cost advantage of monotherapy with a fluoroquinoloneover combination treatment with a ß-lactam and a macrolidehas recently been confirmed by an independent cost analysis(5). That work examined the results of a prospective randomizedclinical trial of sequential i.v. or oral gatifloxacin therapycompared with ceftriaxone and erythromycin or clarithromycintherapy in patients with CAP. The analysis took account of thecosts of drug acquisition, dispensing, administration, treatmentfailure, adverse events, and hospitalization. Despite the similarperiods of i.v. drug administration and the lack of statisticallysignificant differences in clinical and microbiological outcomes,gatifloxacin was shown to be more cost-effective.

When one is selecting an antimicrobial agent for the treatmentof CAP, it is important to consider not only efficacy but alsohow well it is likely to be tolerated by patients. In the trialdescribed here, sequential i.v. and oral treatment with bothregimens, moxifloxacin or co-amoxiclav with or without clarithromycin,was well tolerated. Most adverse events in both treatment groupswere considered to be of mild to moderate intensity, regardlessof the mode of administration.

There were fewer serious adverse events and deaths in the moxifloxacingroup than in the comparator group. The smaller number of deathsin patients treated with moxifloxacin is consistent with a previousmeta-analysis of three comparative studies of the treatmentof CAP with oral moxifloxacin [M. Niederman, D. Church, J. Kaufmann,and M. Springsklee, abstract from Forum on Respiratory Infections2000, Respir. Med. 94(Suppl. A):A14, 2000]. The analysis wasof all causes of mortality within 30 days posttreatment andshowed that in the cohort of 701 patients treated with moxifloxacinthere were 4 deaths (0.57%), whereas there were 12 deaths (1.70%)in the 705 patients treated with the comparator regimen (P =0.045). The incidences of laboratory abnormalities and prematurediscontinuations were similar in both groups.

The higher incidence of cardiovascular events seen in patientstreated with the comparator regimen was largely due to a higherrate of phlebitis. There were no clinically significant (6)cases of treatment-emergent QT_c prolongation in the moxifloxacingroup, and there was only one case in the comparator group.There were no cases of torsade de pointes. Although it has beenreported that several quinolones cause phototoxic reactions,no patients complained of phototoxicity. This is consistentwith the safety experience reported with oral moxifloxacin (R.Kubin and C. Reiter, Abstr. 40th Intersci. Conf. Antimicrob.Agents Chemother., abstr. 820, 2000).

The present study was in accord with existing European (2, 4,19) and North American (1) guidelines for the evaluation ofantibacterial agents, which allow the use of study designs inwhich assessment times are determined by clinical judgment.Such study designs fail to take account of the rate of responseand favor the demonstration of equivalence rather than differencesbetween treatments. This is further compounded by definitionsof equivalence or statistical hypotheses that allow greaterdifferences between treatment effects. In this study the assessmentperiods were determined by the point at which the patient'scondition had improved to the extent that the patient was wellenough for treatment to be discontinued. Hence, it might beanticipated that the clinical statuses of patients at thesetimes might not be dissimilar. Notwithstanding these considerations,the results of this study, designed to demonstrate equivalence,show that moxifloxacin was statistically significantly moreeffective, in terms of clinical and bacteriological outcomes,than a standard combination regimen included in a number oftreatment guidelines for CAP. This superiority was seen in patientsirrespective of the severity of their pneumonia and whetheror not the combination therapy included a macrolide. The highrate of bacteriological success of moxifloxacin, observed atthe TOC visit, against those organisms known to be importantcauses of CAP, including drug-resistant S. pneumoniae, was maintainedat late follow-up. The results for the primary outcome, clinicalresponse, were supported by the faster resolution of fever andthe earlier conversion from i.v. to oral administration in patientstreated with moxifloxacin. Moxifloxacin treatment was associatedwith fewer deaths, fewer serious adverse events, and fewer otheradverse events. Given that data from prospective, randomizedtrials are the major source of information for evidence-basedmedicine, it is concluded that the results presented here mayhave implications for clinical guidelines.

ACKNOWLEDGMENTS

We acknowledge the work of the following investigators: in Belgium,R. Peleman, D. Galdermans, and E. Janssens; in France, O. Ruyer,G. Dien, I. Caby, C. D. Guyonnaud, L. Vives, X. Forecville,L. Cellerin, D. Debieuvre, P. Delaval, M. David, J. Boyer, andB. Fantin; in Germany, S. Kljucar, H. Koch, P. M. Shah, A. Schwarz,L. Greiner, L. Bölcskei, J. Schauer, H. Lode, W. Kreisel,and O. Cornely; in Greece, H. Bassaris and E. Papadakis; inSpain, F. Álvarez Lerma and G. Juan Samper; in the UnitedKingdom, P. Rafferty, R. C. Read, P. B. Anderson, D. R. T. Shepherd,G. S. Basran, M. Farrington, R. J. O. Davies, K. G. Nicholson,R. N. Davidson, and C. G. Wathen; in Israel, E. Rubinstein,R. K. Cohen, and D. Seltzer; in Switzerland, J.-D. Baumgartner,J. Garbino, and D. Genné; and in South Africa, L. Fourie,J. G. Kilian, C. Feldman, T. J. Gray, J. H. J. van Rensburg,A. Jacovides, M. J. Heystek, N. Du Toit.

FOOTNOTES

* Corresponding author. Mailing address: Division of Microbiology & Infectious Diseases, The Clinical Sciences Building, Nottingham City Hospital, Hucknall Rd., Nottingham NG5 1PB, United Kingdom. Phone: 44 115 840 4741. Fax: 44 115 840 4742. E-mail: r.finch{at}nottingham.ac.uk.

REFERENCES

Top