RESEARCH BMJ: first published as 10.1136/bmj.b2538 on 14 July 2009. Downloaded from Cost effectiveness of COX 2 selective inhibitors and traditional NSAIDs alone or in combination with a proton pump inhibitor for people with osteoarthritis Nicholas Latimer, research fellow in health economics,1 Joanne Lord, reader in health economics,2 Robert L Grant, senior technical adviser,3 medical statistician,4 Rachel O’Mahony, research fellow,3 John Dickson, community physician in rheumatology,5 Philip G Conaghan, professor of musculoskeletal medicine6 on behalf of the National Institute for Health and Clinical Excellence Osteoarthritis Guideline Development Group 1Health Economics and Decision ABSTRACT adding a proton pump inhibitor to a COX 2 selective Science, School of Health and Objectives To investigate the cost effectiveness of cyclo- inhibitor (used at the lowest licensed dose) was a cost Related Research, University of Sheffield, Sheffield S1 4DA oxygenase-2 (COX 2) selective inhibitors and traditional effective option, even for patients at low risk of 2Health Economics Research non-steroidal anti-inflammatory drugs (NSAIDs), and the gastrointestinal adverse events (incremental cost Group, Brunel University, addition of proton pump inhibitors to these treatments, effectiveness ratio approximately £10 000). Uncertainties Middlesex UB8 3PH for people with osteoarthritis. around relative adverse event rates meant relative cost 3 National Collaborating Centre for Design An economic evaluation using a Markov model effectiveness for individual COX 2 selective inhibitors and Chronic Conditions, Royal College of Physicians of London, London and data from a systematic review was conducted. traditional NSAIDs was difficult to determine. NW1 4LE Estimates of cardiovascular and gastrointestinal adverse Conclusions Prescribing a proton pump inhibitor for http://www.bmj.com/ 4Royal College of Physicians of events were based on data from three large randomised people with osteoarthritis who are taking a traditional London, London NW1 4LE controlled trials, and observational data were used for NSAID or COX 2 selective inhibitor is cost effective. The 5Redcar and Cleveland Primary Care Trust, Guisborough Primary sensitivity analyses. Efficacy benefits from treatment were cost effectiveness analysis was sensitive to adverse event Care Hospital, North Yorkshire estimated from a meta-analysis of trials reporting total data and the specific choice of COX 2 selective inhibitor or TS14 6HZ Western Ontario and McMaster Universities (WOMAC) NSAID agent should, therefore, take into account 6 Section of Musculoskeletal osteoarthritis index score. Other model inputs were individual cardiovascular and gastrointestinal risks. Disease, University of Leeds, Leeds LS7 4SA obtained from the relevant literature. The model was run on 28 September 2021 by guest. Protected copyright. Correspondence to: P G Conaghan for a hypothetical population of people with INTRODUCTION [email protected] osteoarthritis. Subgroup analyses were conducted for Traditional non-steroidal anti-inflammatory drugs people at high risk of gastrointestinal or cardiovascular Cite this as: BMJ 2009;339:b2538 (NSAIDs) and the newer cyclo-oxygenase-2 (COX 2) doi:10.1136/bmj.b2538 adverse events. selective inhibitors are commonly prescribed for peo- Comparators Licensed COX 2 selective inhibitors ple with osteoarthritis. Approximately half of the peo- (celecoxib and etoricoxib) and traditional NSAIDs ple with osteoarthritis in the United Kingdom who (diclofenac, ibuprofen, and naproxen) for which suitable require medication are treated with an NSAID or a data were available were compared. Paracetamol was COX 2 selective inhibitor.1 COX 2 selective agents also included, as was the possibility of adding a proton are currently prescribed much less often than tradi- pump inhibitor (omeprazole) to each treatment. tional NSAIDs; in 2007, for example, the COX 2 selec- Main outcome measures The main outcome measure was tive inhibitors celecoxib and etoricoxib accounted for cost effectiveness, which was based on quality adjusted approximately 5.8% of total NSAID prescriptions in life years gained. Quality adjusted life year scores were England and approximately 20% of the total spend.2 calculated from pooled estimates of efficacy and major Although traditional NSAIDs and COX 2 selective adverse events (that is, dyspepsia; symptomatic ulcer; inhibitors seem similar in terms of symptom relief in complicated gastrointestinal perforation, ulcer, or bleed; such patients, traditional NSAIDs are associated with myocardial infarction; stroke; and heart failure). gastrointestinal side effects. COX 2 selective agents Results Addition of a proton pump inhibitor to both COX 2 were developed to reduce the gastrointestinal side selective inhibitors and traditional NSAIDs was highly effects of this drug class. In addition, concerns have cost effective for all patient groups considered been raised over the cardiovascular safety of both (incremental cost effectiveness ratio less than £1000 COX 2 selective inhibitors and traditional NSAIDs.34 (€1175, $1650)). This finding was robust across a wide New data indicate that co-prescribing gastroprotective range of effectiveness estimates if the cheapest proton agents with both traditional NSAIDs and COX 2 selec- pump inhibitor was used. In our base case analysis, tive agents is beneficial.5-7 BMJ | ONLINE FIRST | bmj.com page 1 of 9 RESEARCH The latest National Institute for Health and Clinical multinational etoricoxib and diclofenac arthritis long- Excellence clinical guidance for the management of term (MEDAL) study.20-22 Two of these trials (CLASS BMJ: first published as 10.1136/bmj.b2538 on 14 July 2009. Downloaded from osteoarthritis provides an update to previous recom- and the MEDAL study) included a minority of people mendations on the use of COX 2 selective with rheumatoid arthritis; however, the National Insti- inhibitors.8-11 The previous guidance recommended tute for Health and Clinical Excellence Osteoarthritis that these agents should not be used routinely for Guideline Development Group considered that the patients with osteoarthritis or rheumatoid arthritis relative risks of adverse events would be similar in peo- and should only be used in patients at high risk of ple with osteoarthritis and rheumatoid arthritis developing serious gastrointestinal adverse events on because there is no clear evidence of a relationship traditional NSAIDs. In addition, the guidance stated between drug induced adverse event rates and arthritis that there was no evidence to justify the simultaneous type. Table 1 gives an overview of the characteristics of prescription of gastroprotective agents with COX 2 the CLASS, TARGET, and the MEDAL study. These selective inhibitors. This National Institute for Health studies allow comparisons between the currently avail- and Clinical Excellence guidance and other published able COX 2 selective inhibitors (celecoxib and etori- economic analyses in this area preceded the latest evi- coxib) and three traditional NSAIDs (diclofenac, dence on adverse events and gastroprotection, ibuprofen, and naproxen), which together account for however.5912 In addition, drug prices have recently over 80% of NSAID prescriptions in England. 2 “No changed—particularly for proton pump inhibitors— treatment,” paracetamol, and the addition of a proton and the cost effectiveness of gastroprotective agents pump inhibitor (omeprazole) to each NSAID were also could, therefore, also change.13 considered. Topical NSAIDs were not included owing As part of the development of the latest National to data limitations. Institute for Health and Clinical Excellence guideline, we performed an economic evaluation of COX 2 selec- Model design tive inhibitors and traditional NSAIDs, and of the addi- The economic model is described in detail elsewhere.23 tion of gastroprotective agents to these treatments. Here, we give an overview of the model design and parameters. The principal inputs to the model are METHODS shown in tables 2, 3, and 4. We conducted a cost utility analysis according to the The model estimates the net impact of the treatment methods recommended by the National Institute for options on patient outcomes and expenditure, taking http://www.bmj.com/ Health and Clinical Excellence.14 The primary out- account of effects on the incidence of gastrointestinal come measure for the economic analysis was quality and cardiovascular adverse events as well as improve- adjusted life years. A healthcare payer perspective ments in the control of osteoarthritis symptoms. It is was taken—that of the NHS in England and Wales. made up of a series of health states between which a theoreticalcohortofpatientscanmove.Theratesoftran- Comparators sition between these states are estimated using clinical Despite the growth in the evidence base, data are still evidence. The health states represent the most frequent on 28 September 2021 by guest. Protected copyright. sparse regarding the adverse events associated with and severe adverse events: dyspepsia; symptomatic some NSAIDs. Amalgamating data from observa- ulcer; complicated gastrointestinal perforation, ulcer, tional trials with data from randomised controlled or bleed; myocardial infarction; stroke; and heart failure. trials was not feasible because of the differences in In addition, a patient can experience no adverse event, patient groups, drug doses, and adverse event or death. Death rates are based on age specific mortality