SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

Background : It is unclear what the best therapeutic approach is in patients who require non-steroidal anti-inflammatory drugs. In clinical practice, choice of prescriptions are often based on drug costs.

Aim : To evaluate costs per upper gastrointestinal bleeding avoided with different prevention strategies.

Methods : Two major strategies have been considered (coxibs vs. non-steroidal anti-inflammatory drugs plus generic/brand gastroprotective agent). The number of patients needed to treat to prevent a bleeding event, the cost of the drug and duration of treatment were used to estimate costs.

Results : Based on hospitalization costs of a bleeding event, no therapeutic strategy is cost-effective in patients without risk factors. All strategies (including omeprazole + coxib) are cost-effective in patients with bleeding ulcer history. With other risk factors, all strategies are cost-effective but prevention of events is twice as expensive in patients <75 years of age. No strategy shows superiority unless the cheapest generics are prescribed or a 50% reduction in the incidence of lower gastrointestinal complications with coxibs is confirmed.

Conclusions : Current prevention strategies to reduce serious non-steroidal anti-inflammatory drug-associated gastrointestinal events are only cost-effective in patients with risk factors. No strategy shows superiority, but coxib strategy would be more cost-effective if it were associated with a reduction of events of the lower gastrointestinal tract.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

It is well-known that non-steroidal anti-inflammatory drugs (NSAIDs) are widely used worldwide. Their therapeutic use, far from declining, is expected to increase in a gradually ageing Western population.1 The main problem involved in the use of NSAIDs is the occurrence of side-effects, especially gastrointestinal (GI) side-effects, which can be serious and fatal. It is known that over 3% of the patients using NSAIDs for 1 year develop GI complications located both in the upper and lower GI tract.1–4 It has also been estimated that one death occurs for every 1200 patients taking NSAIDs for a period of 2 months.5 In our country and from another standpoint, it has been estimated that NSAIDs were associated with no less than 1200 deaths in 1998.6

In recent years, several approaches have been developed to reduce the toxicity associated with the use of NSAIDs. Today, these approaches are essentially limited to two; on the one hand, the administration of gastroprotective agents and NSAIDs, and on the other, the use of cyclo-oxygenase-2 (COX-2) selective inhibitors without gastroprotective agents.7 Multiple variables may influence the choice of one or the other strategy, including efficacy, safety, potential drug interactions, cost-benefit ratio, existence of other non-GI side-effects, etc. In the past 2 years, major studies and a great deal of information have been published and raised a substantial interest, debate, and even confusion among doctors treating patients with rheumatic diseases. This debate particularly includes the costs resulting from one or the other approach for health systems with an ever-increasing pharmaceutical expenditure. A recent study which estimated the costs of GI side-effects of NSAIDs to one of the national health systems in Europe concluded that the costs incurred by these side-effects were no <186 million € and that two-thirds of this expenditure was due to the prescription of gastroprotective agents.8 On the contrary, penetration in the market of COX-2 selective inhibitors has been limited, especially in Europe, since coxib prescription is more expensive than the alternative proton pump inhibitor (PPI) plus common NSAIDs, which are available as generic drugs in several European countries. The objective of this study was therefore to analyse both approaches from an economic viewpoint, valid in general for any situation, but more specifically focused on the particular case of European countries. The comparative tool has been the cost of preventing a GI-bleeding event, as this is a tool that can be easily quantified and is based on direct scientific evidence.

Definition of test variables

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

The GI side-effects associated with the use NSAIDs range from acute mucosal lesions to endoscopic ulcers, dyspepsia, complications and death.1 It is widely agreed that the most clinically relevant side-effects are the presence of GI complications and death.1, 9 Since the death rate is low and there are no adequate records about it, the presence of GI complications is considered to have the greatest relevance and impact and, accordingly, this will be the adverse effect addressed in the cost-benefit evaluation in this study.

Definitions of variables.  Gastrointestinal complications: Two settings have been assessed. Upper GI complications: These included upper GI bleeding, upper GI perforation, and pyloric stenosis because of peptic, non-neoplastic lesions, not related to portal hypertension. Gastrointestinal complications of the entire GI tract: These included the events defined as bleeding, perforation, or stenosis in the entire GI tract. This figure was obtained by adding the upper and lower events defined in the studies with adequate and reliable information. GI bleeding, (haematemesis, melaena or lower – rectal bleeding) leading to hospital admission were considered. GI perforation excluded perforation resulting from any inflammatory condition (e.g. appendicitis, diverticulitis, etc.). The presence of pyloric or duodenal stenosis of a peptic origin was accepted as a complication when it had been assessed together with all other complications in the studies. The qualification of an ‘upper’ or ‘lower’ origin was determined by the presence or absence of lesions responsible for the event using routine diagnostic techniques (endoscopy, radiology).

Risk factors for the development of complications in patients taking NSAIDs.  Based on widely accepted criteria, the predefined risk factors considered for this study were, a history of ulcer complication, a history of ulcer, age over 65 years, concomitant use of two NSAIDs, concomitant use of anticoagulants, concomitant use of aspirin, and concomitant use of corticoids.9 For other variables, such as antisecretory drugs, mucosal protecting drugs, NSAIDs, and coxibs, definitions previously reported were used.10

Search strategy

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

A specific search was made for clinical trials published in journals meeting the following criteria: (i) specified end-points for analysis consistent with the above definitions; (ii) a population over 18 years of age and required treatment with NSAIDs or coxibs for at least 4 weeks; (iii) a clear specification of the types of procedures; (iv) studies should be randomized, double-blind, and placebo- or comparator-controlled, and published in full in peer-reviewed journals; (v) the information available in the article allowed for absolute values to be obtained for each variable analysed.

The studies were identified on-line using Medline in the period ranging from January 1980 to June 2003, and the keywords used were the same as described in previous studies.10

Review method and tools used in the analysis

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

The articles identified in Medline meeting the required conditions were first selected. The data on analysis given in Tables, text, or Figures corresponding to the intention-to-treat analysis (if specified) were then selected. Otherwise, they had to be obtained from the available information. Among the different possibilities for cost calculation, it was decided to use the number of patients needed to treat to prevent an event (NNT), which is a variable widely accepted by the scientific community that allows specific figures to be estimated in order to prevent an event.11 The NNT variable for each study arm was obtained from the article itself and was calculated using the formula NNT = 1/absolute risk reduction with the treatment used.11, 12 For this purpose, the incidence of complications in each treatment branch had to be clearly specified in the studies. This method allows for an easy conversion of the NNT figure in terms of pharmaceutical expenditure required to prevent an event. The analysis has been carried out from the perspective of the Spanish Health System. Since, in most cases, this type of medication is used in patients with rheumatic conditions and older than 65; the co-pay for these patients is 0% in the system.

In order to estimate the costs of treatment, which are given in Euros (€), the daily costs of several combinations and doses were considered, always including the cheapest, the most expensive and the most common combinations for each possibility. This information was obtained from prices to the public of the preparations in Spain in the first 6 months of the year 2003. Finally, sensitivity studies covering the conditions of minimum and maximum incidence of potential events with each approach were also made.

Studies included in the economic analysis

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

Seven studies meeting the minimum requirements were selected (Table 1); three of these studies addressed the primary prevention of complications. The first of them13 compared an approach of standard NSAID doses + placebo vs. maximum doses of misoprostol and NSAIDs; the other two studies compared doses twofold higher than the maximum recommended dose for the treatment of rheumatic conditions of rofecoxib2 or celecoxib vs. standard doses of non-selective NSAIDs.14 Two additional studies also focused on the incidence of complications with variable doses of coxibs and non-selective NSAIDs,15, 16 but their primary objective was not an evaluation of GI complications. Finally, two additional studies17, 18 compared the strategies consisting of PPIs plus NSAIDs vs. standard doses of celecoxib or Helicobacter pylori eradication for secondary prevention of complications in high-risk patients with a prior history of upper GI bleeding.

Table 1.  Description of studies analysing different strategies for prevention/reduction of complications
Study referenceType of preventionType of patientsTreatments evaluatedEvent analysedMaximum follow-up time (months)
  1. * Complications include: upper GI bleeding of peptic origin, perforation of the upper GI tract and pyloric-antral stenosis. UGIB, upper GI bleeding; PU, peptic ulcer; NSAID, non-steroidal anti-inflammatory drugs; GI, gastrointestinal.

MUCOSA13Primary (n = 8843)Rheumatoid arthritis (RA) (>52 years)(a) NSAIDs (standard dose) vs. (b) NSAIDs (standard dose) + MPH 200 mcg/6–8 hComplications* 6
VIGOR2Primary (n = 8076)RA (>50 years)Naproxen 500 mg/12 h vs. Rofecoxib 50 mg/24 h(1) Symptomatic ulcer + complications (primary objective) (2) Complications (secondary objective)12
CLASS14Primary (n = 8059)Osteoarthritis/RA (27.5 and 27.3%) >18 years (>65 = 37.3 and 39.1%)Ibuprofen 800 mg/8 h; diclofenac 75 mg/12 h vs. Celecoxib 400 mg/12 h(1) Complications (primary objective) (2) Symptomatic ulcer + complications (secondary objective)12
Langman et al.15Primary (n = 5435)Osteoarthritis  Age: 38–94NSAIDs (standard dose) vs. rofecoxib 12.5–25 mg/12 h(1) Symptomatic ulcer + complications (primary objective) (2) Complications (secondary objective)12
Goldstein et al.16Primary (n = 11 008)Osteoarthritis/RA (37.5–33%) Age: 18–89 (58.9 ± 12 and 59.5 ± 12)NSAID (standard dose) vs. celecoxib 100–400 mg/day(1) Complications 3 (annualized data are provided)
Chan et al.17Secondary (n = 150)High risk (previous history of UGIB × PU)Naproxen 500 mg/12 h (Helicobacter pylori-negative) vs Naproxen 500 mg/ 12 h + omeprazole 20 mg/ day (Helicobacter pylori- positive)(1) Complications 6
Chan et al.18Secondary (n = 287)High risk (Helicobacter pylori- negative) (history of UGIB × PU)Diclofenac 75 mg/12 h +  omeprazole 20 mg/day vs. Celecoxib 200 mg/12 h(1) Complications (2) All GI complications and serious events 6

Prevention of an upper GI complication event. Table 2 shows the absolute frequencies of complications for each arm of the studies shown in Table 1, and calculations of NNTs for the strategies with the lowest complication rate. There are marked differences between the NNTs of primary prevention studies and secondary prevention studies (NNT = 125 in VIGOR and 263 in MUCOSA vs. seven in Chan's studies). As regard primary prevention studies, the least adequate estimates appear to be those reported in the studies by Langman et al.15 and Goldstein et al.,16 since although they are within the range seen in the other three studies, the studies had other primary objectives and did not show consistent designs. It should be noted that the data reported in the CLASS study14 are provided despite the fact that the overall differences between the arms were not significant. Differences occurred when comparing ibuprofen, but not diclofenac.

Table 2.  Number (of patients) needed to treat (NNT) with the different strategies for prevention of an upper gastrointestinal (GI) complication in the standard population studied in the trials
ReferenceComplication rate (%)Absolute risk reductionNNT
Treatment 1Treatment 2
  1. * FDA hearing. The overall differences between treatments were not significant.

  2. † Due to the identical design of this study with reference 162 of the same author, where a recurrence rate of 18.8% was seen in patients (Helicobacter pylori-negative) with a history of bleeding and placebo, this figure was used to estimate the NNT for celecoxib vs. placebo in patients with eradicated Helicobacter pylori.

  3. ‡ This figure is similar to that estimated for the diclofenac + omeprazole combination.

  4. NSAID, non-steroidal anti-inflammatory drugs.

MUCOSA13NSAID + placebo = 0.94/6 monthsNSAID + misoprostol = 0.56/6 months0.0038/6 months 263
VIGOR2Naproxen = 1.4/12 monthsRofecoxib = 0.6/12 months0.008/12 months125
CLASS*14NSAID global = 0.8/12 months Ibuprofen = 1.14/12 months Diclofenac = 0.48/12 monthsCelecoxib = 0.4/12 months0.004/ 12 months 0.007/12 months 0.0008/12 months250 143 1250
Langman et al.15NSAIDs: 0.7/12 monthsRofecoxib = 0.18/12 months0.0052/12 months192
Goldstein et al.16NSAID = 0.33/3 months 1.68% (annualized)Celecoxib = 0.03%/3 months 0.20% (annualized)0.003/ 3 months 0.0148/12 months333 68
Chan et al.17Naproxen = 18.8/6 monthsNaproxen + omeprazole = 4.4/6 months0.144/12 months7
Chan et al.18Diclofenac + omeprazole = 6.4/6 monthsCelecoxib = 4.9/6 months0.007/6 months7‡

Prevention of complication events in the entire GI tract.  There are no studies specifically designed to obtain information in this regard, but the studies shown in Table 1 do provide adequate information because all potential complications of the GI tract were routinely recorded. Furthermore, lower GI events from the VIGOR study have been recently published.2 The data in Table 3 give the estimated frequencies of complications in the entire GI tract reported in the studies where this information could be reliably obtained. From this standpoint, the NNTs among primary prevention studies were consistent; the NNTs estimated for studies with coxibs were lower than the NNT seen in the MUCOSA study with misoprostol (77 and 65 for coxibs vs. 227 for misoprostol).13

Table 3.  Estimates of number needed to treat (NNT) with different strategies for the prevention of a gastrointestinal (GI) complication of the entire GI tract
ReferenceType of patientRate (%)Absolute risk reduction (months)NNT
Treatment 1Treatment 2
  1. * The rates are obtained from the data provided in the publication. The events reported by the investigators but not adjudicated by the committee as events occurring in the upper GI tract have been counted. The quality of this information is unknown, since the study was not designed to evaluate this aspect.

  2. † This is the only study formally reporting data related to complications of the lower GI tract. However, unlike events in the upper GI tract, events in the lower GI tract were not adjudicated specifically by a committee but by the authors of the study.

  3. ‡ The rates are obtained from data provided in the publication. The events that were reported by the investigators but not considered by the adjudication committee as events occurring in the upper GI tract have been counted. For this analysis, the events described as ‘non-ulcer bleeding’, ‘anaemia’, and ‘colonic or small bowel disease’ have been included. The quality of these diagnoses is also unknown, since the study was not designed to evaluate this aspect.

  4. § The rates are obtained from the data provided in the publication, which are clearly specified as reported as complication but not adjudicated as upper GI events. The differences in rates in the study were not significant because the study was not designed to evaluate this aspect. Considering the design of Chan's studies, they can be taken as reference of the complication rate.

  5. ¶ NNT vs. the diclofenac + omeprazole combination (non-significant incidence differences).

  6. ** Estimate vs. placebo in patients not infected by Helicobacter pylori or with eradicated infection (data estimated from references 22 and 23).

  7. NSAID, non-steroidal anti-inflammatory drugs.

MUCOSA*13Rheumatoid arthritis (>52 years)NSAID + placebo = 1.99/6 monthsNSAID + misoprostol  = 1.55/6 months0.0044/6227
VIGOR2Rheumatoid arthritis (>50 years)Naproxen = 2.3/12 monthsRofecoxib = 1.0/12 months0.013/1277
CLASS14Osteoarthritis/RA (>18 years)NSAID global = 4.04./12 monthsCelecoxib = 2.49/12 months0.015/1265
Chan et al.§18High riskDiclof + omeprazole = 10.5/6 monthsCelecoxib = 6.2/6 months0.043/6 0.166/6** 23¶ 6**

Prevention of upper GI events in patients with different risk factors.  Only two studies provide adequate information for calculating the NNTs in patients with different risk factors, and do so from different perspectives. Specific information has been provided from the VIGOR study which refers to complications and symptomatic ulcers.19 On the contrary, relevant information dependent on the presence of risk factors has been provided from the MUCOSA study,20 but encompassed all events (upper and lower) reported. The risk factor with the lowest NNT is a prior history of ulcer bleeding, ulcer history and age above 75 years (Table 4). The NNTs reported for the age group from 65 to 75 years are twofold higher than those of patients aged >75 years.

Table 4.  Number needed to treat (NNTs) for different strategies for prevention of an upper gasctrointestinal (GI) event in different situations in patients at risk
Risk factorStudyType of eventRate (%)NNT
Treatment 1Treatment 2
  1. NSAID, non-steroidal anti-inflammatory drugs.

Age <65 yearsLaine et al.3Symptomatic ulcer + complicationsNaproxen 1000 mg/day 3.15/12 monthsRofecoxib 50 mg/day 1.64/12 months66
Simon et al.20All reported GI eventsNSAIDs 3.53/6 monthsMisoprostol 2.70/6 months120
Age 65–74Laine et al.3Symptomatic ulcer + complicationsNaproxen 1000 mg/day 7.37/12 monthsRofecoxib 50 mg/day 3.33/12 months25
Age >75 yearsLaine et al.3Symptomatic ulcer + complicationsNaproxen 1000 mg/day 14.46/12 monthsRofecoxib 50 mg/day 4.51/12 months10
Simon et al.20All reported GI eventsNSAIDs 4.97/6 monthsMisoprostol 3.74/6 months81
History of ulcerLaine et al.3Symptomatic ulcer + complicationsNaproxen 1000 mg/day 13.54Rofecoxib 50 mg/day 5.2412
Simon et al.20All reported GI eventsNSAIDs 7.84/6 monthsMisoprostol 3.73/6 months24
History of bleedingLaine et al.3Symptomatic ulcer + complicationsNaproxen 1000 mg/day 18.84Rofecoxib 50 mg/day 10.0011
Simon et al.20All reported GI eventsNSAIDs 8.9/6 monthsMisoprostol 4.45/6 months23

Prevention of an upper GI complication event.  The estimates of the cost to prevent an upper GI ulcer complication event due to a peptic ulcer is shown in Table 6 are derived from the NNTs reported in Table 1 and from the prices of the drugs in Table 5. There were no specific studies looking at the effect of PPIs in primary prevention of upper GI complication events; however, a reduction in the incidence of upper GI complications similar to that obtained with misoprostol has been considered, which is also supported by data from epidemiological studies.21

Table 6.  Estimated costs of treatment with the different strategies for preventing an upper gastrointestinal (GI) tract complication and a complication of the entire GI tract, depending on the different co-prescription levels
TreatmentType of patientCost of preventing an upper GI event (€) (range)Treatment period (months)Cost of preventing a GI event (€) (range)Treatment period (months)
  1. * High risk includes the presence of a history of ulcer with bleeding.

  2. † Includes the cheapest generic proton pump inhibitor (PPI) in the market.

  3. ‡ Estimates based on an additional 50% reduction with PPIs.

  4. § Based on VIGOR (high dose).

  5. ¶ Minimum cost based on annualized frequency in the Goldstein et al.16 study (standard dose). Intermediate and maximum cost based on CLASS for ibuprofen and global non-steroidal anti-inflammatory drug (NSAID) (high-dose celecoxib).

  6. ** Estimate based on obtaining results for rofecoxib similar to those reported for celecoxib in high-risk patients.

  7. †† Estimate based on an additional 50% reduction with PPI from figures in Chan's studies (NNT = 4).

NSAID (generic) + misoprostolStandard66 749–78 111657 612–67 419 6
High risk*6091–71286No studies –
NSAID (generic) + PPI (generic)Standard‡51 127–71 483644 128–61 698 6
High risk*1360–190262376–3261 6
NSAID + PPIStandard‡ 6  6
 Minimum cost†23 670–99 414 20 430–85 806 
 Maximum cost95 153–106 515 82 128–114 816 
 High risk* 6  6
 Minimum cost†630–2646 1080–4536 
 Maximum cost2205–2835 4384–6048 No specific studies‡ 
CoxibStandardRofecoxib§12Rofecoxib12
 12.5 mg: 59 768  36 817 
 25 mg: 73 000  44 968 
Celecoxib¶ Celecoxib 
 200 mg: 33 507–70463- 123 187  32 028 
 400 mg: 67014- 246 375  64 056 
High risk*Rofecoxib**6Rofecoxib 
 12.5 mg: 1650  1257 
 25 mg: 2016  1728 
Celecoxib Celecoxib 
 200 mg: 1701  1458 
 400 mg: 3402  2916 
Coxib + PPIHigh risk*1548–1756††6  
Table 5.  Daily costs of drug treatment with different strategies based on prices of the first 6 months of 2003 that have been used to estimate the costs of preventing a gastrointestinal (GI) bleeding event with the different strategies
Drug 1 (non-steroidal anti-inflammatory drugs, NSAID)Drug 2 (anti-ulcer agent) (mg/day) Cost/day (€)Tablet/day
 1. Rofecoxib 25 mg/day (most common dose)1.601
 2. Rofecoxib 12.5 mg/day1.311
 3. Celecoxib 200 mg/day (most common dose)1.351
 4. Celecoxib 400 mg (2 × 200)2.702
Generics (mg/day)
 5. Diclofenac generic (150)Omeprazole generic (20)1.114
 6. Ibuprofen generic (2400)Omeprazole generic (20)1.355
 7. Naproxen generic (1000)Omeprazole generic (20)1.193
 8. Aceclofenac (200)Omeprazole generic (20)1.513
 9. Piroxicam generic (20)Omeprazole generic (20)1.082
Generics + minimum price of PPI (proton pump inhibitor, mg/day)
 10. Diclofenac generic (150)Omeprazole (20)0.54
 11. Ibuprofen generic (2400)Omeprazole (20)0.745
 12. Diclofenac generic (150)Lansoprazole (30)0.54
 13. Ibuprofen generic (2400)Lansoprazole (30)0.55
 14. Diclofenac generic (150)Pantoprazole 40)1.864
 15. Ibuprofen generic (2400)Pantoprazole (40)2.15
Generics + maximum price PPI/misoprostol (mg/day)
 16. Diclofenac generic (150)Omeprazole (20)2.244
 17. Ibuprofen generic (2400)Omeprazole (20)2.815
 18. Diclofenac generic (150)Lansoprazole (30)2.014
 19. Ibuprofen generic (2400)Lansoprazole (30)2.255
 20. Diclofenac generic (150)Pantoprazole (40)2.034
 21. Ibuprofen generic (2400)Pantoprazole (40)2.275
 22. Diclofenac generic (150)Rabeprazole (20)2.014
 23. Ibuprofen generic (2400)Rabeprazole (20)2.255
 24. Diclofenac generic (150)Esomeprazol (20)1.754
 25. Ibuprofen generic (2400)Esomeprazol (20)1.995
 26. Diclofenac generic (150)Misoprostol 800 mcg/day1.417
 26. Ibuprofen generic (2400)Misoprostol 800 mcg/day1.658
Coxib + PPI combinations (mg/day)
 27 Rofecoxib (25)Omeprazole generic (20)2.442
 28. Rofecoxib (12.5)Omeprazole generic (20)2.152
 29. Celecoxib (200)Omeprazole generic (20)2.192
 30. Celecoxib (400)Omeprazole generic (20)3.543

Taking as reference the hospital costs of treating a GI bleeding event in the same country, estimated at approximately 2786€ in 2003 (data provided by IASIST; http://www.solucient.com), it may be observed that the cost of preventing an upper GI complication event with all strategies is far higher than the cost of the treatment in the patients with no risk factors. These costs are similar for all strategies and are only reduced by 50% if the cheapest generic of the market is used. However, even in this situation the cost of preventing one event is approximately 10 times higher than the cost required for treating the event.

The opposite is seen when the cost of preventing an upper GI complication event in the high-risk patient (history of GI bleeding) is estimated. All approaches, including the combination of a PPI with a coxib, are cost-effective, since the costs of preventing an upper complication event range from 630€ with the cheapest generic PPI of the market to 3402€ with the highest doses of celecoxib.

Prevention of a complication from the entire GI tract.  Based on the NNT data given in Table 3, the costs of preventing a complication event in any part of the GI tract are significantly reduced for the strategies including prescription of a coxib (Table 6). In patients with standard risk, the reduction of costs of preventing a GI complication event would be between 38.4 and 54.5% with the most common doses of rofecoxib and celecoxib (vs. intermediate cost in the upper GI tract) respectively. In the high-risk patient, costs per GI complication event avoided drop from 2016€ to 1728€ and from 1701€ to 1458€ with most common doses of rofecoxib and celecoxib respectively.

The cost of preventing an event in the entire GI tract with coxibs is a figure similar to the one estimated for the cheapest PPI in prevention of an upper GI complication event (Table 6). In patients at risk, the cost of preventing a GI complication event is generally lower with rofecoxib and celecoxib than with the other options.

Prevention of upper GI events in patients with different risk factors.  The inclusion of symptomatic ulcers as clinically significant events markedly reduces the costs of preventing an event (Table 7). Thus, in patients aged <65 years, the costs are decreased with most approaches to figures around 30 000€. These figures gradually decrease for all strategies from that age, the cost being approximately three times lower in the age range from 65 to 75 years, and six times lower above 75 years. All strategies are cost-effective in patients older than 75 years, patients with ulcer history or ulcer bleeding history. Costs are lower when the cheapest PPI in the market is prescribed.

Table 7.  Costs of preventing a clinically significant upper gastrointestinal (GI) event (symptomatic ulcers + complications) depending on the risk factor (age, history of ulcer or bleeding)
Risk factorRange (minimum–maximum)NSAID + PPI
Rofecoxib* Celecoxib†NSAID + misoprostolNSAID + PPI genericMinimum price PPIMaximum price PPI
  1. Calculations are based on Table 4, assuming the same effect on event reduction among the different strategies. Data provided in €.

  2. * Doses of 12.5 and 25 mg/day; † doses of 200 and 400 mg/day; ‡ includes the cheapest generic in the market.

  3. The most common doses of rofecoxib and celecoxib sold in the market are the 25 mg/day and 200 mg/day doses respectively.

  4. NSAID, non-steroidal anti-inflammatory drugs; PPI, proto pump inhibitor.

Age <65 years31 557–38 54432 521–87 80633 967–50 58926 741–36 37612 405–50 589‡42 157–67 692
Age 65–7411 953–14 60012 318–24 63712 866–15 05610 128–13 7784562–19 16215 968–25 641
Age >75 years4781–58404927–98555146–60224051–55111825–55116387–10 256
History of ulcer4380–57375913–11 8264862–66144861–66132190–91987665–12 307
History of ulcer bleeding5259–64245420–10 8405661–66244456–60622007–84317026–11 282

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References

The economic perspective of this study suggests and confirms that current approaches for the prevention of NSAID-induced upper GI complications (addition of PPI or misoprostol vs. prescription of COX-2 selective inhibitors) are only cost-effective in patients with risk factors. The data also suggest that no strategy seems clearly superior to the others, and that the strategies are particularly cost-effective in patients older than 75 years, patients with an ulcer history and, particularly, patient with a history of GI bleeding, in whom even the coxib + PPI combination is cost-effective. In addition, the study provides relevant data and new perspectives that should be considered.

The first issue refers to the option of prescribing the cheapest PPI in the market (omeprazole) together with a generic non-selective NSAID. With this approach, the cost of preventing an upper GI complication event is lower than with the other options. However, the reduction of costs is not great enough to recommend extension of this approach to patients with no risk factors, and does not prevent other alternatives from being also cost-effective in patients with risk factors. Similarly, as the NNT value is significantly reduced in patients with risk factors, the price differences between drugs are not sufficient to cause cost differences resulting in a clear shift of the balance in favour of one of the options. This is so even when the most commonly prescribed generic option, generic PPI +generic NSAID is included. It should also be noted that the cheapest option may not be the most realistic or adequate, as the generics market is changing, the cheapest generic may not be in a position to appropriately supply the market and its efficacy has been challenged.22, 23 Moreover, when compared with the alternative option of COX-2 selective inhibitors, the presence in the market of PPIs other than omeprazole with an increased capacity to inhibit gastric secretion and with a proven capacity to prevent lesions induced by NSAIDs,24–26 warrants consideration of the NSAID + gastroprotective agent option in this setting of diversity in PPI prescription.

The other relevant aspect is the fact that complications in the GI tract associated with the use of NSAIDs occur throughout the GI tract.3, 4 Complications in the lower GI tract can account for 30–50% of all GI complications (excluding those due to portal hypertension). There is evidence from healthy volunteers that, unlike non-selective NSAIDs, coxibs are not associated with intestinal damage.27 The clinical data available suggest that the use of rofecoxib is associated with a 50% reduction in clinically significant events (mainly lower GI bleeding events) in patients with rheumatoid arthritis (RA).3 Although no specific data have been reported with celecoxib for this event, the analysis of the published information summarized in this study suggests a reduction of a magnitude similar to rofecoxib. From this wider perspective, and if these data related to the lower GI tract are confirmed, the coxib option reduces further the cost of preventing an episode of bleeding, which does not occur with the NSAID + gastroprotective agent option. However, as with the option of the cheapest PPI in the market, the cost reduction does not appear to be sufficient to make it cost-effective in patients with no risk factors.

The study data reported here are relevant for the prescribing doctors in so much as they analyse the costs of preventing a GI complication event induced by NSAIDs. The use of NNT as an analysis tool reinforces this suggestion, which is much more direct, next to the evidence, and easily understandable by the clinical doctors. Furthermore, the easy transfer of these figures to different price scenarios makes it attractive, because the data allow for re-analysing strategies and keeping them continuously updated.

However, it should also be noted that this approach involves a number of problems and has limitations that should be recognized in order to put the data and conclusions given here into context. First, the paucity of the available data and the heterogeneity of the studies analysed suggests that some of the data and conclusions should be taken with caution (e.g. when extrapolations on the costs of lower GI events from the post hoc analysis of VIGOR and CLASS are considered). This makes comparisons between different strategies and different scenarios even more difficult, as therapies and patients were different in the studies. Under this perspective it must be outlined that the goal of the study was not to compare and obtain rapid economic conclusions from a heterogeneous number of studies, but to summarize the evidence available in order to have the best estimate in one aspect (cost of preventing a complication event in different type of patients using NSAIDs) of a complex issue which is the cost-benefits of different strategies. Therefore, the study has not evaluated all potential clinical scenarios and some of the estimates were based on indirect evidence (e.g. effect of PPI in primary prevention)14, 21 that might change if studies directly answering the proposed question are published. Another assumption was that all PPIs have the same efficacy in their effect on NSAID-induced injuries. There is no evidence for rejecting the opposite, but the multiple generics available and their dispensing in different vials and dosage forms may cause their efficacy to be different, particularly as regard excipient degradation over time, which can affect the effective dose released.22, 23

Another limitation of the study is that results are based on drug prices and cost from the prices of a single country in Europe, which may not be easy to extrapolate to other countries or scenarios, especially if the Health System is very different from those existing in Europe and Canada. A further issue to be considered when assessing the data is that NNTs were estimated in some of the studies based on high doses of coxibs (two- or fourfold higher than those usually prescribed) compared with standard anti-inflammatory doses of non-selective NSAIDs.2, 14 This may have played a role against coxibs, as at the lower doses used in practice clinical, they appear to have an even better safety profile.28 Finally, this study was performed from the perspective of GI safety, and its conclusions should be considered from that viewpoint. The adverse effects of NSAIDs and coxibs can also occur in other systems and organs, such as the kidney and the cardiovascular system.2. An overall assessment of all of them may lead to changing the conclusions of this study, but we must not loose the perspective that the most common side-effects are GI effects. The data currently available suggest that, at standard doses, both types of drugs have a similar frequency of non-GI side-effects,29 but new studies are expected to dispel the question as to whether coxibs are associated with a higher risk of cardiovascular events.

Other considerations not addressed in this study refer to the degree of compliance with one or the other approach, which in clinical practice may affect costs and the frequency of adverse effects. In this regard, the higher the number of tablets, the lower the compliance. This effect may be particularly relevant for the combination NSAID + PPI or misoprostol.30 Another aspects that may affect costs is the presence of dyspepsia which may lead to prescription of PPI with either NSAIDS or coxibs.9, 25

Finally, it must be noted that, despite the high frequencies of co-prescription (NSAID-gastroprotective agent) in the setting where the data were obtained (close to 50%), the proportion of patients with risk factors requiring NSAIDs and not receiving preventive therapy may be as high as 80%, which suggests that current clinical practice in prevention strategies is still far from being appropriate and not cost-effective.26, 31 Thus, it appears natural to think that the first action required for containing drug expenditure should be focused on the promotion of adequate prescription for the target population with risk factors rather than on restriction of the use of some drugs for apparent price reasons.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Definition of test variables
  6. Search strategy
  7. Review method and tools used in the analysis
  8. Results
  9. Studies included in the economic analysis
  10. NNTs in the prevention of complications with the different therapeutic strategies
  11. Cost to prevent a GI complication event with the different prevention strategies
  12. Discussion
  13. Acknowledgements
  14. References
  • 1
    Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of nonsteroidal anti-inflammatory drugs. N Engl J Med 1999; 340: 188899.
  • 2
    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: 15208.
  • 3
    Laine L, Connors LG, Reicin A, et al. Serious lower gastrointestinal clinical events with nonselective NSAID or coxib use. Gastroenterology 2003; 124: 28892.
  • 4
    Lanas A, Sekar MC, Hirschowitz BI. Objective evidence of aspirin use in both ulcer and non-ulcer upper and lower gastrointestinal bleeding. Gastroenterology 1992; 103: 8629.
  • 5
    Tramer MR, Moore RA, Reynolds DJ, McQuay HJ. Quantitative estimation of rare side effects which follow a biological progression: a new model applied to chronic NSAID use. Pain 2000; 85: 16982.
  • 6
    Lanas A. Clinical experience with cyclooxygenase-2 inhibitors. Rheumatology (Oxford) 2002; 41(Suppl. 1): 1622; discussion 35–42.
  • 7
    Lanas A, Martin-Mola E, Ponce J, Navarro F, Pique JM, Blanco FJ. Clinical strategy to prevent the gastrointestinal adverse effects of nonsteroidal anti-inflammatory agents. Gastroenterol Hepatol 2003; 26: 485502.
  • 8
    Lanas A. Impacto económico de los efectos secundarios gastrointestinales asociados a antiinflamatorios no esteroideos en el Servicio Nacional de Salud. Med Clin (Barc) 2000; 114(Suppl. 3): 4653.
  • 9
    Lanas A, Hirschowitz BI. Toxicity of NSAIDs in the stomach and duodenum. Eur J Gastroenterol Hepatol 1999; 11: 37581.
  • 10
    Lanas A, Arroyo MT. Antisecretory drugs and prophylaxis of secondary gastrointestinal effects associated with the consumption of non steroid anti-inflammatory drugs. Med Clin (Barc) 2000; 114(Suppl. 2): 5762.
  • 11
    Schoenfeld P. An evidence-based approach to the gastrointestinal safety profile of COX-2-selective anti-inflammatories. Gastroenterol Clin North Am 2001; 30: 102744.
  • 12
    Thompson AJ, Yeomans ND. COX-2 inhibition versus gastroprotection with dual COX inhibitors: an evidence-based approach. Curr Pharm Des 2003; 9: 22218.
  • 13
    Silverstein FE, Graham DY, Senior JR, et al. Misoprostol reduces serious gastrointestinal complications in patients with rheumatoid arthritis receiving nonsteroidal anti-inflammatory drugs. A randomized, double-blind, placebo-controlled trial. Ann Intern Med 1995; 123: 2419.
  • 14
    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. JAMA 2000; 284: 124755.
  • 15
    Langman MJ, Jensen DM, Watson DJ, et al. Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs. JAMA 1999; 282: 192933.
  • 16
    Goldstein JL, Silverstein FE, Agrawal NM, et al. Reduced risk of upper gastrointestinal ulcer complications with celecoxib, a novel COX-2 inhibitor. Am J Gastroenterol 2000; 95: 168190.
    Direct Link:
  • 17
    Chan FK, Chung SC, Suen BY, et al. Preventing recurrent upper gastrointestinal bleeding in patients with Helicobacter pylori infection who are taking low-dose aspirin or naproxen. N Engl J Med 2001; 344: 96773.
  • 18
    Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac and omeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. N Engl J Med 2002; 347: 210410.
  • 19
    Laine L, Bombardier C, Hawkey CJ, et al. Stratifying the risk of NSAID-related upper gastrointestinal clinical events: results of a double-blind outcomes study in patients with rheumatoid arthritis. Gastroenterology 2002; 123: 100612.
  • 20
    Simon LS, Hatoum HT, Bittman RM, Archambault WT, Polisson RP. Risk factors for serious nonsteroidal-induced gastrointestinal complications: regression analysis of the MUCOSA trial. Fam Med 1996; 28: 20410.
  • 21
    Lanas A, Bajador E, Serrano P, et al. Nitrovasodilators, low-dose aspirin, other nonsteroidal antiinflammatory drugs, and the risk of upper gastrointestinal bleeding. N Engl J Med 2000; 343: 8349.
  • 22
    Navarro MA, Raei N, Torres F, et al. Differences in the release of omeprazole in 4 commercial preparations: influence of pH and ionic concentration. Gastroenterol Hepatol 1998; 21: 6370.
  • 23
    Palummo M, Cingolani A, Dall L, Volonte MG. Stability of capsules containing omeprazole in enteric coated pellets. Boll Chim Farm 2000; 139: 1248.
  • 24
    Graham DY, Agrawal NM, Campbell DR, et al. Ulcer prevention in long-term users of non-steroidal anti-inflammatory drugs. Arch Intern Med 2002; 161: 16975.
  • 25
    Yeomans H, Hawkey C, Jones R, et al. Esomeprazole provides effective control of NSAID-associated upper GI symptoms in patients continuing to take NSAIDs. Gastroenterology 2003; 124: A-107 (Abstract 797).
  • 26
    Kimmey M, Lanas A. Review article: Appropriate use of proton pump inhibitors with traditional nonsteroidal anti-inflammatory drugs and COX-2 selective inhibitors. Aliment Pharmacol Ther 2004; 19(Suppl. 1): 605.
  • 27
    Hunt RH, Bowen B, Mortensen ER, et al. A randomized trial measuring fecal blood loss after treatment with rofecoxib, ibuprofen, or placebo in healthy subjects. Am J Med 2000; 109: 2016.
  • 28
    Mamdani M, Rochon PA, Juurlink DN, et al. Observational study of upper gastrointestinal haemorrhage in elderly patients given selective cyclo-oxygenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs. BMJ 2002; 325: 62430.
  • 29
    Schwartz JI, Vandormael K, Malice MP, et al. Comparison of rofecoxib, celecoxib, and naproxen on renal function in elderly subjects receiving a normal-salt diet. Clin Pharmacol Ther 2002; 72: 5061.
  • 30
    Sturkenboom MC, 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 Ther 2003; 18: 113748.
  • 31
    Sturkenboom MC, Burke TA, Dieleman JP, Tangelder MJ, Lee F, Goldstein JL. Underutilization of preventive strategies in patients receiving NSAIDs. Rheumatology (Oxford) 2003; 42(Suppl. 3): iii2331.