Skip Navigation


Rheumatology Advance Access originally published online on December 20, 2005
Rheumatology 2006 45(5):606-613; doi:10.1093/rheumatology/kei241
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
45/5/606    most recent
kei241v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Elliott, R. A.
Right arrow Articles by Symmons, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Elliott, R. A.
Right arrow Articles by Symmons, D.
Related Collections
Right arrow Health Economics
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2005. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Preventing non-steroidal anti-inflammatory drug-induced gastrointestinal toxicity: are older strategies more cost-effective in the general population?

R. A. Elliott, L. Hooper3, K. Payne4, T. J. Brown3, C. Roberts1 and D. Symmons2

School of Pharmacy and Pharmaceutical Sciences, 1 School of Epidemiology and Health Sciences and 2 ARC Epidemiology Unit, University of Manchester, Oxford Road, Manchester M13 9PL, 3 University of Manchester and Central Manchester and Manchester Children's University Healthcare Trust, MANDEC, University Dental Hospital, Higher Cambridge Street, Manchester M15 6FH and 4 North West Genetics Knowledge Park (Nowgen), The Nowgen Centre, University of Manchester, Grafton Street, Manchester M13 9WU, UK.

Correspondence to: R. Elliott, Clinical Senior Lecturer in Pharmacy, School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Oxford Road, Manchester M13 9PL, UK. E-mail: rachel.elliott{at}manchester.ac.uk


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 References
 
Objectives. To assess the relative cost-effectiveness of five gastroprotective strategies for patients in the general population not judged to be at high gastrointestinal (GI) risk requiring regular traditional (t) non-steroidal anti-inflammatory drugs (NSAIDs) for over 3 weeks: tNSAID/H2 receptor antagonists (H2RAs); tNSAID/proton pump inhibitors (PPIs); tNSAID/misoprostol; COX-2 preferential NSAIDs or COX-2-specific NSAIDs (COXIBs).

Methods. A systematic review of outcomes and UK cost data were combined in an incremental economic analysis. Incremental cost-effectiveness ratios were generated for quality-adjusted life years (QALYs) gained.

Results. Cost–utility analysis showed a tNSAID with a H2RA is safer and less costly than tNSAIDs alone, and equally effective and less costly than COXIBs. tNSAID/misoprostol was also dominated by tNSAID/H2RA due to withdrawal caused by side-effects reducing overall health status. The incremental increase in QALYs gained by using COXIBs instead of tNSAID/H2RA would cost £670 000 per QALY gained. The incremental increase in QALYs gained by using tNSAID/PPI instead of COXIBs would cost £26 000 per QALY gained. If the decision-maker will pay up to £140 000 per extra QALY, the optimal strategy is tNSAID/H2RA. If the decision-maker will pay over this the optimal strategy is tNSAID/PPI.

Conclusion. The economic analysis suggests that there may be a case for prescribing H2RAs in all patients requiring NSAIDs. Our recommendations are tentative due to the quality of the data available and the assumptions we have had to make in our model, and it is possible that other strategies may be preferred in patients with higher baseline GI risk.

KEY WORDS: Health economics, Outcomes, COX-2-specific inhibitors, Health resource utilization, NSAIDs, Preventive strategies, Gastrointestinal events, Drug utilization


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 References
 
Traditional non-steroidal anti-inflammatory drugs (tNSAIDs) may cause gastrointestinal (GI) toxicity that results in significant morbidity and mortality [1]. Clinicians may protect against NSAID-induced GI toxicity by co-prescribing gastroprotective agents (GPA) such as H2-receptor antagonists (H2RAs), proton pump inhibitors (PPIs) or misoprostol, or using preferential COX-2 inhibitors (etodolac, meloxicam, nabumetone and nimesulide) or specific COX-2 inhibitors (COXIBs) (celecoxib, etoricoxib).

Our systematic reviews of randomized controlled trials (RCTs) [2] assessed the effectiveness of these strategies for long-term use. We separated preferential COX-2 inhibitors from tNSAIDs and COXIBs because they are marketed as ‘safe’ NSAIDs. Each strategy except tNSAID plus H2RAs, where there were few event data, showed significant protection against symptomatic ulcers. Each strategy except preferential COX-2 inhibitors, where there were few event data, showed significant protection against endoscopic ulcers. Misoprostol was associated with a high drop-out level, but reduced incidence of serious GI events. The evidence was limited by a shortage of trials on H2RAs and PPIs, and of head-to-head comparisons between the strategies. There were insufficient data on serious events (such as death) or other outcomes (such as cardiovascular events) to make comparisons. The review was not able to identify an optimal GPA strategy for decision-makers based on the clinical indicators reported.

Newer medicines generally have a higher acquisition cost than older medicines. Claims of increased effectiveness and safety in newer medicines may prematurely reduce the use of older, effective medicines with a proven safety record. Reductions in costs associated with management of GI toxicity are often claimed to offset costs of COX-2 inhibitors. We found over 40 economic analyses of NSAIDs and GPAs published between 1989 and 2005, mostly comparing preferential COX-2 inhibitors, COXIBs or misoprostol with tNSAIDs. The 10 head-to-head economic analyses found were modelling studies, synthesizing clinical and cost data from a range of sources, none from the UK. Despite the limitations of models, in the absence of alternatives they allow decision-makers to consider, explicitly, the costs and outcomes associated with their decisions. A recent US modelling analysis suggests that a tNSAID plus PPI is a preferred strategy over COXIBs [3]. There were no head-to-head economic evaluations of all five strategies. We performed an economic analysis to establish which strategies are cost-effective interventions to prevent NSAID-induced GI toxicity in a general population not judged to be at high risk.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 References
 
A systematic review and meta-analysis of outcomes [2] was combined with up-to-date UK resource use and unit costs in an incremental economic analysis. Due to lack of primary data we had to carry out a modelling study. To prevent addition of yet another NSAID/GPA model [4], we based our study approach on models used by Maetzel et al. [5] and Moore et al. [6]. Combining data from different sources assumes that the source populations are comparable, which is a limitation of all models. We have presented our model explicitly, with a clear explanation of data sources and any assumptions made.

Strategies compared in the economic analysis
The economic analysis compared tNSAID alone, tNSAID + PPI, tNSAID + H2RA, tNSAID + misoprostol and COX-2 preferential inhibitor + COXIB. Agents and doses used are summarized in Table 1. In common with National Institute for Clinical Excellence (NICE) guidelines, tNSAIDs, COX-2 preferential inhibitors and COXIBs were assumed to have equal analgesic efficacy in a group of patients, although it is recognized that there is inter-patient variability in efficacy [7]. The most commonly prescribed drugs in each class were selected and a class effect was assumed for H2RAs and PPIs with regards to gastroprotective efficacy. In the meta-analysis we only included studies that used the therapeutic doses of GPAs required for efficacy (such as 300 mg ranitidine per day) [2]. Diclofenac and ibuprofen are the two most commonly prescribed NSAIDs in the UK [8]. This analysis assumed equivalent safety for equivalent doses of ibuprofen and diclofenac [9, 10]. Risk of adverse GI outcomes from tNSAIDs and COX-2 preferential inhibitors increases with dose [11, 12]. In this study, it was assumed that a dose in the middle of the prescribing range for each agent was used, in line with the defined daily dose (DDD) for that drug [13].


View this table:
[in this window]
[in a new window]
  		TABLE 1. Probabilities from meta-analysis used in economic analysis [2]

 
Patient population
The meta-analysis included patients with chronic arthritis (primarily rheumatoid arthritis and osteoarthritis) who require regular NSAIDs for more than 3 weeks, with and without risk factors for adverse GI events. The review included 112 RCTs (74 666 participants). Two-thirds of patients were women, baseline GI risk was not known in 62/112 studies and 101/112 studies reported a mean age below 65 yr of age. Only 138 deaths and 248 serious GI events were reported.

Cost and benefits identified in the economic analysis
Figure 1 represents the decision-analytic model (‘decision tree’). Tables 1and 2 summarize the clinical and cost data.


Figure 1
View larger version (9K):
[in this window]
[in a new window]
  		FIG. 1. Decision analytic model for economic analysis of tNSAIDs vs tNSAIDs + GPAs vs COX-2 preferential inhibitors or COXIBs.

 

View this table:
[in this window]
[in a new window]
  		TABLE 2. Treatment pathways [5]

 
Data extraction
Clinical data
Outcome data are reported in the systematic review as relative risk, with 95% confidence intervals (CIs), for gastroprotective strategies (arms 2 to 6) vs tNSAIDs (arm 1) [2]. Due to insufficient head-to-head comparisons, indirect comparisons were used. Lack of comparative non-GI outcomes meant that we assessed GI outcomes only (no GI adverse event, GI discomfort, uncomplicated ulcer and serious GI complication). Symptomatic ulcers, serious GI events and death rates were reported rarely and/or unreliably. These outcomes were recorded particularly poorly in earlier studies, which principally recorded endoscopic ulcers. It is likely that such patients were withdrawn, rather than being allowed to progress to symptomatic ulcers or GI complications. To be able to compare all strategies we had to use the parameter ‘endoscopic ulcers’ (see Table 1) and assume a proportion of those ulcers will become symptomatic.

The MUCOSA study estimated that 85% of endoscopic ulcers remained silent [5]. Therefore, P[symptomatic ulcer, given endoscopic ulcer] in this analysis was assumed to be 15% of P[endoscopic ulcer]. This probability was assigned a range (0–30%) and distribution in the analysis.

Recently, rofecoxib, a COXIB, has been suggested to have reduced cardiovascular safety compared with naproxen [14], which led to its voluntary worldwide withdrawal. The cardiovascular safety of other COXIBs is now under investigation to examine whether this is a ‘class effect’. Celecoxib and valdecoxib studies have not demonstrated increased risk of thrombosis [15–17], but long-term studies are needed. In contradiction to recent events, alternative hypotheses suggest that COX-2 enzyme inhibition may actually reduce the risk of cardiac events [18–21]. In a climate of such uncertainty, drug manufacturers must assess cardiovascular safety urgently. We used celecoxib in our study and, due to the substantive uncertainty around relative cardiovascular safety, assumed equivalent cardiac outcomes between arms.

Treatment pathways, such as proportion and management of patients requiring hospital admission, occurring due to these outcomes were obtained primarily from two sources, which provided the most comprehensive follow-up of treatment pathways [5, 6] (see Table 2). The most up-to-date UK information on gastric bleed death rate (17.3%) was used [1].

We assumed that minor and major adverse events occur independently of one another. Evidence suggests little correlation between abdominal symptoms and the presence of NSAID-induced gastric lesions [22, 23]. Furthermore, the clinical course of major events such as duodenal ulcer haemorrhage is not significantly different in patients with and without minor events such as dyspepsia [24]. Intolerance to misoprostol leads to many patients withdrawing from treatment. This was dealt with explicitly in the analysis.

We assumed that GI discomfort rates reported in RCTs in the meta-analyses related to persistent, not transient, GI discomfort, requiring patients to return to their GP. Most patients who experience GI discomfort do so within the first 2 weeks of initiating therapy [25]. We assumed that patients with persistent GI discomfort will not remain on the same therapy but will be changed to the most commonly prescribed alternative, a tNSAID plus PPI [26]. We assumed that patients already on a tNSAID plus PPI would be switched to a COXIB.

Quality-adjusted life years (QALYs) were the outcome measure used in the analysis. To calculate QALYs we incorporated validated utilities for the health states experienced along each treatment pathway in the model (see Table 3). We used published estimates for time spent in each health state and calculated the utility lost (‘disutility’) associated with each treatment pathway. Where a published range was not available, we assigned wide ranges to each health state and a distribution to reflect uncertainty around that parameter in the analysis.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Utility and duration of GI adverse event health states for cost–utility analysis

 
Due to the lack of clinical data, we assumed that adverse event rates and associated costs remain constant over a 6-month period, which was used as the time horizon.

Resource use data
The perspective of the analysis was the UK National Health Service (NHS). Costs were calculated for the 2003 price year. No detailed resource use data were collected in RCTs in the meta-analysis. We used recent, relevant, patient-based and disaggregated cost data (see Table 4).


View this table:
[in this window]
[in a new window]
  		TABLE 4. Resource use and unit costs used in economic evaluation

 
Cost data were attached to the 12 possible treatment pathways for each of the six strategies (see Table 5). Each cost variable was assigned a range and gamma distribution in the analysis. We assumed that if a cohort of patients was assigned to one of the six strategies, we knew from the probabilities for that strategy how many patients would go along each pathway. As we knew how much each pathway cost for one patient, we were able to calculate how much the total cost would be for a cohort of patients for each of the six strategies. This cost is the ‘expected cost’, usually expressed as ‘expected cost per patient’.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Resource use associated with each treatment pathway (adapted from Moore et al.) [6]

 
Uncertainty around costs and outcomes
Probabilistic sensitivity analysis was used to generate measures of uncertainty (variance) because these were not available from data sources. Each variable had a mean value. The type of distribution was informed by the nature of each variable (probability, cost, QALY). This allowed us to generate 95% CIs (see Table 6). The probabilistic analysis summed the results of multiple analyses (iterations). Each iteration sampled the values for the variables at random from the specified distributions. The main areas of uncertainty in the model were estimates of effectiveness of each strategy. The sampling method used was Monte Carlo, expected value. The simulation software used was @RISK, as an add on to Microsoft Office Excel v.7.0 [27]. The number of iterations for each simulation to generate reliable statistics was determined by the software, which halted the simulation when convergence at less than 1.5% in percentile values, mean and standard deviation was achieved.


View this table:
[in this window]
[in a new window]
  		TABLE 6. Summary of incremental economic analysis for cost per QALY gained

 
Cost-effectiveness analysis
Expected costs and outcomes were generated for each of the six arms for outcome measures derived directly from the meta-analysis and for QALYs. We generated incremental cost-effectiveness ratios (ICERs), a series of cost-effectiveness acceptability curves (CEACs) and a cost-effectiveness acceptability frontier (CEAF) to reflect uncertainty in costs, effects and the maximum willingness to pay for a decision-maker.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 References
 
The least safe to most safe strategies were: tNSAID alone, COX-2 preferential inhibitor, tNSAID + H2RA, COXIB, tNSAID + misoprostol and tNSAID + PPI (see Table 1). The least costly to most costly strategies were: tNSAID + H2RA, tNSAID + misoprostol, tNSAID alone, COXIB, COX-2 preferential inhibitor, and tNSAID + PPI (see Table 6).

Cost per QALY gained
ICERs were generated for each GPA strategy against tNSAIDs alone:

Formula
Non-parametric bootstrapping with replacement was used to simulate the variance of these ICERs sampling from the simulated distribution obtained above.

The decision as to which strategies were cost-effective depends on the maximum willingness to pay for each QALY gained, and the uncertainty that the decision-maker is willing to tolerate. This is shown in the CEACs for QALY gained for tNSAID vs each strategy (Fig. 2). From this, we can say that there is a 95% probability that tNSAID + H2RA dominates tNSAIDs alone (less costly and more effective). There is also a 28% probability that tNSAID + misoprostol is dominant. However, this information does not tell us which is the preferred strategy.


Figure 2
View larger version (24K):
[in this window]
[in a new window]
  		FIG. 2. Cost-effectiveness acceptability curve for tNSAID + GPA, COX-2 preferential inhibitor or COXIB vs tNSAID alone for cost per QALY gained.

 
Cost–utility analysis showed that a tNSAID with a H2RA is safer and less costly than tNSAIDs alone, and equally effective and less costly than COX-2 preferential NSAIDs. tNSAID/misoprostol was also dominated by tNSAID/H2RA due to withdrawal caused by side-effects reducing overall health status. The incremental increase in QALYs gained by using COXIBs instead of tNSAID/H2RA would be £670 000 per QALY gained. The incremental increase in QALYs gained by using tNSAID/PPI instead of COXIBs would be £26 000 per QALY gained (Table 6).

The optimal strategy is the intervention with the highest net benefit. In a multiple intervention decision such as in this analysis, the CEAC for each intervention can be established by calculating the proportion of the iterations from the simulation where that intervention is optimal over the other interventions, at a given cost per outcome. Repeating this process for each intervention and simultaneously graphing the CEAC allows the derivation of a CEAF [28]. A CEAF was generated for tNSAID + GPA, COX-2 preferential inhibitors, and COXIB vs tNSAID alone for cost per QALY gained (see Fig. 3).


Figure 3
View larger version (20K):
[in this window]
[in a new window]
  		FIG. 3. Cost-effectiveness acceptability frontier for tNSAID + GPA, COX-2 preferential inhibitor or COXIB vs tNSAID alone for cost per QALY gained.

 
The frontier is the uppermost combination of lines. The switch points show where the optimal strategy changes, and is equivalent to the ICER between these two options. If the decision-maker will pay up to £140 000 per extra QALY, the optimal strategy is COX-1/H2RA. If the decision-maker will pay over this the optimal strategy is tNSAID/PPI. tNSAID/misoprostol, COX-2 preferential inhibitors and COXIBs are never optimal.


   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
References
 
This study examined the relative cost-effectiveness of newer against older gastroprotective strategies. Our model is unique in that it compares all strategies. This has limited our ability to stratify our patients by risk group. However, our economic analysis suggested that tNSAIDs + H2RAs are among the least effective of the GPAs, but may reduce gastrointestinal risk for the general population at a lower overall cost to the health-care provider than tNSAIDs alone. H2RAs are not as effective at reducing GI effects as the other GPA strategies, but they are effective. Other strategies are equally, or more, effective, but at an increased overall cost to the health-care provider than NSAIDs alone. Therefore, the decision-maker has to decide how much extra they are prepared, or are able, to pay to avoid undesirable outcomes.

The data for endoscopic ulcers were the most robust data obtained over all, although this analysis suggests that there is still an unacceptable level of uncertainty as far as decision-making is concerned. This was not the most clinically relevant measure because most endoscopic ulcers do not develop into clinically significant ulcers or bleeds [5]. If we had not included this parameter in the model, we would not have been able to identify the dominant effect of H2RAs.

In our systematic review, only 138 deaths and 248 serious GI events were reported for 74 666 participants in 112 trials [2]. It is likely that serious GI outcomes are underreported in trials as patients may be withdrawn before events occur. The use of validated health status data allowed us to generate estimates of changes in utility associated with all the GI outcomes in the model.

The model was constructed to provide a conservative estimate of cost-effectiveness and was biased in favour of COXIBs because we did not extend the effect of treatment beyond the length of the trials. Also, due to current widespread uncertainty associated with differences in cardiovascular outcomes caused by these medicines [21], we assumed equivalent cardiovascular safety. Despite this, COXIBs were not the optimal strategy. Of the older strategies, misoprostol was dominated by H2RAs due to differences in acquisition costs and side-effect profile, and the extra effectiveness afforded by PPIs is only obtained at a large incremental cost per extra unit of outcome in a population with average risk of adverse GI events. A different outcome might have been obtained if only patients at high risk of GI adverse events had been considered.

Despite the large body of evidence in this area, there is very little information on the relative effectiveness of the five strategies due to the lack of head-to-head studies. This meant using indirect comparisons in the economic analysis, a method that can provide useful results [29] but which is not as robust as direct comparisons, and dependent on similarity of baseline risk in all groups. We were not able to assess the effect of baseline risk on cost-effectiveness, such as age, previous GI morbidity or aspirin taking due to the poor quality of subgroup reporting in the trials in the meta-analysis.

Direct healthcare costs were available only as reported estimates from clinicians [6]. There has been no head-to-head trial of gastroprotective strategies that has included an economic evaluation that collects prospective patient-based observational data. There is little or no patient-based information about the resource use consequences of NSAID-related GI events.

Recommendations for health care
The economic analysis suggests that there may be a case for prescribing H2RAs in all patients requiring NSAIDs. Other GPAs are more effective, but are associated with a greater cost. Misoprostol is recommended by the National Prescribing Centre [30], but patient acceptability is so low that prescribers may be reluctant to follow this advice. Our recommendations are tentative due to the quality of the data available and the assumptions we have had to make in our model, and it is possible that other strategies may be preferred in patients with higher baseline GI risk.

Recommending H2RAs in all patients requiring NSAIDs will only be effective if prescribers prescribe, and patients adhere to, the medicines. However, low levels of prescribing [31] and non-adherence [32] to dual therapies have been reported, and patients’ exposure to increased costs from two medicines instead of one is likely to contribute to this. This low uptake of GPAs increases cost-effectiveness ratios substantially [33]. Of related concern is the co-prescribing of GPAs with COXIBs, reportedly as high as 20% [33], and this non-evidence-based practice should be discouraged.

Implications for further research
The prescribing of medicines involves the patient, the prescriber, the payer and society. The challenge for prescribers and policy-makers is to use drugs appropriately and cost-effectively. Information for all decision-makers needs to be greatly improved through use of head-to-head comparisons in pragmatic trials, rigorously reported major outcomes and patient-centred outcomes and acquisition of observational cost data and follow-up data about the treatment of patients suffering adverse events.
Figure 4


   Acknowledgments
 
Many thanks to the kind and expert project steering group, Linda Davies (University of Manchester), Andrew Herxheimer (Cochrane Collaboration), Qasim Aziz (Hope Hospital, Salford), to Dr Andrew Moore for the detailed information he provided (Pain Research and Nuffield Department of Anaesthetics, University of Oxford) and to Professor Milton Weinstein for helpful input in earlier drafts of this paper (Harvard Center for Risk Analysis, Harvard University).

LH, RAE, DS, TJB and CR conceived and designed the study; RAE, KP and LH analysed the data and interpreted the results, TJB collected and collated data, CR provided statistical support, RAE drafted the manuscript. RAE, LH, KP, DS, TJB and CR critically revised the manuscript. RAE will act as guarantor for the paper.

The NHS Research and Development Health Technology Assessment Programme funded this study. The views expressed in this paper are those of the authors and not of the funding body.

DS has been reimbursed by Pharmacia for attending a conference. We have no further conflicts of interest to declare.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 

  1. Blower AL, Brooks A, Fenn GC et al. Emergency admissions for upper gastrointestinal disease and their relation to NSAID use. Aliment Pharmacol Therap 1997;11:283–91.[CrossRef][Web of Science][Medline]
  2. Hooper L, Brown TJ, Elliott RA, Payne K, Roberts C, Symmons D. The effectiveness of five strategies for the prevention of gastrointestinal toxicity induced by non-steroidal anti-inflammatory drugs: systematic review. Br Med J 2004;329:948–52.[Abstract/Free Full Text]
  3. Spiegel BMR, Chiou C-F, Ofman JJ. Minimizing complications from nonsteroidal antiinflammatory drugs: cost effectiveness of competing strategies in varying risk groups. Arthritis Rheum 2005;53:185–97.[CrossRef][Medline]
  4. Goldstein JL. Yet another model looking at the cost effectiveness of COX-2 selective inhibitors. Evidence-based Gastroenterol 2003;4:133–4.
  5. Maetzel A, Ferraz MB, Bombardier C. The cost-effectiveness of misoprostol in preventing serious gastrointestinal events associated with the use of nonsteroidal antiinflammatory drugs. Arthritis Rheum 1998;41:16–25.[CrossRef][Web of Science][Medline]
  6. Moore RA, Phillips CJ, Pellissier JM, Kong SX. Health economic comparisons of rofecoxib versus conventional nonsteroidal antiinflammatory drugs for osteoarthritis in the United Kingdom. J Med Econ 2001;4:1–17.
  7. National Institute for Clinical Excellence. Guidance on the use of cyclo-oxygenase (COX) II selective inhibitors, celecoxib, rofecoxib, meloxicam, and etodolac for osteoarthritis and rheumatoid arthritis, Technology Appraisal Guidance 27. London, NICE, 2001.
  8. Department of Health. Prescription cost analysis 2000. DoH national statistics, 2004 (http://www.dh.gov.uk/assetRoot/04/10/76/26/04107626.pdf).
  9. MacDonald TM, Morant SV, Robinson GC et al. Association of upper gastrointestinal toxicity of non-steroidal anti-inflammatory drugs with continued exposure: cohort study. Br Med J 1997;315:1333–7.[Abstract/Free Full Text]
  10. Rodriguez LAG, Cattaruzzi C, Troncon MG, Agostinis L. Risk of hospitalization for upper gastrointestinal tract bleeding associated with ketorolac, other nonsteroidal anti-inflammatory drugs, calcium antagonists and other antihypertensive drugs. Arch Intern Med 1998;158:33–9.[Abstract/Free Full Text]
  11. Henry D, Lim L, Rodriguez LAG et al. Variability in risk of gastrointestinal complications with individual non-steroidal anti-inflammatory drugs: results of a collaborative meta-analysis. Br Med J 1996;312:1563–6.[Abstract/Free Full Text]
  12. Hernandez-Diaz S, Rodriguez LAG. Association between nonsteroidal anti-inflammatory drugs and upper gastrointestinal tract bleeding/perforation. Arch Intern Med 2000;160:2093–9.[Abstract/Free Full Text]
  13. WHO Collaborating Centre for Drug Statistics Methodology. Complete ATC index 2003 (www.whocc.no/atcdd).
  14. Bombardier C, Laine L, Reicin A et al. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. N Engl J Med 2000;343:1520–8.[Abstract/Free Full Text]
  15. Silverstein FE, Faich G, Goldstein JL et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. J Am Med Assoc 2000;284:1247–55.[Abstract/Free Full Text]
  16. White WB, Strand V, Roberts R, Whelton A. Effects of the cyclooxygenase-2 specific inhibitor valdecoxib versus nonsteroidal antiinflammatory agents and placebo on cardiovascular thrombotic events in patients with arthritis. Am J Therap 2004;11:244–50.
  17. Solomon DH, Schneeweiss S, Glynn RJ et al. Relationship between selective cyclooxygenase-2 inhibitors and acute myocardial infarction in older adults. Circulation 2004;109:2068–73.[Abstract/Free Full Text]
  18. Hankey GJ, Eikelboom JW. Cyclooxygenase-2 inhibitors: are they really atherothrombotic, and if not, why not? Stroke 2003;34:2736–40.[Abstract/Free Full Text]
  19. Bogaty P, Brophy JM, Noel M et al. Impact of prolonged cyclooxygenase-2 inhibition on inflammatory markers and endothelial function in patients with ischemic heart disease and raised C-reactive protein: a randomized placebo-controlled study. Circulation 2004;110:934–9.[Abstract/Free Full Text]
  20. Baigent C, Patrono C. Selective cyclooxygenase 2 inhibitors, aspirin and cardiovascular disease. Arthritis Rheum 2003;48:12–20.[CrossRef][Web of Science][Medline]
  21. Hippisley-Cox J, Coupland C. Risk of myocardial infarction in patients taking cyclo-oxgenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs: population-based nested case-control analysis. Br Med J 2005;330:1366–72.[Abstract/Free Full Text]
  22. Taha AS, Dahill S, Sturrock RD, Lee FD, Russell RI. Predicting NSAID related ulcers: assessment of clinical and pathological risk factors and importance of differences in NSAID. Gut 1994;35:891–5.[Abstract/Free Full Text]
  23. Kubba AK, Whyman MR. Upper gastro-intestinal disease in Scotland: a survey of practice amongst Scottish gastroenterologists. J R Coll Surg Edin 1996;41:302–6.[Medline]
  24. Ming-Yu L, Hsien-Hong L, Pang-Chi C. Duodenal ulcer haemorrhage with or without dyspepsia. Am J Gastroenterol 1990;85:1343–5.[Medline]
  25. Peris F, Martinez E, Badia X, Brosa M. Iatrogenic cost factors incorporating mild and moderate adverse effects in the economic comparison of aceclofenac and other NSAIDs. Pharmacoeconomics 2001;19:779–90.[Medline]
  26. British Medical Association, Royal Pharmaceutical Society of Great Britain. British National Formulary, 45th edn. London: Pharmaceutical Press, 2003.
  27. @RISK: risk analysis and simulation add-in for Microsoft Excel, version 4.5. New York: Palisade Corporation, 2002.
  28. Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost effectiveness acceptability curves. Health Econ 2001;10:779–87.[CrossRef][Web of Science][Medline]
  29. Song F, Altman D, Glenny A, Deeks JJ. Validity of indirect comparison for estimating efficacy of competing interventions: empirical evidence from published meta-analyses. Br Med J 2003;326:472–5.[Abstract/Free Full Text]
  30. National Prescribing Centre. Vioxx (rofecoxib) withdrawal. MeReC EXTRA 15. Liverpool: National Prescribing Centre, 2004.
  31. Sturkenboom MCJM, Burke TA, Dieleman JP, Tangelder MJD, Lee F, Goldstein JL. Underutilization of preventive strategies in patients receiving non-steroidal anti-inflammatory drugs. Rheumatology 2003;42:iii23–iii31.[Abstract]
  32. Sturkenboom MCJM, Burke TA, Tangelder MJD, Dieleman JP, Walton S, Goldstein JL. Adherence to proton pump inhibitors or H2-receptor antagonists during the use of non-steroidal anti-inflammatory drugs. Aliment Pharmacol Therap 2003;18:1137–47.[CrossRef][Medline]
  33. Cox ER, Motheral B, Mager D. Verification of a decision analytic model assumption using real-world practice data: implications for the cost effectiveness of Cyclo-oxygenase 2 inhibitors (COX-2s). Am J Manag Care 2003;9:785–94.[Medline]
  34. Groeneveld PW, Lieu TA, Fendrick AM et al. Quality of life measurement clarifies the cost-effectiveness of Helicobacter pylori eradication in peptic ulcer disease and uninvestigated dyspepsia. Am J Gastroenterol 2001;96:338–47.[Web of Science][Medline]
  35. Ebell MH, Warbasse L, Brenner C. Evaluation of the dyspeptic patient: a cost-utility study. J Fam Pract 1997;44:545–55.[Web of Science][Medline]
  36. Sonnenberg A, Everhart JE. Health impact of peptic ulcer in the United States. Am J Gastroenterol 1997;92:614–20.[Web of Science][Medline]
  37. National Health Service England and Wales. Drug tariff, July edn. London: The Stationery Office, 2003.
  38. British Society of Gastroenterology Endoscopy Committee. Non-variceal upper gastrointestinal haemorrhage: guidelines. Gut 2002;51:iv1–iv6.[Free Full Text]
  39. Netten A, Curtis L. Unit costs of health and social care 2002. Canterbury: PSSRU, 2003.
  40. National Health Service Executive. NHS reference costs 2002 (http://www.dh.gov.uk/PublicationsandStatistics/Publications/PublicationspolicyandGuidancearticle/fs/en?CONTENTID=4069646&chk=kk52).
  41. Heyland D, Gafni A, Griffith L et al. The clinical and economic consequences of clinically important gastro-intestinal bleeding in critically ill patients. Clin Intens Care 1996;7:121–5.
  42. Gralnek IM, Jensen DM, Gornbein J et al. Clinical and economic outcomes of individuals with severe peptic ulcer hemorrhage and nonbleeding visible vessel: an analysis of two prospective clinical trials. Am J Gastroenterol 1998;93:2047–56.[Medline]
Submitted 22 July 2005; revised version accepted 1 November 2005.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
45/5/606    most recent
kei241v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Elliott, R. A.
Right arrow Articles by Symmons, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Elliott, R. A.
Right arrow Articles by Symmons, D.
Related Collections
Right arrow Health Economics
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?