Technology Report. New Fluoroquinolones in Community- Acquired Pneumonia: A Clinical and Economic Evaluation. Issue 20 November 2001

Transcription

1 Technology Report New Fluoroquinolones in Community- Acquired Pneumonia: A Clinical and Economic Evaluation Issue 20 November 2001

2 Publications can be requested from: CCOHTA Green Valley Crescent Ottawa, Ontario, Canada, K2C 3V4 Tel. (613) Fax. (613) or download from CCOHT A s web site: Cite as: Metge CJ, Vercaigne L, Carrie A, Zhanel GG. New fluoroquinolones in communityacquired pneumonia: a clinical and economic evaluation. Ottawa: Canadian Coordinating Office for Health Technology Assessment ; Technology report no 20. Reproduction of this document for non-commercial purposes is permitted provided appropriate credit is given to CCOHTA. Legal Deposit National Library of Canada ISBN (print) ISBN (online) Publications Mail Agreement Number:

3 Canadian Coordinating Office for Health Technology Assessment New Fluoroquinolones in Community-Acquired Pneumonia: A Clinical and Economic Evaluation Colleen J. Metge, B.Sc.(Pharm.), Ph.D. 1,2 Lavern Vercaigne, Pharm.D. 1 Anita Carrie, B.Sc.(Pharm.) 1 George G. Zhanel, Ph.D. 1,2,3 November Faculty of Pharmacy, University of Manitoba, Winnipeg, MB 2 Faculty of Medicine, University of Manitoba, Winnipeg, MB 3 Microbiology and Medicine, Health Sciences Centre, Winnipeg, MB

4

5 REVIEWERS These individuals kindly provided comments on this report. CCOHTA Scientific Advisory Panel Dr. Gina Bravo Sherbrooke University Geriatric Institute Sherbrooke, Quebec Dr. Phillip Jacobs University of Alberta Department of Public Health Sciences Faculty of Medicine and Oral Health Sciences Edmonton, Alberta Dr. Ruth Collins-Nakai University of Alberta Edmonton, Alberta Dr. Murray Krahn The Toronto Hospital Toronto, Ontario External Reviewers Dr. John Conoly University of Toronto University Health Network Toronto, Ontario Dr. Wayne Gold Toronto Hospital General Hospital Toronto, Ontario Dr. Thomas Marrie University of Alberta Department of Medicine Edmonton, Alberta Bristol-Myers Squibb Québec Bayer Ontario Jansen-Ortho Inc. Ontario Smith Kline Beecham Ontario This report is a review of existing public literature, studies, materials and other information and documentation (collectively the source documentation ) which are available to CCOHTA. The accuracy of the contents of the source documentation on which this report is based is not warranted, assured or represented in any way by CCOHTA and CCOHTA does not assume responsibility for the quality, propriety, inaccuracies, or the reasonableness of any statements, information or conclusions contained in the documentation. CCOHTA takes sole responsibility for the final form and content of this report. The statements and conclusions in this report are those of CCOHTA and not of its Panel members or reviewers. i

6 Authorship The integrity of this work in its entirety is the responsibility of Colleen Metge, Lavern Vercaigne and George Zhanel. All authors participated in one or more of the following (1) a substantial contribution to conception and design (Metge, Vercaigne, Zhanel), or acquisition of data (Metge, Carrie), or analysis and interpretation of data (Metge, Vercaigne, Zhanel, Carrie); (2) drafting the report or revising it critically for important intellectual content (Metge, Vercaigne, Zhanel, Carrie); and (3) final approval of the version to be published (Metge, Vercaigne, Zhanel, Carrie). Acknowledgments Vikram Sarveiya (University of Manitoba) should be recognized for substantively contributing to the design and authoring of the economic analysis in this review. The authors are also grateful to Michel Boucher (CCOHTA), Bruce Brady (CCOHTA), Janet Joyce (CCOHTA), Dr. Nicolaas Otten (previously at CCOHTA), Kelly Cugnet and Kimi Guilbert (University of Manitoba) and Dr. Tom Einarson (University of Toronto) for their assistance in the preparation of this manuscript. Disclosures of Conflicts of Interest Colleen Metge holds a chair in evaluating drug therapy effectiveness from Bristol-Myers Squibb Canada Inc., the manufacturer of a new fluoroquinolone, gatifloxacin, which is included in the clinical review but not in the economic analysis. Lavern Vercaigne completed a two-year fellowship in 1999 which was sponsored by GlaxoSmithKline (formerly Glaxo Wellcome), the developer of grepafloxacin. George Zhanel is sponsored in one or more endeavours (research grants, a position on a medical advisory board or paid honoraria for speaking) by the following developers of new fluoroquinolones: Bayer Inc. (moxifloxacin), Bristol-Myers Squibb Canada Inc. (gatifloxacin), Janssen-Ortho Inc. (levofloxacin),glaxosmithkline (gemifloxacin), Aventis Pharma formerly Rhone-Poulenc Rorer Inc. (sparfloxacin), Pfizer Canada Inc. (trovafloxacin). ii

7 EXECUTIVE SUMMARY Objectives: To critically review and compare, both clinically and economically, the new fluoroquinolones a for the empiric treatment of community-acquired pneumonia (CAP). Clinical Efficacy Review Methods: A systematic review and meta-analyses of randomized controlled trials were used to compare the bacteriological and clinical efficacy of new fluoroquinolones (FQ) with current comparator antibiotics for the treatment of CAP. Across all studies, the main outcome measures were both the number of subjects designated clinically cured or improved [=clinical success] at the end of treatment and the number of subjects designated clinically cured or improved based on causative pathogen [S. pneumoniae, H. influenzae and atypical organisms (M. pneumoniae, C. pneumoniae, Legionella pneumophilia)]. Using meta-analyses, estimates of the differences in clinical success rates between new fluoroquinolones (FQs) and comparator antibiotics were expressed as risk differences (RD, 95% confidence interval (CI)). Results: A total of 16 studies met the selection criteria. Twelve studies with a total of 3634 subjects compared orally-administered new fluoroquinolones with comparative orallyadministered antibiotics for CAP; eight of these studies reported (n=3131) intention-to-treat [ITT] analyses. Three studies (n=914) compared intravenous (IV) or intravenous with step down to oral (IV/oral) fluoroquinolone therapy for CAP with a comparative IV or IV/oral CAP antibiotic; none reported an ITT analysis. One study compared an IV/oral fluoroquinolone to another IV/oral fluoroquinolone (gatifloxacin versus levofloxacin) (n=417). Only four of the studies involved fluoroquinolones approved for use in Canada (gatifloxacin, levofloxacin and moxifloxacin), and a further five studies involved the fluoroquinolone trovafloxacin, which is restricted for use in Canada because it is associated with liver toxicity. Quality assessment of the studies showed that one study was of high quality, nine were of moderate quality, and six were of low quality; all three studies of IV/oral treatment were of low quality. No statistically significant difference in clinical success was found [RD=1.7%; 95% CI ( 1.4, 4.8%), ns] between orally administered FQs and comparators using a meta-analysis of the ITT results. An evaluable subjects meta-analysis found the new FQs administered orally to be slightly more effective than comparative antibiotics [RD= 2.9%; 95% CI ( %)], and the FQs administered IV/orally were found to be more effective than comparative IV/oral antibiotics [RD=5.4%; 95% CI ( %)]. An evaluable subjects meta-analysis showed no significant risk difference for studies involving oral administered FQs that are approved for use in Canada (gatifloxacin and moxifloxacin) [RD = 0.2%; 95% CI ( 4.8 to 4.5%)], and for those studies involving oral administered trovafloxacin, which is restricted for use in Canada [RD = 2.2%; 95% CI ( 0.7 to 5.1%)]. a The new or respiratory fluoroquinolones for the purposes of this study included: gatifloxacin, grepafloxacin, levofloxacin, moxifloxacin, sparfloxacin and trovafloxacin (see description in Table 4.) iii

8 Economic Analysis Methods: The primary economic analysis of the oral and IV/oral treatments were based on a cost-minimization analysis (CMA), which assumes equivalent efficacy between the new FQs and comparator treatments because of the uncertainty and limitations of the clinical evidence. Outpatients were treated for 10 days with one of the two comparators. In this type of analysis, the least costly drug treatment is selected as the treatment of choice. Also, a clinical decision model was constructed to compare the cost-effectiveness of oral and IV/oral treatments from a provincial government perspective. In the cost-effectiveness analysis (CEA), the efficacy results (risk differences) for the new FQs were taken from the evaluable subjects analysis, and levofloxacin and clarithromycin were selected as the basis for comparing the costs of FQ vs the comparator. The incremental cost per additional successfully treated patient was used as the primary economic outcome measure (cases showing savings are reported in aggregate). Results: In the CMA, oral treatment for 10 days with one of the three fluoroquinolones (gatifloxacin, levofloxacin or moxifloxacin) approved for use in Canada provides savings of $10 per patient treated over oral treatment with clarithromycin, and $18 savings compared to oral treatment with cefuroxime axetil plus erythromycin. However, the cost advantage of oral treatment with these three fluoroquinolones is lost (i.e. no savings or net cost) when compared to oral treatment with amoxicillin / clavulanic acid plus erythromycin. IV/oral treatment with levofloxacin provides savings of $49 per patient over IV/oral treatment with the comparator, ceftriaxone/erythromycin. Considering that the CEA was based on the results of the evaluable subjects analysis and a more expensive comparator (clarithromycin), the new FQs used orally and IV/orally for CAP are costeffective. For 1000 individuals at low risk of death from CAP, results show 38 more cures for $39,000 less when they are treated initially with FQs versus a comparator antibiotic. Savings per treated low-risk patient are substantially greater for those treated in-hospital ($91) than out of hospital ($13). For 1000 individuals at moderate risk of death from CAP, results show 51 more cures for $83,000 less when they are treated initially with FQs versus a comparator antibiotic. Savings per treated moderate-risk patient are somewhat greater for those treated in-hospital ($85) than out of hospital ($61). The results are sensitive to assumptions about the relative efficacy and cost of the new FQs and comparator treatments. Conclusion Clinical Efficacy: Analysis of the trials on an intention-to-treat basis, which preserves the value of randomization and prevents the introduction of bias, indicates that the orally-administered new fluoroquinolones offer no statistically significant additional clinical successes when compared to current comparative antibiotics for the empirical treatment of CAP. The evaluable subjects analysis found new fluoroquinolones to be slightly more effective in treating individuals diagnosed with CAP than comparative antibiotics. However, considering the limitations associated with the available evidence, we conclude that the new fluoroquinolones are at least as effective as, and may be slightly more effective than, comparator antibiotics for the treatment of empiric CAP. Finally, our review did not point to any overall differences in terms of serious adverse effects, but these would be unlikely to occur in relatively small studies. iv

9 Cost analysis: The CEA from a provincial government payer perspective, suggests that the new fluoroquinolones used orally and IV/orally for CAP may be more effective and cost less than comparative antibiotics. However, these results should be interpreted with much caution given that this analysis is based on only one comparator (clarithromycin) and the results of the less rigorous evaluable subjects analysis, which may exaggerate the effect of the intervention, since we cannot account for study dropouts. For the economic analysis, we prefer the new fluoroquinolones to be considered "neutral" in terms of efficacy relative to comparators, because of the uncertainty and limitations of the available clinical evidence. Therefore, we consider the CMA to be the primary economic analysis, based on the assumption of therapeutic equivalence of the fluoroquinolones and alternative antibiotic treatments. The CMA, also conducted from a provincial government perspective, indicates that the new fluoroquinolones approved for use in Canada have a cost advantage when compared to some recommended alternative antibiotics for the outpatient (oral) treatment of CAP, but that this cost advantage is lost when other recommended alternative treatments are considered. Given that differences in drug acquisition costs may be relatively minor, decisions about choice of empirical treatment of CAP may involve other considerations, such as the regional pattern of bacterial resistance, the adverse drug reaction profile and patient convenience. Concerns about potential cross-resistance among fluoroquinolones resulting from overuse should also be considered. v

10 vi

11 TABLE OF CONTENTS EXECUTIVE SUMMARY...iii 1. INTRODUCTION Study Objectives Community-acquired pneumonia Diagnosis of community-acquired pneumonia Treatment of Community-acquired Pneumonia Pharmacokinetic/dynamic review of new fluoroquinolones Outcomes of Community-acquired Pneumonia Economic evaluations in community-acquired pneumonia METHODOLOGY Clinical Review Inclusion criteria Literature search Methods for review Data extraction Statistical analysis Economic Review Inclusion of economic evaluation studies Critical appraisal Economic analysis RESULTS Clinical Review Results Selection and quality assessment of studies Studies comparing new fluoroquinolones to new fluoroquinolones Studies comparing orally administered new fluoroquinolones and comparator antibiotics Studies comparing intravenously/orally-administered new fluoroquinolones and comparator antibiotics Economic Analysis Results Selection and quality assessment of economic studies Original economic analysis results DISCUSSION Clinical Efficacy Review Economic Analysis CONCLUSION REFERENCES...34 vii

12 APPENDIX I: MEDLINE / EMBASE DATABASE SEARCH STRATEGY...41 APPENDIX II: Standardized Data Quality Form...43 APPENDIX III: Abstraction Form...44 APPENDIX IV: Estimating Inpatient RDRG Costs...47 APPENDIX V: Studies included in Clinical Review of New Fluroquinolones...48 APPENDIX VI: List of Excluded Studies...51 FIGURE 1: Decision Tree for Treatment of Community-Acquired Pneumonia...53 FIGURE 2: Decision Model for the Management of Community-Acquired Pneumonia...54 FIGURE 3: Decision Model for the Management of Community-Acquired Pneumonia...55 TABLE 1: Compilation of CAP Treatment Guidelines...56 TABLE 2: Selection criteria for clinical review...57 TABLE 3: Costs used in economic analysis...58 TABLE 4: A description of the orally-administered fluoroquinolones and comparators...59 TABLE 5: Results of meta-analysis of different studies comparing orrallyadministered fluoroquinolones and comparator antibiotics...60 TABLE 6: Results of meta-analysis of different studies comparing I/V / orallyadministered fluoroquinolones and comparator antibiotics...61 TABLE 7: Consequences of patients at LOW RISK (Level I to III) for CAP mortality expressed as probability (Pr) of occurrence...62 TABLE 8: Costs attributable to patients at LOW RISK (Level I to III) and MODERATE RISK (Level IV) for CAP mortality...63 TABLE 9: Costs and consequences for each of the decision tree branches (patients at LOW RISK for CAP mortality)...64 TABLE 10: Costs and consequences (effectiveness) of treating patients at LOW RISK (Level I to III) for CAP mortality with fluoroquinolones...65 TABLE 11: Consequences of patients at MODERATE RISK (Level IV) for CAP mortality expressed as probability (Pr) of occurence...66 TABLE 12: Costs and Consequences for each of the decision tree branches (patients at MODERATE RISK for CAP mortality)...67 TABLE 13: Cost-consequence (effectiveness) of treating patients at MODERATE RISK (Level IV) for CAP mortality with fluoroquinolones...68 TABLE 14: Sensitivity Analyses (LOW RISK)...69 TABLE 15: Sensitivity Analyses (MODERATE RISK)...70 TABLE 16: Results of the one-way sensitivity analyses comparing FQ to Other for those at LOW RISK of mortality from CAP and hospitalized...71 TABLE 17: Results of the one-way sensitivity analyses comparing FQ to Other for those at LOW RISK of mortality from CAP and treated as an outpatient...72 viii

13 TABLE 18: Results of the one-way sensitivity analyses comparing FQ to Other for those at MODERATE RISK of mortality from CAP and hopitalized...73 TABLE 19: Results of the one-way sensitivity analyses comparing FQ to Other for those at MODERATE RISK of mortality from CAP and treated as an outpatient...74 TABLE 20: Results for the cost minimization analysis for outpatient (oral) treatment for 10 days*...75 ix

14 x

15 1. INTRODUCTION The continual reassessment of antibiotic use is needed due to concerns about the increasing resistance to currently available antibiotics. 1-3 Due diligence in the prescribing of antibiotics is necessary. Due diligence means appropriate use; that is, both not prescribing an antibiotic when one is not needed, as well as prescribing an antibiotic appropriately according to morbidity and mortality risk when an antibiotic is needed. However, appropriate antibiotic prescribing for community-acquired pneumonia (CAP) is particularly difficult. Diagnostic cultures are rarely done in the community setting making due diligence a difficult prospect for the clinician treating an individual with CAP. In addition, CAP can be multifactorial, present differently and follow a different course depending on the presumed causative factors (e.g., viral or bacterial etiology). It is, therefore, a disease subject to and dependent upon, empiric or trial-and-error initial treatment choice. A new group of antibiotics, the respiratory or new fluoroquinolones is now available. Their improved in vitro activity against respiratory pathogens (e.g. Streptococcus pneumoniae or S.pneumoniae) and improved pharmacokinetic properties necessitates study to determine their place in the treatment of CAP. 1.1 Study Objectives The overall goal of this report is to compare and contrast the use of new fluoroquinolones, including their cost-effectiveness, against usual or current best treatment for CAP in Canada. b The outcomes of treatment of CAP by each class of antibiotic (new fluoroquinolone versus comparators) were quantified in terms of clinical success (cure or improvement). Secondarily, these outcomes were further stratified, where possible, with respect to: (a) the specific antibiotic used for CAP treatment, (b) route of drug administration (oral, or intravenous only or intravenous followed by oral), (c) causative organism, and (d) patient type (e.g., elderly and comorbidity). The specific objectives of this study were to: 1. Compare the efficacy and safety of new fluoroquinolones for the treatment of CAP by performing a meta-analysis (or undertaking a critical literature synthesis) of randomized controlled trials. 2. Compare the efficacy and safety of new fluoroquinolones with other antibiotics (e.g. macrolides, penicillins, cephalosporins, tetracyclines) in the treatment of CAP by performing a meta-analysis of randomized controlled trials (or undertaking a critical literature synthesis of the best available evidence). b The new or respiratory fluoroquinolones include gatifloxacin, gemifloxacin, grepafloxacin, levofloxacin, moxifloxacin, sparfloxacin and trovafloxacin. When this report was written, only levofloxacin was approved for use in Canada; gatifloxacin and moxifloxacin have been subsequently approved. Grepafloxacin has been withdrawn from the market because of concerns about cardiac toxicity, and trovafloxacin is restricted becaus e of concerns about severe liver toxicity. 1

16 3. Estimate the rates of adverse events (AEs) that occur following treatment with various antibiotics for CAP by performing a meta-analysis (or undertaking a critical literature synthesis) of randomized clinical trials. 4. Evaluate the available studies comparing the cost-effectiveness of the new fluoroquinolones to other antibiotics for the treatment of CAP by performing a critical appraisal of available cost-effectiveness analyses. 5. Compare the cost-effectiveness of new fluoroquinolones with other antibiotics used to treat CAP where effectiveness is based upon rates of clinical success and special population attributes (e.g. elderly, comorbidity). A cost minimization analysis (CMA) was also performed because the rates of clinical success between the two treatment strategies were assumed to be equivalent. 1.2 Community-acquired pneumonia Despite the availability of potent new antimicrobials and vaccines, CAP is a common and serious illness. CAP is associated with significant morbidity, mortality and use of health care resources. It is an acute, lower respiratory tract infection of the pulmonary parenchyma (or cells lining the lung) caused by bacteria, viruses and/or fungi. Causative bacterial pathogens are often grouped into typical (e.g., S.pneumoniae, Haemophilus influenzae) and atypical organisms (Mycoplasma pneumoniae, Chlamydia pneumoniae, Legionella pneumophilia). Since CAP is not a reportable disease in Canada, most of the statistics available on the incidence of this condition come from the United States and Europe. It is estimated that as many as four million cases of CAP occur annually in the United States, 4 this is about 12 cases per thousand per year. Figures from the United Kingdom and Sweden suggest rates of CAP between 1 and 5 cases per thousand per year. 5 Assuming that the Canadian incidence of CAP is somewhere between Europe's and the United States, then between 146,000 and 350,000 cases occur annually in Canada. The most recent Canadian statistics available report that 78,400 cases of pneumonia were treated in hospital in , accounting for 2.2% of all hospital stays. 6 Pneumonia is the seventh leading cause of death in Canada and is ranked directly behind ischemic heart disease, lung, breast and prostate cancer, stroke and chronic lung disease. According to Statistics Canada, 7,728 deaths were attributed to pneumonia in A number of Canadian studies have reported significant morbidity and mortality from pneumonia in the elderly and in those with comorbid conditions such as diabetes mellitus, chronic obstructive pulmonary disease, chronic renal failure, congestive heart failure, chronic alcohol abuse and acquired immuno-deficiency syndrome, as well as individuals with multiple medication use According to Canadian mortality statistics, an average of 3.2% deaths were due to pneumonia and influenza for the last quarter century. In 1997, deaths from pneumonia alone (abbreviated death codes A ) were 25.8 per 100,000 deaths. 7 2

17 1.3 Diagnosis of community-acquired pneumonia Despite remarkable advances in the identification of new microbial pathogens and antimicrobial agents, few diseases are so characterized by disputes about diagnostic evaluation and therapeutic decisions. 13 Bartlett and Mundy s words about new microbial pathogens (atypical bacteria), new antimicrobial agents (including respiratory fluoroquinolones) and CAP are echoed in almost all articles that review the infectious constellation we call CAP However, the latest guidelines from the Infectious Diseases Society of America define it accordingly: 14 Community-acquired pneumonia associated with at least some symptoms of acute infection, accompanied by the presence of an acute infiltrate on a chest radiograph or auscultatory findings consistent with pneumonia (altered breath sounds and/or localized rales), in a patient who is not hospitalized or residing in a long-term care facility for 14 days before the onset of symptoms. Symptoms of acute lower respiratory tract infection may include several (in most studies, at least 2) of the following: fever hypothermia, rigors, sweats new cough with or without sputum production change in the color of respiratory secretions in a patient with chronic cough chest discomfort the onset of dyspnea (shortness of breath). Most persons also have nonspecific symptoms such as fatigue, myalgias, abdominal pain, anorexia, and headache. At the point of diagnosis in the community setting, the pathogen causing CAP for a specific patient is unknown. Indeed, it has been difficult to establish an association between individual symptoms, physical findings or laboratory test results and specific causative pathogens; thus, the empiric approach for the initial treatment of CAP. 18 Regardless, it is recommended that physicians should suspect pneumonia in individuals with newly acquired lower respiratory symptoms like cough, sputum production and/or shortness of breath, especially if these symptoms are accompanied by fever, altered breath sounds and crackles or rales. 14 Ideally, the diagnosis of CAP should require a combination of clinical, laboratory and microbiological data prior to treatment. The minimal requirement for diagnosis careful history including comorbid conditions and physical examination is included in both of the recognized North American guidelines for the diagnosis and treatment of CAP. 5,14,17 In addition, both the Canadian guidelines 5 and the Infectious Diseases Society of America (IDSA) practice guidelines 14 recommend the use of chest radiography in persons whose symptoms and physical examination suggest the possibility of pneumonia. 19 3

18 Practically however, CAP may still be treated empirically in Canada without radiographic or laboratory confirmation. Therefore, it is incumbent on the diagnosing physician to clinically differentiate between upper and lower respiratory tract infections, bacterial versus viral pathogens, and non-infectious causes of the symptoms presented. More importantly, the ability to diagnose pneumonia is desirable given the increasing concern for the overuse of antibiotics. 20,21 In prospective cohort studies and clinical trials, usually less than 50% of responsible pathogens are defined or identified as causative. 8,18 The Pneumonia Patient Outcome Research Team s (PORT) study was observational in nature and not designed to examine the use of diagnostic tests. It was observed in this study, however, that 29.7% of outpatients with CAP had one or more microbiologic tests performed and only 5.7% of CAP outpatients had an assigned microbiologic cause. 22 In comparison, a microbiologic cause was identified in 29.6% in the inpatient population. The absence of a single test to identify all potential pathogens is a barrier to identifying the most responsible pathogen. A sputum Gram stain and culture may not show the presence of typical pathogens such as S.pneumoniae. In addition, Gram stain is not able to detect atypical bacteria like M.pneumoniae, C.pneumoniae and respiratory viruses. 14 These stains require a serum sample to search for antibodies and, in the case of Legionella pneumophilia, a urine sample is also required for its antigen. S.pneumoniae still accounts for a large portion of pneumonias. 23,24 However, atypical organisms are more commonly presumed causative pathogens than S. pneumoniae in those less than 65 years of age. 25 It is difficult to say with any precision which the most common pathogen is in Canada that causes CAP as measured by the severity of presenting symptoms because of differing patterns in the etiology for CAP throughout the developed world. In addition, no etiologic pathogen is found in 30% to 60% of patients. 14 The failure to identify an etiologic pathogen in a large number of CAP patients may necessitate the continued use of empiric antibiotic therapy. Echoing Bartlett and Mundy s 13 words then, CAP is a difficult disease to diagnose and treat because of multiple etiologic pathogens, the difference in severity of presenting individuals and the lack of tests for CAP with adequate sensitivity and specificity that would definitively indicate the right treatment. 1.4 Treatment of Community-acquired Pneumonia Antibiotics used to treat CAP both inside and outside of hospital have included: penicillins, cephalosporins, macrolides, tetracyclines and trimethoprim-sulfamethoxazole (TMP/SMX). Appropriate antibiotic treatment is used for a specific pathogen when it is known. However, there is little likelihood of accurately determining the causative pathogen in an individual presenting CAP-like symptoms in the community setting. Clinicians are left with trying to categorize individuals by age, comorbidity and severity of symptoms in order to make an appropriate empiric decision about treatment. 4

19 Individuals working on the Agency for Health Care Policy and Research s Pneumonia PORT developed a prediction rule for prognosis. This rule would help clinicians to accurately identify individuals with CAP who have a low, medium and high risk of dying within 30 days of presentation with CAP. 26 The use of such a prediction rule can also help with treatment choices and it is now commonly recommended that an individual s risk be taken into consideration when choosing a treatment path for CAP. 26 Both of the currently used North American treatment guidelines use an approximation of severity (age, comorbidity and place of treatment) to assist clinicians with their empiric decisions. 5,14 The Infectious Diseases Society of America and the Canadian Infectious Diseases/Thoracic Societies each recently have published (2000), a comprehensive document for the empirical selection of antibiotics for persons with CAP. 5,14 A compilation of the two guidelines is outlined in Table 1. There is rising concern about penicillin-resistant pneumoccocci, but its clinical implications are still uncertain. 1 The Canadian Community-acquired Pneumonia Working Group states in its most recent guidelines that, the current laboratory definitions of penicillin resistance for non-csf isolates of S. pneumoniae may not be clinically relevant, and this issue is being re-examined by the National Committee for Clinical Laboratory Standards. 5 In their opinion, there is substantial evidence that penicillins remain effective in the treatment of nonmeningitic pneumococcal infections caused by strains for which MICs are 4 mg/l. 5 Therefore, with the development of new macrolides like clarithromycin and azithromycin and respiratory fluoroquinolones, a reassessment of the role of these new antibiotics in the initial management of CAP is needed. The new or respiratory fluoroquinolones include gatifloxacin, gemifloxacin, grepafloxacin, levofloxacin, moxifloxacin, sparfloxacin and trovafloxacin. When this study was done, only levofloxacin was approved for use in Canada; moxifloxacin (October 2000) and gatifloxacin (January 2001) have been subsequently approved. Rare cases of cardiotoxicity resulted in the removal of grepafloxacin from the Canadian market in October 1999, c while a limited number of hepatotoxicity events lead Pfizer Canada and Health Canada to change the labeling of trovafloxacin in May d Trovafloxacin use has been limited to instances where treatment is initiated in a hospital. e In October of this year, Pfizer Canada announced it will withdraw trovafloxacin from the Canadian market, due to its limited use and the presence of alternative antibiotics on the market. f However, at the time this review was conducted, trovafloxacin was still available in Canada. The respiratory or new fluoroquinolones have a broader spectrum than the new macrolides, and their usefulness in CAP is the subject of this report. A discussion of pharmacokinetic/dynamic properties across the various fluoroquinolones provides some justification for deriving a class effect and combining the safety and efficacy data for this clinical and economic review, although this cannot be confirmed at this time given the relative lack of clinical efficacy/safety trials comparing the individual new fluoroquinolones. c Dear Pharmacist Letter, GlaxoWellcome, Mississauga, ON, October 29, d Dear Pharmacist Letter, P fizer Canada, Kirkland, QC, May 17, e CPS: Compendium of pharmaceuticals and specialities. 36 th ed. Ottawa: Canadian Pharmacists Association;2001. f Dear Pharmacist Letter, P fizer Canada, Kirkland, QC, October

20 1.4.1 Pharmacokinetic/dynamic review of new fluoroquinolones A traditional review was undertaken to determine whether the new fluoroquinolones are likely of equal efficacy and safety from a pharmacokinetic/pharmacodynamic perspective. This review relies heavily on a paper authored by two of the co-investigators (GGZ and LV). 27 Pharmacokinetics is the study of what the body does to the drug. A review outlining the differences and similarities of the new fluoroquinolones around absorption, distribution, metabolism, and elimination parameters is helpful to understand, clinically, the premise for equivalency. Pharmacodynamics is what the drug does to the pathogen. This describes the (1) structure/ antibiotic activity of the various analogs of the basic quinolone molecular structure, (2) mechanism of action, (3) mechanism of resistance, and (4) in vitro activity. a) Pharmacokinetic review of new fluoroquinolones All fluoroquinolones have high oral bioavailability. 27 The time to peak concentration is approximately 1 to 2 hours with the exception of sparfloxacin (time to peak concentration is approximately 4.5 hours). Administration with food results in slower absorption and decreased peak levels, however, this reduction is thought to be non-significant. 27 Significant data on the tissue distribution of the new fluoroquinolones are available only for grepafloxacin, levofloxacin, sparfloxacin and trovafloxacin. Grepafloxacin and sparfloxacin exhibit the best penetration into respiratory tissues and fluids. All the newer fluoroquinolones have longer half-lives than ciprofloxacin and may be dosed once daily. Excretion for gatifloxacin and levofloxacin is predominantly renal, while excretion for grepafloxacin, moxifloxacin, sparfloxacin and trovafloxacin is predominately non-renal. The major and limiting adverse drug reactions relevant to the new fluoroquinolones class include: cardiovascular, hepatotoxic and phototoxic reactions. Prolongation of the QTc interval has been observed with grepafloxacin and sparfloxacin putting individuals at risk for a rare, but potentially fatal, ventricular arrhythmia known as torsades de pointes. The potential for this effect is greater in those persons with pre-existing cardiovascular abnormalities and those taking concomitant medications (e.g., class I and III antiarrhythmics, terfenadine, astemizole, cisapride, tricyclic antidepressants) known to prolong the QTc interval. Grepafloxacin (Raxar ) was removed from the Canadian market in October 1999 following reports of torsades de pointes associated with its use. Sparfloxacin (Zagam ) remains on the market (in the United States) but its labeling was changed on October 19, 1999 to reflect caution for its use in persons with preexisting cardiac abnormalities or taking concomitant medications known to prolong the QTc interval. Hepatoxicity has been largely reported with trovafloxacin (Trovan ) and the US Food and Drug Administration (FDA) and Health Canada consequently issued a public health advisory to physicians in June The FDA and Pfizer agreed to limit the use of trovafloxacin and its IV prodrug alatrofloxacin to persons who receive initial treatment in a hospital or long-term nursing facility. Phototoxic reactions are possible with all fluoroquinolones and appear to result from an interaction between ultraviolet A (UVA) light and the fluoroquinolone molecule. This interaction produces photodegradation products and/or reactive oxygen species that may cause tissue damage. 28 Sparfloxacin exhibits the highest incidence of phototoxicity. 27 Phototoxicity 6

21 due to sparfloxacin has resulted in the need for hospitalization in approximately 10-15% of cases. 28 Levofloxacin (Levaquin ) and trovafloxacin have relatively low phototoxic potential. Limited information is available regarding new fluoroquinolones and drug interactions. Grepafloxacin is known to inhibit cytochrome P450 (1A2) decreasing the clearance of theophylline Ciprofloxacin is the only fluoroquinolone reported to reduce the metabolism of warfarin and phenytoin by inhibiting cytochrome P450; similar decreases in warfarin and phenytoin metabolism have not been seen with the new fluoroquinolones. 27,29-31 b) Pharmacodynamic review of new fluoroquinolones The new fluoroquinolones were developed in response to increasing antimicrobial resistance to common pathogens, especially to S.pneumoniae. The use of specific structural modifications to improve microbiologic spectrum and potency was recognized in the mid-1980s. 32 However, the structural modification has also resulted in a revised adverse events profile for the class of new fluoroquinolones. 33 The differences in safety profiles among the newer fluoroquinolones need to be understood and weighed when interpreting efficacy findings. Fluoroquinolones are broad-spectrum antibiotics exhibiting antibacterial activity against a wide variety of Gram positive and Gram negative bacteria. Structural modifications in the newer fluoroquinolones expanded their antibacterial spectrum for enhanced coverage of Gram positive organisms (e.g., S.pneumoniae). 33 All fluoroquinolones share the same basic quinolone structure, with the exception of trovafloxacin which possesses a naphthyridine structure. Fluoroquinolones are bactericidal; target sites are bacterial DNA gyrase and topoisomerase IV. Through interference with bacterial DNA gyrase, fluoroquinolones prevent the supercoiling of bacterial DNA. 34 Interference with topoisomerase IV inhibits decatenation of DNA, thus inhibiting replication Fluoroquinolones may act at both target sites, however, the primary site may be dependent on bacteria type In addition, different fluoroquinolones may selectively target one enzyme over another. 36 The mechanisms of resistance to fluoroquinolones include mutations of the target sites and reduced accumulation of fluoroquinolones within the bacteria. 36,37 Decreased accumulation may be accomplished through a decrease in uptake into bacteria or through an increase in the efficiency of the efflux mechanism. 38 The mechanism of resistance may be dependent on bacteria type and/or the specific fluoroquinolone as different mechanisms of resistance do not affect all of the fluoroquinolones equally. 36 Barry et al. have found that pneumoccocci resistant to penicillin, cefotaxime, erythromycin, clindamycin, trimethoprim/sulfa-methoxazole, tetracycline and chloramphenicol were still susceptible to sparfloxacin and trovafloxacin, suggesting that new fluoroquinolones may offer therapeutic solutions in the treatment of bacteria resistant to many drugs. 27 New fluoroquinolones offer an expanded antibacterial spectrum over ciprofloxacin. An examination of the in vitro activity of the new fluoroquinolones reveals improved coverage of Gram-positive organisms (e.g., S.pneumoniae). 27 Trovafloxacin exhibits the greatest improvement (MIC only) over ciprofloxacin, while levofloxacin is the least improved. 27 Specifically, for penicillin-sensitive and penicillin-resistant S.pneumoniae, moxifloxacin and 7

22 trovafloxacin have the greatest activity (MIC only), followed by the other new fluoroquinolones. 27 Gains in Gram-negative coverage compared to ciprofloxacin are limited. Ciprofloxacin s coverage of anaerobes is marginal but many of the new fluoroquinolones, with the exception of grepafloxacin and levofloxacin, have significantly improved anaerobic coverage. Similarly, coverage of atypical organisms is marginal with ciprofloxacin. All the new fluoroquinolones offer improved coverage over ciprofloxacin of atypical organisms (e.g., C.pneumoniae) Outcomes of Community-acquired Pneumonia In addition to the outcomes of antibiotic treatment for CAP that are usually reported (clinical success g, clinical failure h and mortality), two other morbidity-related outcomes have been reported by the Pneumonia PORT: (1) the rate of return to usual activity and /or work for inpatients and outpatients treated for CAP and (2) the rate of persisting pneumonia-related symptoms 30 days post-start of antibiotic treatment. 22 A description of these other morbidityrelated outcomes are important and, in the words of the Pneumonia PORT investigators, allow physicians to inform their patients about the likelihood of potential complications, (the) natural history of symptom resolution, and expected time to return to usual activities. 22 Unfortunately, published reports of clinical trials (examining efficacy and safety of treatments for the purposes of registration) for antibiotics do not include a description of these outcomes by treatment arm. This makes a clinical and economic review of a new class of CAP treatment, like the respiratory fluoroquinolones, difficult to describe in other than the simplest terms (% clinical success and mortality). Little was known about these important CAP outcomes before the Pneumonia (PORT) study formed under the aegis of the U.S. National Institute of Health s Agency for Health Care Policy and Research (grant number R01 HS06468). The processes and outcomes of care for CAP are described by the Pneumonia PORT project for over 900 outpatients and 1300 inpatients. 22 Although an understanding of the medical outcomes of CAP was the main objective of the Pneumonia PORT s investigation; the study extended beyond simply short-term mortality and the usual clinical trial CAP treatment outcomes of clinical success or clinical failure. For example, the study reports that three deaths (0.3%) attributable to pneumonia occurred in the Pneumonia PORT s outpatient population. At 30 days, 95.6% of employed outpatients had returned to usual activities, yet 76.0% of all outpatients still had one or more persisting pneumonia-related symptoms. 22 g Clinical success: Cure is a resolution of signs and symptoms of pneumonia, improved or resolved on chest radiography and no need for further antibiotic therapy; improvem ent is a resolution of fever, improvement in other signs and symptoms, improved or resolution on chest radiographic findings, and no further need for antibiotic therapy. h Clinical failure: Failure is a lack of resolution of chest radiographic findings, signs and symptoms, and/or requirement for additional antibiotics for inadequate response; relapse is a clinical cure/improvement at treatment end with failure at a subsequent visit, or clinical cure or improvement at treatment end but receipt of additional antibiotic therapy for the CAP before study end. 8

23 Regardless, if one extrapolates from what we know of CAP-related death and morbidity in the United States, then we might assume that CAP results in 6.4 million days of restricted activity, 3.9 million days of bed confinement and 1 million days of work loss annually in Canada Economic evaluations in community-acquired pneumonia Economic evaluations of treatments for specific conditions aid in formulating the best decisions within the confines of limited monetary resources. However, few studies have been published which focus on the empirical outpatient treatment of CAP and its costs and consequences Several burden of illness or cost-only analyses of CAP exist, but all point to the need to fully evaluate the costs and consequences of the various empirical treatment alternatives Ideally, the costs of CAP treatment and the enumeration of its consequences (clinical success, adverse events, morbidity days, etc.) should be specifically identified through the prospective, controlled, randomized and blinded trials designed to describe efficacy and safety for the purposes of licensing or registration. Traditionally, however, clinical trials intended for the evaluation of efficacy and safety alone, have been used as the source of information for an economic evaluation of CAP. There are rapidly changing antibiotic resistance patterns developing throughout the world. Therefore, using both the historical results from efficacy and safety trials for consequences as well as modelled data for costs, could risk false conclusions on cost-effectiveness of new treatment modalities for CAP. A CCOHTA-sponsored analysis (1997) used retrospective claims data and estimated the efficacy of three macrolides by using evaluable clinical success rates from a synthesis of published papers. 41 The study found erythromycin to dominate the newer macrolides (clarithromycin and azithromycin) in the treatment of CAP. However, the economic impact of hospitalization or referral to specialists associated with treatment failure was not considered in the analysis. This type of information is usually only available from a prospective study design that was not available for this analysis. In a study by Backhouse et al., a decision analytic model was developed to compare three common antibiotics (amoxicillin, cefaclor and amoxicillin-clavulanate) for the community-based treatment of CAP. 43 This analysis concluded that the main driver of the economic model was antibiotic clinical efficacy rates and that even small differences in efficacy could have a significant impact on the rank order of the various agents for cost-effectiveness. As well, the health care costs associated with treatment failure were more important than the drug price alone. Again, however, historical estimates from the clinical trials literature were used to itemize costs and consequences of treating CAP, making the general use of its results to specific geographic areas subject to question. Several authors have pointed to potential methodologic deficiencies in arriving at an economic evaluation that is believable given the present lack of prospectively collected economic data. Specifically, these deficiencies may be preventing a true look at the burden of CAP: A difference in clinical efficacy rates by patient type, by region and by place of treatment (hospital or community). 47 9

24 Consideration of only the immediate costs of alternate treatments without measurement of downstream costs and consequences associated with treatment failure (e.g., measurement of days of lost productivity or morbidity-days). 48 An inclusion of a no treatment arm in the cost-effectiveness of alternatives (this is seen to be unethical treatment in clinical trials; however, economic studies of CAP may help to identify costs and consequences of no treatment). 47 Lack of intent-to-treat analyses: few clinical trials report clinical success and failure based on the intent-to-treat population; failing to do so, however, may result in the inclusion of biased rates of clinical efficacy for calculation of the cost-effectiveness using clinical efficacy rates. 49 Use of clinical trial or observational data from retrospective studies; this evidence will continue to have limited value because of the changes to antibiotic susceptibility over time. 49 Lack of a reporting of morbidity-related outcomes such as days of restricted activity, days of bed confinement, days of work lost, rate of return to usual activities or work and rate of persisting pneumonia-related symptoms 30 days post-start of antibiotic treatment. 10

25 2. METHODOLOGY 2.1 Clinical Review Inclusion criteria The following criteria were used when considering studies for this review: Type of participants: Adults with a diagnosis of community-acquired pneumonia (CAP) treated in or out of hospital. Types of interventions: At least one group treated with one of six fluoroquinolones (gatifloxacin, grepafloxacin, levofloxacin, moxifloxacin, sparfloxacin, trovafloxacin) and compared to one or more antibiotic comparative treatments. The duration of treatment had to be equal to or greater than seven days. Types of outcome measures: The primary outcome measure of treatment for CAP was clinical success (cure or improvement). Secondly, measures were further stratified, where possible, with respect to the following outcomes: (a) the specific antibiotic (fluoroquinolone or other), (b) presumed causative organism, and (c) patient type (e.g., elderly, comorbid state). Types of studies: Randomized controlled trials (RCTs) [Results from RCTs (or meta-analysis of such trials) of adequate size to ensure a low risk of incorporating false-positive or false-negative results are considered to be Level I evidence]. A summary of inclusion and exclusion criteria can be found in Table Literature search A comprehensive search of electronic bibliographic databases (MEDLINE and EMBASE ) was conducted for articles on the treatment of community-acquired pneumonia published between January 1993 and December The search strategy used can be found in Appendix I. The major subject heading (MeSH) search terms used and exploded included: Pneumonia,-Bacterial Bronchopneumonia Pneumonia,-Viral Community-Acquired Infections Pleuropneumonia Bronchitis Influenza Common-Cold Legionellosis Ciprofloxacin Ofloxacin Anti-Infective-Agents,- Quinolones Anti-Infective-Agents,- Fluoroquinolone Erythromycin Amoxicillin Amoxicillin-Potassium- Clavulanate-Combination Cefaclor Cefuroxime Cephalexin Tetracycline Penicillin-V Doxycycline Several non-mesh headings were added to the search strategy including lower-respiratory infection, respiratory tract infection and individual antibiotic names (e.g., cefuroxime axetil, TMP-SMX, levofloxacin, moxifloxacin, etc.) Articles from the intersection of these topics in 11