Rheumatology Advance Access originally published online on June 15, 2006
Rheumatology 2007 46(1):135-140; doi:10.1093/rheumatology/kel195
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Celecoxib compared with sustained-release paracetamol for osteoarthritis: a series of n-of-1 trials
1School of Medicine, Logan campus, Griffith University, University Drive, Meadowbrook, 2Discipline of General Practice, 3N-of-1 trial service, Discipline of General Practice, The University of Queensland, Herston, Queensland, 4Faculty of Health Science and Medicine, Bond University, Gold Coast, Queensland, Australia, 5Faculty of Health and Environmental Sciences, Auckland University of Technology, Auckland, New Zealand and 6School of Medicine, The University of Queensland, Herston, Queensland, Australia
Correspondence to: Prof. Michael J. Yelland, Associate Professor of Primary Health Care, School of Medicine, Logan campus, Griffith University, University Drive, Meadowbrook Qld 4131, Australia. E-mail: m.yelland{at}griffith.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Objective. To assess the use of n-of-1 trials for short-term choice of drugs for osteoarthritis, with particular reference to comparing the efficacy of sustained-release [SR] paracetamol with celecoxib in individual patients.
Methods. Evaluation of community-based patients undergoing n-of-1 trials which consisted of double-blind, crossover comparisons of celecoxib 200 or 400 mg/day with sustained-release paracetamol 1330 mg three times a day in three pairs of 2 week treatment periods per drug with random order of the drugs within pairs. Outcomes evaluated were pain and stiffness in sites nominated by the patient, functional limitation scores, preferred medication, side effects and changes in drug use after an n-of-1 trial. Participants were 59 patients with osteoarthritis in multiple sites (hip 6, knee 24, hand 6, shoulder/neck 8, back 14, foot 5), with pain for 
1 month severe enough to warrant consideration of long-term use of celecoxib but for whom there was doubt about its efficacy. Forty-one n-of-1 trials were completed.
Results. Although on average, celecoxib showed better scores than SR paracetamol [0.2 (0.1) for pain, 0.3 (0.1) for stiffness and 0.3 (0.1) for functional limitation], 33 of the 41 individual patients (80%) failed to identify the differences between SR paracetamol and celecoxib in terms of overall symptom relief. Of the eight patients who were able to identify the differences, seven had better relief with celecoxib and one with SR paracetamol. In 25 out of 41 [61%] patients, subsequent management was consistent with their trial results.
Conclusions. N-of-1 trials may provide a rational and effective method to best choose drugs for individuals with osteoarthritis. SR paracetamol is more useful than celecoxib for most patients of whom management is uncertain.
KEY WORDS: n-of-1 trials, osteoarthritis, celecoxib, SR paracetamol
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Osteoarthritis (OA) is one of the 10 most disabling diseases in the developed world [1], inflicting joint pain and stiffness among 10% of men and 20% of women aged 45–60 yrs in the West [2]. This morbidity presents a significant healthcare burden and comes at enormous cost, mostly for analgesic and anti-inflammatory drugs. Paracetamol, relatively inexpensive and safe, is the agent of first choice of advisory guidelines based on good evidence from trials [3–5]. But non-steroidal anti-inflammatory drugs (NSAIDs) are better for some individuals, especially for moderate-to-severe OA pain [4] or where the pain is unresponsive to paracetamol [3].
One solution for the safety issue seemed to be cyclooxygenase enzyme, subtype 2 (COX-2) specific NSAIDs, which were increasingly used until 2004 when the reports of elevated risk of cardiovascular events changed this perception [6, 7].
Paracetamol has the disadvantage of requiring 4 doses/day to maintain therapeutic serum levels. The recent introduction of sustained-release (SR) paracetamol has reduced this requirement to 3 doses/day [8, 9].
N-of-1 trials provide empirical data of individual responses to the treatment. These are within-patient randomized, double-blind, crossover trials, in which patients act as their own controls, and provide the most rigorous information available for any individual patient [10–14]. In the n-of-1 trial, the unit of randomization is the treatment sequence for an individual patient, and a single n-of-1 treatment cycle includes an exposure to each therapy. Data are usually analysed for individual patients. In contrast, in randomized crossover trials, where the individual is randomized to one group or another, each participant receives each intervention at different time frames of the study, and the data are analysed for each group as a whole. The question answered by an n-of-1 trial is ‘which therapy is better for this patient?’ and in classic crossover designs, ‘in population x, which drug is better?’
There are no n-of-1 trials comparing paracetamol and celecoxib, and only three published reports of randomized controlled trials comparing paracetamol and celecoxib. For knee or hip OA, celecoxib was more efficacious than paracetamol [15]. For knee OA, celecoxib showed significantly greater efficacy [16] and faster onset of efficacy [17] than paracetamol.
Our main hypothesis was that for each individual patient there was no difference between the two medications. We also hypothesized that the results of n-of-1 trials influenced drug use in the short-term for patients with OA, with particular reference to the efficacy of SR paracetamol compared with celecoxib in individual patients.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
We offered an n-of-1 trial service for celecoxib (Pfizer, PO Box 57, West Ryde NSW, Australia 2114) [10–12] compared with SR paracetamol [GlaxoSmithKline, 82 Hughes Avenue, Locked Bag 3, Ermington, NSW 2115] throughout Australia between December 2003 and December 2004, communicating as we have previously described [13, 18, 19] by post, telephone, fax and e-mail. The process was similar to requesting a pathology test: we sent packs of test medications by post to patients on request from their family physicians; patients completed a daily symptom diary; and we followed-up patients by telephone, while the clinician continued to provide usual clinical care. At the end of the n-of-1 trial, diaries were analysed and a report sent to the doctor within 2 weeks. When the patient next consulted their doctor, the results were available to inform management decisions.
Recruitment
Recruitment was through a network of participating doctors and a print media and radio publicity campaign. Potential patients were able to contact our service directly, and information packs which we sent out could be taken to the doctor for them to request an n-of-1 trial.
Subjects
Eligibility was restricted to adults providing written informed consent with a clinical diagnosis of OA pain for at least 1 month (in the opinion of their attending general practitioner) of sufficient severity to consider long-term use of anti-inflammatory drugs or SR paracetamol. Contra-indications to either of these, or sulphas, concomitant disease (such as peptic ulcer, hepatic or renal dysfunction) which increases the risk of side effects, and depot corticosteroid injection in the last two months, were exclusions. Subjects did not need to have radiographic OA.
Intervention
Patients took either SR paracetamol [as 2 x 665 mg tablets (i.e. 1.33 g) 3 times a day] or celecoxib [200 mg daily, or 200 mg twice a day for those who were already using this dose], and a placebo identical to the alternative drug. The celecoxib and its placebo were encapsulated; the paracetamol placebo was manufactured. There were three cycles of paired treatment periods (2 weeks for each treatment and a total of 12 weeks). The order of the drugs in each pair was randomly assigned using a computer-generated schedule. Patients, doctors and the research assistant were blinded to medication order. We sent the drugs to participants fortnightly in pre-prepared blister packs. We recommended that they request tramadol (the dose was prescribed by their general practitioner) for additional analgesia as required.
Ethics approval for this study was provided by The University of Queensland Human Research Ethics Committee. The subjects’ written consent was obtained according to the Declaration of Helsinki [20].
At commencement, patients provided demographic information and a drug history; and recorded side effects weekly. At the end of each treatment period they guessed which treatment they had received. After the n-of-1 trial, patients were interviewed by telephone about subsequent management decisions.
Outcome measures
There were five primary outcomes to which we decide to give equal weight: pain, stiffness and functional limitation scores, medication preference and adverse effects.
We used pain/stiffness intensity rating scales to assess pain and stiffness scores daily with visual analogue scales marked 0–10 [21, 22]. We omitted the first week of data from each period to negate any carry-over effects [the effects of celecoxib wear off after multiple dosing within 5 days (10 half-lives)]. To assess functional limitation, we used the patient-specific functional scale [23] with up to five patient-nominated functions on a 0–10 visual analogue scale. Differences in mean scores between treatments were analysed using hierarchical Bayesian random effects models [24]. The Bayesian method which we employed allows for individual assessment and group meta-analysis. Assuming a minimum detectable difference in pain and stiffness scores of 1.0 [25], a definite response was defined as an adjusted mean absolute difference 1.0, a probable response as a difference of 0.5 but <1.0, and all other responses as no difference. Assuming a minimum detectable difference in functional limitation scores of 2.0 [26], a definite response was defined as an adjusted mean absolute difference 2.0, a probable response as a difference of 1.0 but <2.0, and all other responses as no difference.
At the end of each cycle, we assessed the medication preference. A definite response was defined as a preference for one medication in all three cycles, a probable response as a preference in two cycles and no response as a preference in no cycle or one cycle. We assessed adverse events weekly in each treatment period. Here, a ‘definite response’ was defined as fewer events on one medication in all three cycles, a ‘probable response’ as fewer events in two cycles and ‘no response’ as fewer events in no cycles or one cycle. All patients who completed the trial took at least 96% of their tablets.
Statistical analysis
To describe the overall response, we created an aggregate response variable, composed from an equally weighted linear combination of the five variables (each arbitrarily defined on a 5-point scale from –2 favouring celecoxib to +2 favouring SR paracetamol). An individual with an aggregate response absolute value 6 was considered a definite responder, a value 3 but <6 was considered a probable responder, and a value <3 was considered a non-responder.
The Kappa statistic was used to measure the agreement between matched categorical outcome variables.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Recruitment was stopped prematurely in December 2004 because the Australian Therapeutic Goods Administration (TGA) directed all research involving celecoxib to stop in view of newly discovered increased risk of cardiovascular events.
We enrolled 79 patients: 20 did not start their n-of-1 trials; [13 because of the TGA directive, two because of a prior recall of rofecoxib (a COX-2 inhibitor in the same class), and five for other reasons, mainly sulpha allergies], 18 completed only one or two cycles (one because of adverse reactions to celecoxib; six due to severe pain; five due to the TGA directive; three due to the large number of tablets and one each because of concern relating to side-effects of celecoxib, failure to complete diaries, and impending admission for surgery). Marker joints were knee 24, back 14, shoulder/neck 8, hand 6, hip 6, foot 5 (some had multiple sites). Demographic and clinical characteristics of the 59 enrolled patients were unremarkable (Table 1).
|
Blinding
The dose of celecoxib used during the celecoxib periods was 200 mg once a day for 32 patients and 200 mg twice a day for nine patients. Only one of the 41 patients guessed which medication they were using in 6/6 treatment periods correctly, one guessed 5/6, four 4/4 and the remainder 0/6, 1/6 or 2/6 correctly, no different from what could be expected from chance alone. All patients who completed the trial took at least 96% of their tablets.
Pain, stiffness and functional limitation scores
Of the 41 completers, 12 had detectable differences in pain scores (10 in favour of celecoxib), 14 in stiffness scores (12 in favour of celecoxib) and two in functional limitation scores (both in favour of celecoxib). The number of patients with no detectable differences between medications for these scores was 24, 22 and 26, respectively (Table 2).
|
Using hierarchical Bayesian random effects models to meta-analyse differences in mean scores within pairs for all 41 trials, the mean (S.D.) scores for the group were lower for celecoxib: 0.2 (0.1) for pain, 0.3 (0.1) for stiffness and 0.3 (0.1) for functional limitation.
Medication preference
Three patients preferred celecoxib over SR paracetamol in all three cycles and five in two of three cycles. The remaining 33 had no obvious preference.
Adverse events
Only one adverse event—severe foot/ankle swelling on celecoxib—resulted in withdrawal. Nine patients reported more adverse events while on SR paracetamol than on celecoxib, and five reported more while on celecoxib than on SR paracetamol. In the other 25 patients, there was no difference in the prevalence of adverse events reported.
The most common adverse events on celecoxib were headache (54%), loss of energy (54%), indigestion (36%) and constipation (32%); and on SR paracetamol were loss of energy (51%), headache (49%) and constipation and indigestion (44%) (Table 5). There were differences between the two drugs in terms of stomach pain (15% for celecoxib vs 27% for SR paracetamol) and vomiting (2% for celebrex vs 7% for SR paracetamol). This may be due to the small sample size, rather than a real difference in response to the two drugs.
|
Adverse events were mild or moderate in 32 patients. The other nine had between one and three severe symptoms, and one had six severe symptoms. One patient had tinnitus through the trial on both drugs. Two other patients had this with celecoxib only. Other severe events on celecoxib were trembles, upper body rash, loss of energy and indigestion/heartburn.
Two patients had severe loss of energy on SR paracetamol. Other severe events on SR paracetamol were dizziness, diarrhoea, restless leg and poor concentration.
Aggregate scores
The aggregate scores showed no difference between the two medications in 33 (80%) patients; five had scores that showed a probable advantage of celecoxib over SR paracetamol and one had a score that showed a probable advantage of SR paracetamol over celecoxib. Two patients’ scores showed a definite advantage of celecoxib over SR paracetamol. Agreement between all pair-wise comparisons of the five outcome variables contributing to the aggregate score was poor (
< 0.40), except for that between pain and stiffness ( = 0.80). Of the nine patients who were taking celecoxib 200 mg b.i.d., in one patient SR paracetamol was probably better, in three celecoxib was probably better and in the rest there was no difference.
Change in drug use after the n-of-1 trial
Following the n-of-1 trials, there was no change in management in 15/41 (37%) patients; 12/41 (29%) discontinued NSAID/COX-2 inhibitors afterwards; SR paracetamol was added or substituted for 7/41 (17%) patients and SR paracetamol was discontinued in 6/41 (15%) patients (Table 3).
|
Consistency of drug management immediatelyafter the trial with the result of the n-of-1 trial
Among the 33 patients for whom there was no difference between medications, 13 were subsequently managed with SR paracetamol mainly and six with COX-2 inhibitors mainly; three switched to NSAIDs, two ceased drugs and the management was unknown for nine. Of the other eight patients whose results favoured one or the other drug, six were managed consistently with their trial result (i.e. the favoured drug was prescribed). Altogether, in 25/41 (61%) patients, management was consistent with their results, meaning that a logical decision was made based on the results (Table 4); for example, for those with no difference between medications, either medication could be logically prescribed.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
This is the first study to report n-of-1 trials of SR paracetamol vs celecoxib, and one of the few to use Bayesian methods to conduct statistical analysis of n-of-1 trials. It is the first in the pharmacological OA literature to apply Bayesian methods to n-of-1 trial analysis.
The aggregate results showed that most (80%) patients completing an n-of-1 trial had a similar response to celecoxib as to SR paracetamol. Of the remainder, celecoxib was probably better in most. These findings are hardly surprising as they are similar to previous ones for OA [13, 14]. Caution is warranted in generalizing these results to the broader population of patients with OA, as they are derived from a population of patients characterized by uncertainty about the efficacy of their drugs for them as individuals and so may not be representative. Such a population may be less likely to show differences between the drugs as their response rates to both drugs may be lower than in the broader population of patients with OA. Likewise, caution applies to interpreting the post-trial decisions about medications, as the highly publicized problems with COX-2 inhibitors that occurred simultaneously may have led more patients to stop celecoxib than would have otherwise.
The main application of n-of-1 trials in clinical practice could be to guide patients in a rational decision about which of a pair of management options best suits their chronic disease. These data might help medical services decide to adopt this process. We have shown that for one of the most common chronic diseases, the use of n-of-1 trials is entirely feasible. That they are acceptable to many patients can be attested by the fact that they commit to completing the daily-symptom diaries for 12 weeks. The withdrawal rate of 30% is fairly typical of n-of-1 trials [13–15] and indeed typical of many conventional randomized controlled trials [27].
We have shown that the use of n-of-1 trials could promote rational management of chronic OA. The impact of this useful clinical tool on long-term management and subsequent economical consequences needs further evaluation before it becomes more widely accepted.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Thanks to the Australian Health Ministers Advisory Council for funding; GlaxoSmithKline Consumer Healthcare for supply of active and placebo medication and funding and for their invaluable support; Donna-Marie Preston and Grace McBride for collecting the data; and the doctors and patients who took part in this study.
GlaxoSmithKline Consumer Healthcare provided active and placebo medication, and funding; participated in initial planning discussions; and viewed the manuscript prior to publication, but had no involvement in the collection, analysis or interpretation of data, writing the report or the decision to submit the final manuscript for publication. GlaxoSmithKline Consumer Healthcare provided funding that supported the salaries of J.N. and N.M. on the project. M.Y. has been paid by GlaxoSmithKline to present the results of this research at two conferences. The researchers are independent from the funders.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
- World Health Organization. Chronic rheumatic conditions. http://www.who.int/chp/topics/rheumatic/en/ [2005 [cited 5 A.D. Aug. 23].
- March LM, Schwarz JM, Carfrae BH, Bagge E. (1998) Clinical validation of self-reported osteoarthritis. Osteoarthr Cartil 6:87–93.[CrossRef][Web of Science][Medline]
- Jordan KM, Arden NK, Doherty M, et al. (2003) EULAR Recommendations 2003: an evidence based approach to the management of knee osteoarthritis: report of a task force of the Standing Committee for International Clinical Studies Including Therapeutic Trials [ESCISIT]. Ann Rheum Dis 62:1145–55.
[Abstract/Free Full Text] - American College of Rheumatology Subcommittee on Osteoarthritis. Recommendations for the medical management of osteoarthritis of the hip and knee: 2000 update. Arth Rheum (2000) 43:1905–15.[CrossRef][Web of Science][Medline]
- Koes BW, Sanders RJ, Tuut MK. (2004) [The Dutch Institute for Health Care Improvement [CBO] guideline for the diagnosis and treatment of aspecific acute and chronic low back complaints]. Ned Tijdschr Geneeskd 148:310–4.[Medline]
- FDA MedWatch. COX-2 Selective [includes Bextra, Celebrex, and Vioxx] and Non-Selective Non-Steroidal Anti-Inflammatory Drugs [NSAIDs]. Available from http://www.fda.gov/medwatch/SAFETY/2005/safety05.htm#NSAID. Accessed 15 June 2005.
- Day RO GGG. (2004) The vascular effects of COX-2 selective inhibitors. Australian Prescriber 27:142–5.
- Bacon TH, Hole JG, North M, Burnett I. (2002) Analgesic efficacy of sustained release paracetamol in patients with osteoarthritis of the knee. Br J Clin Pharmacol 53:629–36.[CrossRef][Web of Science][Medline]
- Coulthard P, Hill CM, Frame JW, Barry H, Ridge BD, Bacon TH. (2001) Pain control with SR paracetamol from a sustained release formulation and a standard release formulation after third molar surgery: a randomised controlled trial. Br Dent J 191:319–24.[CrossRef][Web of Science][Medline]
- Nikles CJ, Yelland M, Glasziou PP, Del MC. (2005) Do individualized medication effectiveness tests [n-of-1 trials] change clinical decisions about which drugs to use for osteoarthritis and chronic pain? Am J Ther 12:92–7.[CrossRef][Medline]
- March L, Irwig L, Schwarz J, Simpson J, Chock C, Brooks P. (1994) N of 1 trials comparing a non-steroidal anti-inflammatory drug with SR paracetamol in osteoarthritis. Br Med J 309:1041–5.
[Abstract/Free Full Text] - Sheather-Reid RB and Cohen M. (1998) Efficacy of analgesics in chronic pain: a series of N-of-1 studies. J Pain Symptom Manage 15:244–52.[CrossRef][Web of Science][Medline]
- Larson EB. (1990) N-of-1 clinical trials. A technique for improving medical therapeutics. West J Med 152:52–6.[Web of Science][Medline]
- Guyatt GH, Keller JL, Jaeschke R, Rosenbloom D, Adachi JD, Newhouse MT. (1990) The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience. Ann Intern Med 112:293–9.[CrossRef][Web of Science][Medline]
- Pincus T, Koch G, Lei H, et al. (2004) Patient preference for placebo, acetaminophen or celecoxib efficacy studies [PACES]: two randomized placebo-controlled cross-over clinical trials in patients with osteoarthritis of the knee or hip. Ann Rheum Dis 63:931–9.
[Abstract/Free Full Text] - Schnitzer TJ, Weaver AL, Polis AB, Petruschke RA, Geba GP. VACT-1 and VACT-2 (Protocols 106 and 150) Study Groups. (2005) Efficacy of rofecoxib, celecoxib, and acetaminophen in patients with osteoarthritis of the knee. A combined analysis of the VACT studies. J Rheumatol 32:1093–105.
[Abstract/Free Full Text] - Battisti WP, Katz NP, Weaver AL, et al. (2004) Pain management in osteoarthritis: a focus on onset of efficacy—a comparison of rofecoxib, celecoxib, acetaminophen, and nabumetone across four clinical trials. J Pain 5:511–20.[CrossRef][Web of Science][Medline]
- Coxeter PD, Schluter PJ, Eastwood HL, Nikles CJ, Glasziou PP. (2003) Valerian does not appear to reduce symptoms for patients with chronic insomnia in general practice using a series of randomised n-of-1 trials. Complement Ther Med 11:215–22.[CrossRef][Web of Science][Medline]
- Nikles CJ, Clavarino AM, Del Mar CB. (2005) Using n-of-1 trials as a clinical tool to improve prescribing. Br J Gen Pract 55:175–80.[Web of Science][Medline]
- Vollmann J and Winau R. (1996) Informed consent in human experimentation before the Nuremberg code. Br Med J 313:1445–9.
[Free Full Text] - Daut RL, Cleeland CS, Flanery RC. (1983) Development of the Wisconsin Brief Pain Questionnire to assess pain in cancer and other disease. Pain 17:197–210.[CrossRef][Web of Science][Medline]
- Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. (1988) Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 15:1833–40.[Web of Science][Medline]
- Stratford PW and Gill C. (1995) Assessing disability and change in individual patients: a report of a patient-specific measure. Physiother Can 47:258–63.
- Zucker DR, Schmid CH, McIntosh MW, D’Agostino RB, Selker HP, Lau J. (1997) Combining single patient [N-of-1] trials to estimate population treatment effects and to evaluate individual patient responses to treatment. J Clin Epidemiol 50:401–10.[CrossRef][Web of Science][Medline]
- Ehrich EW, Davies GM, Watson DJ, Bolognese JA, Seidenberg BC, Bellamy N. (2000) Minimal perceptible clinical improvement with the Western Ontario and McMaster Universities osteoarthritis index questionnaire and global assessments in patients with osteoarthritis. J Rheumatol 27:2635–41.[Web of Science][Medline]
- Chatman AB, Hyams SP, Neel JM, et al. (1997) The Patient-specific Functional Scale: measurement properties in patients with knee dysfunction. Phys Ther 77:820–29.
[Abstract/Free Full Text] - Martin R, Vogtle L, Gilliam F, Faught E. (2003) Health-related quality of life in senior adults with epilepsy: what we know from randomized clinical trials and suggestions for future research. Epilepsy Behav 4:626–34.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Nikles, G. Mitchell, J. Walters, J. Hardy, P. Good, D. Rowett, T. Shelby-James, and D. Currow Prioritising drugs for single patient (n-of-1) trials in palliative care Palliative Medicine, October 1, 2009; 23(7): 623 - 634. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||