Economic aspects of antimicrobial therapy of acute exacerbations of COPD

Respiratory Medicine (2007) 101, 15 26 REVIEW Economic aspects of antimicrobial therapy of acute of COPD Steven Simoens a,, Marc Decramer b, Gert Laekeman a a Research Centre for Pharmaceutical Care and Pharmaco-economics, Faculty of Pharmaceutical Sciences, K.U. Leuven, Onderwijs en Navorsing 2, Herestraat 49, P.O. Box 521, 3000 Leuven, Belgium b Respiratory Division, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium Received 23 December 2005; accepted 23 March 2006 KEYWORDS Chronic obstructive pulmonary disease; Exacerbations; Antibiotics; Pharmacoeconomics; Cost-of-illness; Economic evaluation Summary This article synthesizes and appraises the methodological quality of the international literature on costs of chronic obstructive pulmonary disease (COPD), economic value of antimicrobial therapy of, and factors affecting the economic value of antibiotics. Included studies had carried out a cost-of-illness, cost effectiveness, cost utility or cost benefit analysis. Exacerbation costs varied between countries and increased with the severity of the exacerbation. Hospitalization costs accounted for more than 45% of healthcare costs of. Drug costs made up 6 21% of healthcare costs in the majority of studies. Cost estimates were biased due to difficulties involved in diagnosing and treating, and because the patient sample studied may not be representative of the population of patients suffering from. Although this needs to be corroborated by future research, evidence is emerging that secondgeneration antibiotics such as fluoroquinolones may have a favourable economic profile as compared with first-generation antibiotics. The higher acquisition costs of fluoroquinolones appear to be balanced by less treatment failure, more time between exacerbation episodes, and lower hospitalization costs. There is a need for prospective economic evaluations alongside clinical trials with a sufficient number of patients and length of follow-up period. The economic value of antibiotics is influenced by difficulties involved in diagnosing the condition, effectiveness, resistance, patient compliance with treatment, and treatment failure associated with antibiotics. The small number of economic evaluations and their methodological limitations precludes the recommendation of a specific antibiotic for use in the management of COPD on economic grounds. & 2006 Elsevier Ltd. All rights reserved. Corresponding author. Tel.: +32 16 323465; fax: +32 16 323468. E-mail address: steven.simoens@pharm.kuleuven.be (S. Simoens). 0954-6111/$ - see front matter & 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.rmed.2006.03.030

16 S. Simoens et al. Contents Introduction.................................................................. 16 Methods.................................................................... 17 Search strategy............................................................. 17 Economic techniques......................................................... 17 Data analysis.............................................................. 17 Assessment of methodological quality............................................... 17 Results..................................................................... 17 Cost-of-illness studies of COPD.......................................... 17 Quality of cost-of-illness studies of COPD................................... 19 Economic evaluations of antimicrobial therapy of COPD.......................... 19 Quality of economic evaluations of antimicrobial therapy of COPD................... 22 Factors affecting the economic value of antimicrobial therapy of COPD.................. 23 Discussion................................................................... 24 Acknowledgements............................................................. 24 References................................................................... 24 Introduction Chronic obstructive pulmonary disease (COPD) is a disease characterized by chronic airflow limitation which is not fully reversible, progressive and linked with an abnormal inflammatory response to noxious stimuli. 1 The decline in lung function of COPD patients tends to be accompanied by intermittent acute. Although there is debate about the exact definition of a COPD exacerbation, a well-accepted definition states that a COPD exacerbation requires the presence of one or more of the following clinical symptoms: increase in sputum purulence, increase in sputum volume, and worsening of dyspnea. 2 COPD are associated with significant morbidity, impaired quality of life, mortality, and are a primary driver of hospital admissions and healthcare costs. Exacerbations worsen the natural history of COPD and lead to poorer functional status and quality of life. 3 An inpatient mortality rate of COPD of 10%, rising to 40% in 1 year, has been reported. 4 The literature indicates that 3 16% of give rise to a hospital admission. 3,5 A British study found that 10% of hospital admissions in the UK originate from COPD. 6 Half of hospitalized patients need to be re-admitted to hospital at least once in the next 6 months. 7 Exacerbations make up around 70% of direct healthcare costs of COPD. 8 COPD are generally triggered by respiratory infections, but also by environmental exposures (e.g. air pollutants, dust) and co-morbidities (e.g. heart failure, pulmonary embolism, pneumothorax and non-pulmonary infections). Forty to fifty per cent of respiratory infections are caused by bacteria, with the most common bacterial isolates being Haemophilus influenzae, Streptococcus pneumoniae, and Moraxella catarrhalis. Other causes of respiratory infections are viruses (30%) and atypical bacteria (5 10%). 9 Although there is some controversy over the role of bacterial infections, literature reviews and meta-analyses have found that antibiotics improve clinical outcomes in COPD and that patients with more severe benefit more from antibiotics. 10 12 There is some debate about the choice of antibiotic in the management of. Guidelines on the diagnosis, management and prevention of COPD issued by the World Health Organization Global Initiative for Chronic Obstructive Lung Disease (GOLD) recommend that the choice of antibiotic reflects local patterns of microbial resistance to H. influenzae, S. pneumoniae, and M. catarrhalis. 1 Increased resistance to traditional first-generation antibiotics (aminopenicillins, macrolides, tetracyclines) needs to be weighed against higher acquisition costs and relative effectiveness of cephalosporins, secondgeneration macrolides and fluoroquinolones. Additionally, in selecting an antibiotic, a physician needs to consider dose form, ease of administration, frequency of dosing, spectrum of activity, duration of treatment, adverse effect profile and potential drug interactions. The aim of this article is to synthesize the international literature on economic aspects of antimicrobial therapy of acute of COPD. This review focuses on three specific aspects: first, the level and distribution of costs associated with ; second, the economic value of antibiotics in the management of COPD ; and third, factors influencing the economic value of antibiotics. Methodological issues surrounding the evaluation of costs and

Antimicrobial use in acute of COPD 17 economic value of antibiotics are also explored. The relevance of the findings is discussed and gaps in the evidence base are identified. Methods Search strategy Studies were identified by searching the following electronic databases up to December 2005: PubMed, EMBASE, Bath Information and Data Services, National Health Service Economic Evaluation Database, Cochrane Library, EconLit, and Social Science and Citation Index. Search terms included COPD, chronic bronchitis (CB), emphysema, acute exacerbation, costs, costs of treatment, cost-of-illness, pharmaco-economics, economic evaluation, cost effectiveness, cost minimization, cost utility, cost benefit alone and in combination with each other. Additionally, the bibliography of included studies was checked for other relevant studies. Economic techniques Information about the costs of COPD was derived from cost-of-illness analyses. A costof-illness analysis was defined as a study quantifying the costs involved in treating an exacerbation. The economic value of antimicrobial therapy was derived from cost effectiveness, cost utility and cost benefit analyses. Cost effectiveness analysis denotes an economic evaluation where a single consequence is quantified in a natural unit, such as the number of successfully treated patients, the number of exacerbation-symptom days averted or the number of patients without recurrence of exacerbation. The incremental cost effectiveness ratio is calculated as the difference in costs between alternatives divided by the difference in the consequence. This technique is of less value when multiple consequences need to be considered. A study that measures consequences by specific health-related quality of life measures, such as quality-adjusted life years (QALYs), is referred to as a cost utility analysis. The QALY takes into account the quantity and quality of life. Quality of life associated with a health state is measured on a scale of 0 (reflecting death) to 1 (reflecting perfect health). Quality of life data are then combined with estimates of the time period for which the health benefits last to generate QALYs. Cost benefit analysis refers to an economic evaluation where consequences are valued in monetary terms. A monetary value can be assigned to health benefits by means of, for instance, the willingness-to-pay technique. As a result, benefits can be compared with costs of antimicrobial therapy and the net worth (benefits minus costs) of each of the treatment alternatives can be estimated. The results of a cost benefit analysis may be stated in the form of the net benefit or loss of one alternative over another or in the form of a ratio of benefits to costs. Data analysis To compare costs between studies, costs were actualized to 2001 values using a rate of inflation based on the evolution of the Consumer Price Index. Costs were converted using US dollar exchange rates. Assessment of methodological quality A qualitative appraisal was carried out of the methodological quality of cost-of-illness studies by investigating study population, data sources, methods of data collection, scope of included costs and time horizon. The quality of economic evaluations was assessed by considering the perspective, scope and measurement of costs and consequences, sources of data on costs and consequences, and application of an incremental analysis. 13 Results Cost-of-illness studies of COPD The literature search generated 52 articles. Articles were excluded if healthcare resource utilization was not converted into costs or if costs of were not distinguished from costs of underlying COPD or co-morbidities. Nine cost-ofillness studies of COPD were included in the review. 5,14 21 Direct healthcare costs of drugs, contacts with healthcare professionals and hospitalization vary substantially from study to study, with estimates ranging from $123 to $666 per exacerbation (see Table 1). 5,15,18,20 This variation in cost estimates originates, amongst other things, from differences between countries in healthcare resource utilization, prices, clinical practice, the organization and funding of healthcare systems. One study explored

18 S. Simoens et al. Table 1 Direct healthcare costs per COPD exacerbation. Country Costs per exacerbation ($) Reference Spain 123 5 Germany 146 15 Sweden 334 18 Netherlands/Belgium 666 20 Argentina, Brazil, Colombia, Ecuador, Mexico, Peru, Venezuela 98 329 19 Table 2 Distribution of direct healthcare costs of COPD. Country Hospitalisation costs (%) Drug costs (%) Other outpatient costs (%) Reference USA 97 0.4 2 14 Netherlands/Belgium 79 8 14 20 Italy 77 10 14 21 Sweden 69 6 25 18 Spain 67 21 11 5 UK 44 11 45 16 Germany 44 34 22 15 the sensitivity of costs to differences in prices between countries. 19 Costs of treating an exacerbation were compared between Argentina, Brazil, Colombia, Ecuador, Mexico, Peru and Venezuela. The authors derived data on healthcare resource utilization from a Spanish study. 5 These data were then combined with price information for each country to calculate costs. The analysis revealed that costs per exacerbation ranged from $98 in Colombia to $329 in Argentina. Direct healthcare costs increase with the severity of the exacerbation. A Swedish study defined severity on the basis of healthcare resource use. 18 Costs per exacerbation amounted to $13 for a mild exacerbation; $226 for a moderate exacerbation; and $2330 for a severe exacerbation. The authors of a Dutch/Belgian study defined exacerbation severity in terms of a physician assessment of the impact of the exacerbation on ability to perform usual activities. 20 Costs per exacerbation amounted to $81 for a mild exacerbation; $532 for a moderate exacerbation; and $3719 for a severe exacerbation. The distribution of direct healthcare costs of is presented in Table 2. The principal cost driver of direct healthcare costs is inpatient costs, accounting for more than 45% of direct healthcare costs. As hospitalization is generally indicative of treatment failure, these estimates highlight the potential savings that can be attained from preventing treatment failure. One study focused specifically on quantifying costs of treatment failure. Costs of treatment failure arising from additional outpatient visits, emergency department visits and hospitalizations accounted for 29 63% of direct healthcare costs, depending on the country. 19 Table 2 also shows that the second most important cost driver of direct healthcare costs tends to be other outpatient costs arising from contacts with healthcare professionals and diagnostic tests (2 45%). Drug costs cover up to 35% of direct healthcare costs, with the majority of studies having estimates between 6% and 21%. Few studies have quantified direct non-healthcare costs of transportation to healthcare professionals associated with treatment of. A Swedish and a Dutch/Belgian study found similar estimates indicating that transportation costs account for less than 3% of direct costs. 18,20 No study explored direct non-healthcare costs associated with home adaptations and educational requirements. The few available estimates of indirect costs of productivity loss associated with vary, depending on the scope of costs included. An Italian cost-of-illness study investigated indirect costs arising from work absenteeism of patients during treatment of. 21 The authors found that indirect costs made up 3% of annual costs per patient. A German study not only considered work absenteeism, but also took into account productivity loss arising from premature death of patients and productivity loss of family/

Antimicrobial use in acute of COPD 19 friends who care for patients. 15 Indirect costs accounted for 47% of costs per exacerbation. Quality of cost-of-illness studies of COPD Some cost-of-illness studies enrolled national samples that are representative of the population, 14,16,17 whereas others analyzed a specific group of patients. 5,15,18 21 Differences in, for example, age and disease severity of the study population influence cost estimates. With respect to age of patients, a US study observed that costs of patients aged 65 years and over were around three times higher than costs of patients aged under 65 years. 14 With respect to disease severity, two studies found that costs per exacerbation were 7 10 times higher for a severe exacerbation than for a moderate exacerbation. 18,20 A combination of survey and claims data were used in studies to obtain information about health resource utilization associated with. The reliability of survey data is hindered by patients ability to recall health resource utilization. Claims data may suffer from missing data and incorrect diagnostic coding of claims. Examining the impact of the data source on cost estimates, a US study found that hospitalization costs per discharge amounted to $6336; $6677 or $7038, depending on the data source used. 17 Studies enrolling a specific group of patients tend to be carried out prospectively using survey data. 5,15,18 21 This type of analysis can be considered to be more reliable than studies focusing on a national sample, which are generally based on a retrospective analysis of clinical databases. 14,16,17 All but one cost-of-illness study identified patients suffering from, but did not have a control group of patients without. Such case series analyses that focus on identified patients only may be misleading in the case of COPD, where diagnosis is complex and attribution of healthcare resource utilization to the disease is difficult. A case control approach seems better suited in that it is more inclusive and allows identification of additional healthcare resource use related to. This approach was taken by one study which compares patients suffering from CB with equivalent healthy patients. 16 The scope of included costs was generally restricted to direct healthcare costs associated with. This refers to costs of drugs, contacts with healthcare professionals and hospitalization. Few studies have considered direct nonhealthcare costs associated with transportation to the healthcare professional, 18,20 or indirect costs arising from lost productivity of patients. 15,21 The omission of indirect costs is of minor concern given that the population of patients that experience COPD tends to be elderly. Moreover, in this population of elderly and frail patients, it is not straightforward to distinguish lost productivity of caregivers related to COPD from lost productivity related to general illness. Nevertheless, in order to gain an insight into the costs of from a societal perspective, future studies need to collect data on all relevant costs generated by patients suffering from. Cost-of-illness studies measured costs of during a certain time period. In studies that reported on patients during 1 month, the duration of the period may have been too short to capture all relevant costs. 5,19 Other studies had a longer follow-up period of up to 6 months, but this period was still too short to account for seasonal variations in the frequency of. 18,21 Such studies thus produced biased estimates of the burden of illness associated with COPD. Economic evaluations of antimicrobial therapy of COPD A total of 57 articles were identified by the literature search. Articles were excluded from the review for the following reasons: treatment of COPD rather than, treatment not involving antibiotics, analysis of a single intervention without a comparator, absence of data on both costs and consequences. The economic literature evaluating antimicrobial therapy of COPD consisted of seven cost effectiveness analyses. 22 28 Two studies examined both cost effectiveness and cost utility. 29,30 No cost - benefit analyses were identified. Characteristics of these studies are reported in Table 3. Three studies set in different countries contrasted the cost effectiveness of aminopenicillins, cephalosporins, macrolides and fluoroquinolones in terms of the cost per successfully treated patient. 22 24 These studies employed simple decision analytic models populated with effectiveness data derived from the literature and expert opinion. Two analyses calculated average cost effectiveness ratios. An Italian study found comparable costs per successfully treated COPD patient for first-line therapy with co-amoxiclav and rufloxacin, which tended to be more cost effective than ciprofloxacin and clarithromycin. 23 Irrespective of severity

20 S. Simoens et al. Table 3 Economic evaluations of antimicrobial therapy of COPD. Country Sample Comparators Design Costs and consequences Ref Cost effectiveness analyses UK Patients with CB Italy Patients with COPD Germany Patients with CB USA 60 patients with CB USA 121 patients with CB First- and second-line therapy of amoxicillin, co-amoxiclav, ciprofloxacin, cefaclor First-line therapy with ciprofloxacin, rufloxacin, clarithromycin, co-amoxiclav Macrolides (roxithromycin, clarithromycin), fluoroquinolones (ciprofloxacin), penicillins (amoxicillin, co-amoxiclav), cephalosporins (second and third generation) First-line agents (amoxicillin, cotrimoxazole, tetracyclines, erythromycin); second-line agents (cephradine, cefuroxime, cefaclor, cefprozil); third-line agents (co-amoxiclav, azithromycin, ciprofloxacin) First-line agents with partial coverage (amoxicillin, cephalexin, erythromycin); firstline agents with full coverage (doxycycline); second-line agents (co-amoxiclav, azithromycin, cefuroxime, clarithromycin, levofloxacin, ofloxacin) Decision analytic model Decision analytic model Decision analytic model Evaluation alongside retrospective cohort study Evaluation alongside retrospective cohort study Compared with two courses of amoxicillin, cost per successfully treated patient was $137 with amoxicillin and co-amoxiclav; $705 with two courses of ciprofloxacin; $328 with amoxicillin and ciprofloxacin. First- and second-line therapy with co-amoxiclav dominated two courses of amoxicillin. Average cost per successfully treated patient was $160 with rufloxacin; $156 with co-amoxiclav; $218 with clarithromycin; $204 with ciprofloxacin. Average cost per successfully treated patient was $184 with macrolides; $168 with fluoroquinolones; $210 with penicillins; $181 with cephalosporins in patients with less than 4 AECBs per year. Average cost per successfully treated patient was $511 with macrolides; $323 with fluoroquinolones; $409 with cephalosporins in patients with four or more AECB per year. Third-line agents reduced the failure rate and need for hospitalization, prolonged the time between AECB episodes. Average drug costs of therapy of AECB were $11 with first-line; $27 with second-line; $51 with thirdline agents. Average costs of AECB episode were $1075 with first-line; $643 with second-line; $619 with thirdline agents. There were no differences between the three antibiotic groups in terms of failure rate, hospitalization rate, time until next AECB. Average drug costs of therapy of AECB were $3 with first-line agents with partial coverage; $3 with first-line agents with full coverage; $39 with second-line agents. Average costs of AECB episode were $1073 with first-line agents with partial coverage; $239 with first-line agents with full coverage; $344 with second-line agents. 22 23 24 25 26

Antimicrobial use in acute of COPD 21 USA 438 patients with CB Spain 1,097 patients with CB and COPD Gemifloxacin, clarithromycin Evaluation alongside prospective RCT Moxifloxacin, co-amoxiclav, clarithromycin Evaluation alongside prospective cohort study Percentage of patients without AECB recurrence after 26 weeks was 73.8% with gemifloxacin and 63.8% with clarithromycin. Direct healthcare costs per patient were $316 with gemifloxacin and $450 with clarithromycin. Costs of medical care and lost productivity per patient were $1538 with gemifloxacin and $1,883 with clarithromycin. The probability that gemifloxacin dominates clarithromycin was 88% from perspective of third-party payer and 84% from societal perspective. Resolution of symptoms at 10 days occurred in 69.2% of patients with moxifloxacin; 65.7% of patients with coamoxiclav; 66.7% of patients with clarithromycin. Costs per exacerbation were $100 with co-amoxiclav; $127 with clarithromycin; $102 with moxifloxacin. Cost effectiveness and cost utility analyses Canada 222 patients with CB Ciprofloxacin versus standard antibiotic care (amoxicillin, cefuroxime, erythromycin, cefaclor and others) Canada 222 patients with CB Ciprofloxacin versus standard antibiotic care (amoxicillin, cefuroxime, erythromycin, cefaclor and others) Evaluation alongside prospective RCT Evaluation alongside prospective RCT The cost per AECB-symptom day averted was $160 with ciprofloxacin as compared with usual care. The cost per QALY gained was $14,199 with ciprofloxacin as compared with usual care. In patients with moderate or severe CB and at least four AECBs in the past year, ciprofloxacin dominated usual care. Cost per AECB-symptom day averted with ciprofloxacin was $160 from societal perspective and $232 from thirdparty payer perspective. Cost per QALY gained with ciprofloxacin was $14,199 from societal perspective and $20,644 from third-party payer perspective. Ciprofloxacin may be dominant in patients with more severe disease. Note: CB ¼ chronic bronchitis; COPD ¼ chronic obstructive pulmonary disease; RCT ¼ randomized controlled trial. 27 28 29 30

22 of disease, a German study noted that ciprofloxacin and cephalosporins were more cost effective than macrolides (roxithromycin, clarithromycin) and aminopenicillins (amoxicillin, co-amoxiclav) in treating CB. 24 A British study explored the incremental cost effectiveness of two courses of therapy with amoxicillin, co-amoxiclav, cefaclor or ciprofloxacin in patients with severe CB. 22 The findings showed that coamoxiclav as first-line and second-line therapy dominated two courses of amoxicillin. The incremental cost effectiveness of amoxicillin followed by co-amoxiclav, two courses of ciprofloxacin, and amoxicillin followed by ciprofloxacin was $137, $705 and $328 per successfully treated patient, respectively. A prospective analysis of 1097 Spanish patients compared treatment of COPD with moxifloxacin (39% of patients), co-amoxiclav (32%) or clarithromycin (29%) during 7 10 days. 28 The results indicated that costs of treatment failure accounted for 44% of costs per exacerbation. This percentage lies within the range of estimates observed in the cost-of-illness literature. 5,14 16,20,21 Moxifloxacin and co-amoxiclav had a similar cost per exacerbation, but resolution of symptoms at 10 days occurred in a higher percentage of patients treated with moxifloxacin. Costs and consequence data were not combined into cost effectiveness ratios. When contrasting the cost effectiveness of different generations of antibiotics, higher drug acquisition costs of newer antibiotics need to be balanced against the relative effectiveness of antibiotics, as this has implications for treatment failure, time between exacerbation episodes, hospitalization costs, and costs of lost productivity. Two economic evaluations reviewed medical records of small samples of US patients suffering from CB who were not randomized to antimicrobial therapy. Although the classification of the different generations of antibiotics varied between studies, one study found that newer antibiotics (fluoroquinolones, azithromycin) were associated with a lower failure rate and need for hospitalization, more time between exacerbation episodes, and a lower cost for treating. 25 The second study found no significant differences between groups of antimicrobial agents in terms of costs or effectiveness. 26 An economic evaluation examined treatment of CB with gemifloxacin, a fluoroquinolone, or with clarithromycin, a macrolide, in a sample of US patients. 27 Effectiveness was defined as the percentage of patients without recurrence of exacerbation requiring antimicrobial therapy after S. Simoens et al. 26 weeks. Costs were calculated from the perspective of the third-party payer (direct healthcare costs) and the societal perspective (direct healthcare costs and costs arising from lost productivity). Treatment with gemifloxacin was associated with lower direct healthcare costs, inferior costs of lost productivity, and a higher percentage of patients without exacerbation recurrence. The authors accounted for uncertainty around the point estimates of incremental cost effectiveness ratios and calculated that the probability of gemifloxacin dominating clarithromycin was 88% from the perspective of the third-party payer and 84% from the societal perspective. Investigating cost effectiveness as well as cost utility, two studies performed an economic evaluation alongside the same randomized controlled trial of Canadian patients that underwent therapy of CB with ciprofloxacin or standard antibiotic care (amoxicillin, cefuroxime, erythromycin, cefaclor and others). 29,30 Incremental costs per exacerbation-symptom day averted with ciprofloxacin as compared with standard care were $232 from the perspective of the third-party payer and $160 from the societal perspective. Incremental costs per QALY gained with ciprofloxacin were $20,644 and $14,199; respectively. Quality of economic evaluations of antimicrobial therapy of COPD The perspective taken in economic evaluations was that of the third-party payer 24 26,28 or the National Health Service. 22,23 These studies considered direct healthcare costs associated with treating. On the one hand, this is acceptable as interest in managing tends to focus on the type of antibiotics, intensity and length of treatment in outpatient and inpatient settings. On the other hand, such a perspective is too restrictive as treatment has wider implications on employment of patients and family and friends who give care to patients. Only three out of the eight economic evaluations adopted a societal perspective. 27,29,30 There was no uniformity in how consequences were quantified. Some studies used intermediate process measures (e.g. treatment compliance, failure rate after first-line treatment), 25,26,28 whilst others opted for final measures (e.g. the number of exacerbation-symptom days averted or the number of patients without recurrence of exacerbation). 22 24,27,29,30 Economic evaluations were carried out alongside randomized controlled trials 27,29,30 or observational

Antimicrobial use in acute of COPD 23 These studies were designed to reflect real-world practices in managing patients suffering from. Investigating the economic profile of treatment alternatives in a trial setting provides a degree of internal validity. This contrasts with the observational studies which may be subject to selection bias. Some studies employed simple decision analytic models based on expert opinion and secondary data taken from the literature. 22 24 Capturing the opinion of experts is a valid technique in the absence of comparative trial results, although empirical confirmation of assumptions used in the model is needed. It should also be noted that many of the primary studies used in modelling studies were uncontrolled, enrolled a small sample, and clinical data used were quite old. Therefore, more attention needs to be paid to the study design and quality of data. Further work is required to standardize methods used in modelling studies. It was possible to determine the economic value of antimicrobial therapy in the four studies which performed an incremental analysis. 22,27,29,30 Two studies computed average cost effectiveness. 23,24 Insight into the economic value of an antibiotic requires that the additional costs and effectiveness of the antibiotic as compared with the best, currently used alternative is calculated. This necessitates the calculation of incremental rather than average cost effectiveness ratios. Finally, three studies measured multiple consequences (so-called cost consequence analysis ), and did not combine costs and consequences into cost effectiveness ratios. 25,26,28 studies. 25,26,28 Factors affecting the economic value of antimicrobial therapy of COPD The economic value of antimicrobial therapy of COPD is influenced by a number of factors including diagnosis, effectiveness, antimicrobial resistance, patient compliance with treatment, and treatment failure. Diagnosing a bacterial exacerbation is rendered difficult by the fact that the diagnosis is generally based on patients self-reported clinical symptoms and high-quality sputum specimens are not always available. 1 This implies that antibiotics may be prescribed to treat that are not bacterial in origin. A recent Belgian survey of general practitioners and pulmonologists compared actual management with GOLD recommendations. 31 The authors concluded that antibiotics were overused in the treatment of COPD and that their use should be limited to those where bacterial infection is likely to be present. Inappropriate prescribing of antibiotics is likely to have significant cost and health consequences. 32 The relative effectiveness of available antibiotics plays a key role in determining the economic value of antimicrobial therapy. Studies suggest that second-generation macrolides and fluoroquinolones are marginally more or equally effective as firstgeneration antibiotics (aminopenicillins, macrolides, tetracyclines). 33,34 However, this literature is limited by the fact that most trials are powered to demonstrate equivalence rather than clinical superiority; have enrolled small samples that are not always representative of the patient population; and do not control for concomitant therapy or for co-morbidities. There is a need to investigate whether second-generation macrolides and fluoroquinolones are more effective than first-generation antibiotics. Given that these drugs have higher acquisition costs, their relative effectiveness will determine whether or not they exhibit a superior economic value than first-generation antibiotics. Antimicrobial resistance compromises the effectiveness and, thus, undermines the economic value of antimicrobial therapy. Increased resistance of the main respiratory pathogens H. influenzae, S. pneumoniae, and M. catarrhalis to first-generation antibiotics has been observed. 35,36 Second-generation antibiotics such as fluoroquinolones exhibit low resistance and a broad spectrum of activity. 37 GOLD guidelines recommend that the choice of antibiotic takes account of local patterns of antimicrobial sensitivity among bacterial pathogens. 1 The success of therapy also depends on patient compliance, with compliance being affected by the frequency of dosing, duration of treatment, adverse events, ease of administering drugs, ease of packaging, and price. 38 Only one economic evaluation of antimicrobial therapy quantified patient compliance: rates of compliance defined as intake of at least 80% of prescribed dose varied between 76% and 83%. 28 Various strategies to enhance patient compliance with antimicrobial therapy in respiratory tract infections have been proposed such as patient education, once-daily dosing schedules, a convenient and acceptable form of medication, easy-to-open packaging, and the choice of an antibiotic with few side effects. 38 Antimicrobial resistance and poor patient compliance need to be considered when choosing an antibiotic because they may lead to treatment failure and further antimicrobial therapy or hospitalization. Treatment failure may also be caused by

24 a number of host factors. The literature suggests that frequency of, presence of comorbidities, impairment in lung function, need for more aggressive bronchodilator therapy, and previous hospitalization predict treatment failure. 12,39,40 The ability to identify patients at a higher risk of failing treatment should aid clinicians in their choice of antibiotic and inform decisions about hospital admission. 41 Discussion This literature review summarized evidence on economic aspects of antimicrobial therapy of COPD. In light of the small number of economic evaluations and methodological limitations of these studies, it is not possible to recommend an antibiotic for preferential use in the management of COPD on economic grounds. Defining an optimal strategy to manage COPD is difficult given that studies use different definitions of an exacerbation. Although variation in definitions may appear slight, such differences matter when interpreting and comparing results across studies. Symptom-based definitions have been linked to therapeutic decisions, 4 but may under-report. 3 Defining an exacerbation in terms of healthcare resource use recognizes that more severe are associated with higher resource use, but inhibits international comparison because treatment patterns and, thus, classification systems of exacerbation severity vary across countries. The cost-of-illness literature has demonstrated that inpatient costs represent the principal component of costs of treating COPD. This indicates that considerable savings can be generated by more effective antibiotics that allow patients to be managed in primary care and that prevent treatment failure and hospitalization. Acquisition costs of antimicrobial drugs make up a limited percentage of treatment costs. However, these cost estimates are biased given the difficulties involved in correctly diagnosing and treating patients suffering from in daily practice. Furthermore, estimates were derived for specific patient samples and may not be generalizable to the population of patients suffering from. Evidence is emerging that second-generation antibiotics such as fluoroquinolones may have a favourable economic profile as compared with firstgeneration antibiotics. This seems to arise from the fact that fluoroquinolones have higher acquisition costs, but are also associated with less treatment failure, more time between exacerbation episodes, and lower hospitalization costs. Fluoroquinolones are likely to have a better economic profile in countries where inpatient costs account for a higher percentage of costs of managing. If corroborated by future research, there may be a role for fluoroquinolones as a first-line treatment for COPD in the presence of increased resistance to first-generation antibiotics or for patients who are at increased risk of treatment failure. The Canadian Thoracic Society and the Canadian Infectious Disease Society recommend fluoroquinolones as first-line treatment for patients suffering from CB who are at risk of failing therapy with first-generation antibiotics such as aminopenicillins, macrolides and tetracyclines. 42 When comparing the economic value of different generations of antibiotics, it is crucial that studies have sufficient statistical power to explore the impact of antibiotics on time between, treatment failure, outpatient and inpatient costs. Studies not only need to enroll a sufficient number of patients, but also need to follow up these patients for a minimum period of time in order to be able to detect potential differences in these outcome parameters and in the economic value of various antibiotics. There is a need for more, better-designed and comprehensive economic evaluations examining the economic profile of antimicrobial therapy of COPD. Studies need to be carried out that collect primary data on cost effectiveness and cost utility. Analyses based on randomized controlled trials provide a degree of internal validity, while analyses based on cohort studies reflect reallife practice. 13 Alternatively, modelling approaches can be employed based on high-quality data taken from the literature. Moreover, to date, there have been no cost benefit analyses that balance costs of antimicrobial therapy against the patient s willingness to pay to avoid. Acknowledgements Financial support for this research was received from Bayer HealthCare Pharmaceuticals. The authors have no conflicts of interest that are directly relevant to the content of this manuscript. References S. Simoens et al. 1. Pauwels RA, Buist AS, Calverley PMA, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and

Antimicrobial use in acute of COPD 25 prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Pulmonary Disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163:1256 76. 2. Anthonisen NR, Manfreda J, Warren CP, Hershfield ES, Harding GK, Nelson NA. Antibiotic therapy in of chronic obstructive pulmonary disease. Ann Intern Med 1987;106:196 204. 3. Seemungal TAR, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:1418 22. 4. Soto FJ, Varkey B. Evidence-based approach to acute of COPD. Curr Opin Pulm Med 2003;9:117 24. 5. Miravittles M, Murio C, Guerrero T, Gisbert R. Pharmacoeconomic evaluation of acute of chronic bronchitis and COPD. Chest 2002;121:1449 55. 6. Ashton CM, Ferguson JA, Goldacre MJ. In-patient workload in medical specialities 2. Profiles of individual diagnoses from linked statistics. Q J Med 1995;88:661 72. 7. Connors AF, Dawson NV, Thomas C, et al. Outcomes of acute of severe chronic obstructive lung disease. Am J Respir Crit Care Med 1996;154:959 67. 8. Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest 2000;117(Suppl.):5 9. 9. Sethi S. Infectious etiology of acute of chronic bronchitis. Chest 2000;117(Suppl.):380 5. 10. Saint S, Bent S, Vittinghoff E, Grady D. Antibiotics in chronic obstructive pulmonary disease. A metaanalysis. J Am Med Assoc 1995;273:957 60. 11. British Thoracic Society Nebulizer Project Group. Nebulizer therapy. Guidelines. Thorax 1997;52:4 24. 12. McCrory DC, Brown C, Gelfand SE, Bach PB. Management of acute of COPD: a summary and appraisal of published evidence. Chest 2001;119:1190 209. 13. Drummond M, O Brien B, Stoddart G, Torrance G. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press; 1997. p. 28 9. 14. Niederman MS, McCombs JS, Unger AN, Kumar A, Popovian R. Treatment cost of acute of chronic bronchits. Clin Ther 1999;21:576 91. 15. Rychlik R, Pfeil T, Daniel D, et al. Zur sozioökonomischen relevanz akuter exazerbationen der chronischen bronchitis in der Bundesrepublik Deutschland (Socioeconomic relevance of acute of chronic bronchitis in the Federal Republic of Germany). Dtsch Med Wochenschr 2001;126:353 9. 16. McGuire A, Irwin DE, Fenn P, et al. The excess cost of acute of chronic bronchitis in patients aged 45 and older in England and Wales. Value Health 2001;4: 370 5. 17. Saint S, Flaherty KR, Abrahamse P, Martinez FJ, Fendrick AM. Acute exacerbation of chronic bronchitis: disease-specific issues that influence the cost effectiveness of antimicrobial therapy. Clin Ther 2001;23:499 512. 18. Andersson F, Borg S, Jansson SA, et al. The costs of in chronic obstructive pulmonary disease (COPD). Respir Med 2002;96:700 8. 19. Miravittles M, Jardim JR, Zitto T, Rodrigues JE, Lopez H. Pharmacoeconomic study of antibiotic therapy for acute of chronic bronchitis and chronic obstructive pulmonary disease in Latin America. Arch Bronconeumol 2003;39:549 53. 20. Oostenbrink JB, Rutten-van Mölken MPMH. Resource use and risk factors in high-cost of COPD. Respir Med 2004;98:883 91. 21. Lucioni C, Donner CF, De Benedetto F, et al. I costi della broncopneumopatia cronica ostruttiva: la fase prospettica dello studio ICE (Italian costs for in COPD). Pharmacoeconomics Ital Res Articles 2005;7:119 34. 22. Backhouse R, Shakespeare A, Hutton J. Economic evaluation of alternative antibiotic regimens in the management of acute of chronic bronchitis. Br J Med Econ 1995;8:11 25. 23. Wool C, Cerutti R, Garbagna N, Grossi E. A cost effectiveness study of four different antibiotics in the treatment of acute of chronic obstructive pulmonary disease. Br J Med Econ 1996;10:159 68. 24. Van Barlingen HJJ, Nuijten MJC, Volmer T, Sanderson DJ, Lacey LF. Model to evaluate the cost effectiveness of different antibiotics in the management of acute bacterial of chronic bronchitis in Germany. J Drug Assess 1998;1:575 728. 25. Destache CJ, Dewan N, O Donohue WJ, Campbell JC, Angelillo VA. Clinical and economic considerations in the treatment of acute of chronic bronchitis. J Antimicrob Chemother 1999;43(Suppl. A):107 13. 26. Peng CC, Aspinall SL, Good CB, Atwood CW, Chang CCH. Equal effectiveness of older traditional antibiotics and newer broad-spectrum antibiotics in treating patients with acute of chronic bronchitis. South Med J 2003;96:986 91. 27. Halpern MT, Palmer CS, Zodet M, Kirsch J. Cost effectiveness of gemifloxacin: results from the GLOBE study. Am J Health Syst Pharm 2002;59:1357 65. 28. Llor C, Naberan K, Cots JM, Molina J, Miravitlles M. Economic evaluation of the antibiotic treatment of of chronic bronchitis and COPD in primary care. Int J Clin Pract 2004;58:937 44. 29. Grossman R, Mukherjee J, Vaughan D, et al. A 1-year community-based health economic study of ciprofloxacin vs usual antibiotic treatment in acute of chronic bronchitis. Chest 1998;113:131 41. 30. Torrance G, Walker V, Grossman R, et al. Economic evaluation of ciprofloxacin compared with usual antibacterial care for the treatment of acute of chronic bronchitis in patients followed for 1 year. Pharmacoeconomics 1999;16:499 520. 31. Decramer M, Bartsch P, Pauwels R, Yernault JC. Management of COPD according to guidelines. A national survey among Belgian physicians. Monaldi Arch Chest Dis 2003;59: 62 80. 32. Spiritus E. Antibiotic usage for respiratory tract infections in an era of rising resistance and increased cost pressure. Am J Manage Care 2000;23(Suppl. 1):1216 21. 33. Ball P, Tillotson G, Wilson R. Chemotherapy for chronic bronchitis: controversies. Presse Med 1995;24:189 94. 34. Gaillat J. Objectives for antibiotic therapy in acute of chronic bronchitis. Presse Med 2001;30: 17 22. 35. Felmingham D. Comparative antimicrobial susceptibility of respiratory tract pathogens. Chemotherapy 2004; 50(Suppl. 1):3 10. 36. Ferrara AM. New fluoroquinolones in lower respiratory tract infections and emerging patterns of pneumococcal resistance. Infection 2005;33:106 14. 37. Martinez FJ, Anzueto A. Appropriate outpatient treatment of acute bacterial of chronic bronchitis. Am J Med 2005;118(Suppl. 7A):39 44. 38. Kardas P. Patient compliance with antibiotic treatment for respiratory tract infections. J Antimicrob Chemother 2002; 49:897 903.

26 S. Simoens et al. 39. Miravittles M, Murio C, Guerrero T. Factors associated with relapse after ambulatory treatment of acute of chronic bronchitis. A prospective multicenter study in the community. Eur Respir J 2001;17:928 33. 40. Miravitlles M, Llor C, Naberan K, Cots JM, Molina J. Variables associated with recovery from acute of chronic bronchitis and chronic obstructive pulmonary disease. Respir Med 2005;99:955 65. 41. Habib MP. Management of clinical failures in non-icu patients with chronic obstructive pulmonary disease. Chemotherapy 2001;47(Suppl. 4): 39 46. 42. Balter MS, La Forge J, Low DE, Mandell L, Grossman RF. Canadian guidelines for the management of acute of chronic bronchitis. Can Respir J 2003; 10(Suppl. B):3 32.