- Email: [email protected]
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee—A randomized, placebo controlled phase II study
Complementary Therapies in Medicine (2010) 18, 113—118
available at www.sciencedirect.com
journal homepage: www.elsevierhealth.com/journals/ctim
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee—–A randomized, placebo controlled phase II study Roman Huber a,∗, Ute Prestel a, Isabel Bloss a, Ulrich Meyer b, Rainer Lüdtke c a
Department of Environmental Health Sciences, University Hospital Freiburg, Breisacher Str. 115b, D-79106 Freiburg, Germany WALA Heilmittel GmbH, Bad Boll, Germany c Karl and Veronica Carstens Foundation, Essen, Germany Available online 1 July 2010 b
KEYWORDS Osteoarthritis; Organ preparation; Anthroposophic medicine; WOMAC; Randomized controlled trial
Summary Objectives and methods: Articulatio Genus D5 (AG5), a cartilage preparation, is widely used within anthroposophic medicine to treat osteoarthritis of the knee (OA). To evaluate its efficacy 91 OA patients were randomized to receive either AG5 (46) or placebo (45) following a fully sequential study design. For a period of 4 weeks AG5 or placebo were injected subcutaneously 3 times a week in the predominantly affected knee. Outcome parameters were the Western Ontario and Mc. Masters University Arthritis Index (WOMAC) scores and the use of concomitant pain medication during a follow up period of 2 months. Results: 89 patients completed the study. After 4 weeks of treatment WOMAC pain scores were significantly reduced compared to baseline (mean from 5.1 ± 1.8 to 3.2 ± 2.2 with AG5 and from 4.3 ± 1.5 to 3.0 ± 1.8 with placebo, p < 0.001) in both groups. This effect was maintained until follow up (mean 3.3 ± 2.6 with AG5 and 3.2 ± 2.4 with placebo). WOMAC function- and stiffness scores and use of concomitant pain medication were also significantly reduced. There was, however, no significant difference between AG5 and placebo (group difference after 4 weeks: −3.3, CI: −11.8 to 5.3, p = 0.46, triangular test). After follow up, WOMAC scores were also not different between AG5 and placebo but the use of pain medication dropped significantly more in the AG5 group (p = 0.036). Severe side effects did not occur, minor side effects in 2 patients (1 in each group) were related to the subcutaneous injection. Conclusions: For the primary outcome parameter pain AG5 was not superior to placebo after 4 weeks of treatment but the equal reduction of pain in combination with higher reduction of pain medication suggests an effect of the preparation. The long term effect of AG5 on sparing pain medication should be further investigated. © 2010 Elsevier Ltd. All rights reserved.
Introduction ∗ Corresponding author. Tel.: +49 761 270 8201; fax: +49 761 270 8323. E-mail address: [email protected] (R. Huber).
Osteoarthritis of the knee (OA) is the most common form of OA,1 associated with a distinct reduction of quality of life2 and enormous costs for the health systems in industrialized
0965-2299/$ — see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ctim.2010.06.001
114 countries. In Germany the direct and indirect costs of OA amount to more than 8 billion Euro.3 Oral NSAIDs are frequently used to treat OA symptoms, but bear the risk of gastrointestinal bleeding and disturbed kidney function especially in elderly patients.4,5 Therefore, there is a need for well tolerable, effective treatments of OA related complaints which consist mainly in pain, associated to physical strain and later in function deficit of the knee. The causes of pain in patients with OA knee are complex, involve central and peripheral mechanisms and are yet not fully understood. Because the joint cartilage is not directly supplied by nerves, surrounding tissues play a major role, including the joint capsule, ligaments, synovium, bone and the outer edge of the menisci.6,7 Changes of the subchondral bone,8 stiffness of ligaments,9 enthesopathy,10 altered proprioception11 and muscle hyperalgesia12 have been found as potential factors generating or exaggerating OA pain. It has therefore been hypothesized that subcutaneous injection of analgetic or anti-inflammatory substances into trigger points at the knee may improve OA symptoms. Local anaesthetics, different plant extracts and organ extracts have been used.13—16 Their effectiveness has, however, not yet been evaluated in larger RCT’s. In this study we investigated a diluted cartilage extract which has been used since decades in anthroposophical medicine. Numerous case reports have suggested its effectiveness in OA claiming that the extract stimulates the self-healing properties in the knee.17—19 According to these reports subcutaneous injections to the knee with extracts from bovine knee cartilage have significantly and sustainably improved OA symptoms after 3—4 weeks of treatment.
Patients and methods The study was prospective, randomized, double blind and placebo controlled with two parallel arms. Patients were sequentially allocated to treatments in incoming order. Allocation was concealed and based on an unstratified block randomization with a block length unknown by the treating physicians. The study medication was provided from the manufacturer (WALA Heilmittel GmbH, Bad Boll, Germany) in sequentially numbered ampoules, identical in appearance. Patients and all investigators were completely blind to treatment allocation throughout the whole study and the code was opened only after all data had been collected. The study was approved by the Ethical Committee of University Hospital Freiburg and registered in the clinical trials register of University Hospital Freiburg, Germany. Patients of both sexes were recruited by an announcement in the local newspaper and were screened when suffering from primary knee OA. Further inclusion criteria were Kellgren and Lawrence osteoarthritis stage 2—3, age between 40 and 85 years, body mass index <35 kg/m2 , pain in the knee on more than 50% of the days of the last 3 months before study onset, and initial pain scores >4 cm in at least one of the Western Ontario and Mc. Masters University Arthritis Index (WOMAC) pain scales. Diagnosis of OA was documented by conventional X-ray of the knee prior to inclusion in the study and physical examination. Exclusion criteria were:
R. Huber et al. - coagulation disorders with INR >2.5, intake of anticoagulants from Cumarine-type or haemophilia, - systemic intake of steroids or other immunosuppressants, - injections into the joint or arthroscopy within the last 3 months before onset of the study, - secondary arthritis as suggested from X-ray, blood sedimentation rate >40 cm in 1 h or from the patients history, - known fibromyalgia or least 11 from 18 positive fibromyalgia trigger points in physical examination, - knee injury in the last 6 weeks before onset of the study, - planned surgical intervention within the next 3 months, - severe concomitant diseases (e.g. decompensated heart, kidney- or liver disease, malignancies, HIV, severe psychiatric disorders), - pregnancy.
Interventions and outcome parameters AG5 is a glycerol extract from bovine knee cartilage which has been diluted 5 times 1:10. AG5 is licensed by WALA Heilmittel GmbH, Bad Boll, Germany, for supportive treatment of degenerative knee disorders according to principles of Anthroposophical Medicine. Placebo was the solvent of AG5, an isotonic solution of sodium hydrogen carbonate and sodium chloride. Each ampoule contained 1 ml. According to recommendations of the manufacturer one ampoule was injected 3 times a week (Monday, Wednesday, Friday) subcutaneously for a period of 4 weeks. The injections were performed at two trigger points (points painful at palpation) if available, otherwise above the medial and lateral ligaments of the knee. 1/2 ampoule was injected at each trigger point. The injections were given by physicians of the study centre or by the general practitioners of the patients. The patients were asked to reduce their pain medication to an on demand non-steroidal anti-inflammatory drug (NSAID) only. Primary outcome parameter was the WOMAC pain score after 4 weeks of treatment.20,21 Secondary outcome parameters were WOMAC pain score at day 84 (follow up), WOMAC stiffness and function scores, concomitant pain medication, which was recorded before onset of the study, day 28 and day 84, tolerability (excellent, good, moderate, bad, very bad), which was assessed day 28, safety parameters (C-reactive protein, prothrombin time, urea, creatinine, bilirubin, alanine aminotransferase, aspartate-aminotransferase, gamma-glutamyl-transferase) which were investigated before onset of the study and at day 28 and side effects, assessed at every visit. Furthermore, the patients were asked at day 84, which medication they believed to have received.
Statistics This study was conducted in a fully sequential design, meaning that group differences were evaluated immediately each time a patient contributes outcome data. Consequently, sequential trials are terminated as early as there is enough evidence for or against the primary study hypothesis. In a sequential trial, the number of patients cannot be determined a priori: even when the anticipated effect sizes and error probabilities are identical, the actual number
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee of patients differs from trial to trial. Instead, sample size calculations are usually based on an average trial, thus estimating the expected number of patients to be included. In this study it was a priori assumed that the patients in the AG5 group improved by 1.8 ± 1.6 cm in the WOMAC pain scale compared to 1.0 ± 1.6 cm in the placebo group. Allowing for a type 1 error of ˛ = 2.5% (one-sided) and a statistical power of ˇ = 80%, we expected to enrol 84 patients (42 per group) on average (calculated with the PEST® software package 4.1, MPS Research Unit, Reading, UK). Effectiveness analyses were based on an intention-totreat (ITT) population, meaning that each patient was evaluated in that group he or she was randomized to regardless of any protocol violations. Missing values were replaced, if necessary, by last-observation-carried-forward. According to its sequential nature the primary analysis in our study was based on a one-sided triangle test (˛ = 2.5%), which included treatment group and WOMAC pain score at baseline as covariables. Secondary parameters were evaluated by analyses of covariance (ANCOVA) again including treatment group and the respective baseline parameter as covariables.
Results Due to the sequential nature of the trial, it was stopped after inclusion of 91 patients. 89 patients finished the study regularly; two (one in each group) dropped out due to reasons not related to the study medication (Fig. 1). Despite randomization the groups differed significantly in WOMAC pain and function scores with higher scores in the AG5 group. All other baseline characteristics were balanced between the groups (Table 1). From baseline to day 28 WOMAC pain, stiffness and function scores improved significantly in both groups (each p ≤ 0.001, Table 2). Mean improvements of the WOMAC pain
Table 1
Figure 1
115
Flow of the study.
score at day 28 and 84 were −1.9 ± 2.1 and −1.9 ± 2.4 cm with AG5 and −1.3 ± 1.8 and −1.1 ± 2.3 cm with placebo. This resulted in a small and insignificant group difference of d = 0.33 cm in favour of AG5 (95% confidence interval (CI): −11.8 to 5.3, p = 0.459) at day 28. Similar results were obtained for day 84 (d = 0.47, CI: −13.3 to 3.9, p = 0.283). Concomitant use of pain medication, mostly oral NSAID could be reduced in both groups and was at day 84 even lower than after treatment at day 28 (Fig. 2). While at day 28 the difference was not yet significant (p = 0.941), the number of patients who stopped pain medication at follow up was significantly higher in the AG5 group (n = 22) than in the placebo group (n = 11, p = 0.036). At day 84, 40% of the placebo but only 15% of the AG5 patients suspected to have received placebo. In contrast 27%
Baseline characteristics of the patients (absolute numbers and percentage or means ± standard deviations). Placebo (n = 45)
AG5 (n = 46)
p-Value
Male/female Age in years Body mass index (kg/m2 )
12/33 64 ± 10 26.9 ± 3.7
17/29 68 ± 8 26.1 ± 4.1
0.292 0.083 0.674
Arthroscopy in the past Clinical signs of effusion Pain medication (NSAID) at baseline Other therapies at baseline
16 (35.6%) 9 (20.0%) 24 (53.3%) 12 (26.7%)
13 (28.3%) 9 (19.6%) 29 (63.0%) 13 (28.3%)
0.455 0.958 0.348 0.856
Kellgren stage 1 2 3 4
2 (4.4%) 23 (51.1%) 20 (44.4%) 0 (0.0%)
0 (0.0%) 28 (60.9%) 15 (32.6%) 3 (6.5%)
0.978
Duration of disease (years)
9.7 ± 9.8
9.2 ± 8.1
0.855
WOMAC pain score WOMAC stiffness score WOMAC function score
4.3 ± 1.5 4.2 ± 2.2 4.3 ± 1.5
5.1 ± 1.8 4.6 ± 2.1 5.1 ± 1.7
0.016 0.472 0.022
116 Table 2
R. Huber et al. WOMAC scores before, during and after treatment with AG5 or placebo: median and mean ± standard deviations. Baseline
WOMAC pain score WOMAC stiffness score WOMAC function score
4 weeks treatment
8 weeks follow up
AG5
Placebo
AG5
Placebo
AG5
Placebo
5.2 5.1 ± 1.8 4.5 4.6 ± 2.1 5.1 5.1 ± 1.7
4.2 4.3 ± 1.5 4.5 4.2 ± 2.2 4.3 4.3 ± 1.5
2.7 3.2 ± 2.2 2.9 3.3 ± 2.5 2.9 3.5 ± 2.4
2.7 3.0 ± 1.8 2.3 2.9 ± 2.1 2.9 3.2 ± 1.9
2.9 3.3 ± 2.6 2.8 3.3 ± 2.7 2.7 3.4 ± 2.6
2.8 3.2 ± 2.4 2.3 3.3 ± 2.6 2.6 3.4 ± 2.5
Figure 2 Percentage of patients with use of concomitant pain medication.
of the placebo but 37% of the verum group suspected to have received verum (p = 0.073). Tolerability was rated by the patients as good or excellent by 98% of both, placebo and AG5 (rest did not answer the question). Mild local irritations occurred after single injections in 1 patient of the AG5 and 1 patient of the placebo group. Severe side effects did not occur. There was no systematic deviation and no significant difference between AG5 and placebo regarding the safety laboratory parameters.
Discussion AG5 is claimed to stimulate self-healing properties in patients with OA knee. We found, that in patients with moderate complaints from osteoarthritis of the knee injections of AG5 or placebo into painful trigger points of the knee resulted in clinically relevant and statistically significant improvements in all WOMAC scales. Superiority of AG5 beyond placebo, however, could not be demonstrated for the primary outcome parameter after 4 weeks of treatment. At follow up AG5 had an effect, because stop of pain medication was more frequent in the AG5 than in the placebo group and an improvement of the WOMAC scores could therefore have been masked. Regarding the major health risks
of NSAID the higher rate of stopping pain medication is a relevant outcome and deserves further investigations. Also the trend to a higher number of patients guessing correctly to have verum (p = 0.073) speaks for an effect of AG5. The ampoules were absolutely indistinguishable from appearance and injection did not cause local reactions which could deblind the patients. So this result could be explained either by chance or because the verum had an effect which is not covered by the questionnaires. Case reports with AG5 in OA patients18,19 suggest a rapid effect after 4 and 3 weeks, respectively which correlates to our results. Therefore, and because it was a recommendation of the manufacturer, we decided to treat for 4 weeks only. Stimulation of selfhealing properties might, however, need longer treatment for maximal effects. In studies with successful acupuncture treatment, also a stimulatory principle, and similar baseline pain scores as in our study22,23 the treatment duration was longer (8—13 weeks) than in our study. It was, however, not longer than in studies with successful symptomatic NSAID treatment and similar baseline pain scores (6—13 weeks, Table 3).24,25 From these data it is unlikely that stimulatory therapeutic principles need much longer time than symptomatic therapies. It cannot be excluded, however, that prolonged treatment would have shown better effects in our study but it is unlikely that differences between placebo and verum were missed after 4 weeks treatment and 8 weeks follow up. Despite randomization the treatment groups were not well balanced in baseline WOMAC scores. It should, however, been noted, that the estimated group differences were not affected by this imbalance because baseline scores entered all statistical models as a covariate. Irrespective of the medication the effects of subcutaneous injections were distinct and comparable to or even better than effects of NSAID in other studies with similar baseline pain scores24,25 and a little less than improvements with acupuncture (Table 3).22,23 Also the excellent tolerability and safety of the subcutaneous injections suggests their relevance as complementary or alternative symptom modifying treatment in OA patients. Numerous study patients of both groups asked because of their good experiences for repetition of the treatment in the months after the study was finished. Most improvements in this study were probably caused by strong unspecific effects. In studies on acupuncture or intraarticular injections ‘‘placebo’’ effects were considerably stronger than the effects of oral placebos (Table 3).26,27
Randomized, placebo controlled studies with primary osteoarthritis of the knee and similar baseline characteristics to our study.
Study
Groups
n
End of treatment
WOMAC baseline score
WOMAC score at end of treatment
Difference baseline-end of treatment
p-Value
Pincus et al.25
Celecoxib 200 mg daily vs. paracetamol 4× 1000 mg daily vs. placebo
524
6 weeks
1. Celecoxib: 4.9 ± 1.5 2. Paracetamol: 5.3 ± 1.5 3. Placebo: 5.0 ± 1.6
3.8 ± 1.5
−1.0
1. vs. 2. n.s.
4.4 ± 1.5
−0.8
2. vs. 3. n.s.
4.6 ± 1.7
−0.5
1. vs. 3. =0.002
3.5 ± 1.7
−1.4
1. vs. 4. <0.01
3.4 ± 1.7
−1.4
2. vs. 4. <0.01
3.5 ± 1.6
−1.3
3. vs. 4. <0.01
4.0 ± 1.6
−0.9
Tannenbaum et al.24
Lumiracoxib 200 mg vs. lumiracoxib 400 mg vs. celecoxib 200 mg vs. placebo
1702
13 weeks
1. Lumiracoxib 200: 9 ± 1.4 2. Lumiracoxib 400: 4.8 ± 1.5 3. Celecoxib: 4.9 ± 1.5 4. Placebo: 4.9 ± 1.4
Raynauld et al.26
Steroids vs. saline intraarticular every 3 months
68
1 year
1. Steroids: 4.0 ± 2.6 2. Placebo: 4.8 ± 2.8
2.6 ± 2.0 3.5 ± 1.9
−1.4 −1.3
0.87
Karlsson et al.27
Hyaluronic acid A vs. hyaluronic acid B. vs. saline intraarticular thrice per week
246
12 weeks
1. Hyaluronic acid A: 4.9 ± 1.3 2. Hyaluronic acid B: 4.9 ± 1.1 3. Placebo: 4.9 ± 1.3
3.5
−1.4
n.s.
3.2
−1.7
3.1
−1.8
1. Acupuncture: 5.1 ± 1.9 2. Minimal acupuncture: 5.3 ± 1.9 3. Waiting group: 5.2 ± 1.9 1. Acupuncture: 5.4 ± 2.0 2. Sham acupuncture: 5.5 ± 1.9 3. Standard therapy: 5.5 ± 2.0
2.7 ± 1
−2.4
1. vs. 2. <0.001
3.6 ± 2
−1.7
1. vs. 3. <0.001
5.0 ± 2
−0.2
Witt et al.23
Scharf et al.22
12 sessions acupuncture vs. minimal acupuncture vs. waiting group
10 sessions acupuncture vs. sham acupuncture vs. standard therapy with NSAID
300
1039
8 weeks
13 weeks
−2.1
1. vs. 3. <0.001
−1.9
2. vs. 3 <0.001
−0.9
1. vs. 2. =0.51
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee
Table 3
vs.: versus and n.s.: not significant.
117
118 This corroborates the hypothesis that injections have stronger placebo effects than oral medications.28 Effects of injections to the knee in osteoarthritis patients may however not merely be attributed to psychological placebo effects. Probably more appropriate they can be classified as unspecific effects, because injections to the knee can affect pain mediating structures as pointed out in introduction. Such unspecific effects could have occurred in this study by penetrating the skin, counter irritation, application of 1 ml fluid subcutaneously and hereby neuronal or metabolic changes at trigger or acupuncture points or other pain mediation structures of the knee. This is also supported by the acupuncture studies which suggested that sham acupuncture at the knee was more effective than sham acupuncture at points distant from the knee.22,23 In conclusion, injections to the knee were accompanied by strong and clinical relevant improvements of the complaints of knee osteoarthritis in both, AG5 and verum group. For the primary outcome parameter pain AG5 was not superior to placebo after 4 weeks of treatment but the equal reduction of pain in combination with higher reduction of pain medication suggests an effect of the preparation. Because subcutaneous injections to the knee are a safe and cheap procedure, their clinical relevance should be further investigated in comparison to standard therapies and acupuncture. AG5 should be further investigated in studies with longer treatment duration.
Conflict of interest The authors declare that there is no conflict of interest.
Acknowledgment Medication and financial compensation for the study was kindly provided by WALA Heilmittel GmbH, Bad Boll, Germany.
References 1. Engelhardt M. Epidemiology of osteoarthritis in Western Europe. Dtsch Z Sportmed 2003;54(6):171—3. 2. Sprangers MA, de Regt EB, Andries F, van Agt HM, Bijl RV, de Boer JB, et al. Which chronic conditions are associated with better or poorer quality of life? J Clin Epidemiol 2000;53(9):895—907. 3. Kessler S. Behandlungsstrategien der Gonarthrose. Med Welt 2003;54:12—7. 4. Gurwitz JH, Avorn J. The ambiguous relation between aging and adverse drug reactions. Ann Intern Med 1991;114(11):956—66. 5. Tramèr MR, Moore RA, Reynolds DJ, McQuay HJ. Quantitative estimation of rare adverse events which follow a biological progression: a new model applied to chronic NSAID use. Pain 2000;85(1—2):169—82. 6. Felson DT. The sources of pain in knee osteoarthritis. Curr Opin Rheumatol 2005;17(5):624—8. 7. Felson DT. Developments in the clinical understanding of osteoarthritis. Arthritis Res Ther 2009;11(1):203. 8. Sulzbacher I. Arthrosis-histology and pathogenetic approaches. Radiologe 2000;40(12):1126—33. 9. Fishkin Z, Miller D, Ritter C, Ziv I. Changes in human knee ligament stiffness secondary to osteoarthritis. J Orthop Res 2002;20(2):204—7.
R. Huber et al. 10. Slobodin G, Rozenbaum M, Boulman N, Rosner I. Varied presentations of enthesopathy. Semin Arthritis Rheum 2007;37(2):119—26. 11. Weiler HT, Pap G, Awiszus F. The role of joint afferents in sensory processing in osteoarthritic knees. Rheumatology (Oxford) 2000;39(8):850—6. 12. Bajaj P, Bajaj P, Graven-Nielsen T, Arendt-Nielsen L. Osteoarthritis and its association with muscle hyperalgesia: an experimental controlled study. Pain 2001;93(2):107—14. 13. Scott NA, Guo B, Barton PM, Gerwin RD. Trigger point injections for chronic non-malignant musculoskeletal pain: a systematic review. Pain Med 2009;10(1):54—69. 14. Sickel K. Plenosol in the treatment of arthrotic changes. Z Allgemeinmed 1971;47(31):1667—9. 15. Roemer F, Roggatz M, Soldner G, Sommer M, Vogel HH. Anthroposophische Medizin in der Praxis 2. 1st ed. Natur Mensch Medizin Verlags GmbH Bad Boll; 2002. 16. Barron MC, Rubin BR. Managing osteoarthritic knee pain. J Am Osteopath Assoc 2007;107(10 Suppl. 6):ES21—7. 17. Roemer F, Sommer M. Zur Bedeutung der potenzierten Organpräparate in der anthroposophischen Therapierichtung. Der Merkurstab; 1999. Sonderheft: 2—5. 18. Roemer F. Kasuistik fortgeschrittene Valgusgonarthrose. Der Merkurstab; 1999. Sonderheft: 33—34. 19. Roemer F. Kasuistik entzündlicher Reizzustand mit massivem Kniegelenkserguss auf der Grundlage einer Gonarthrose. Der Merkurstab; 1999. Sonderheft: 35. 20. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. 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 1988;15(12): 1833—40. 21. Stucki G, Meier D, Stucki S, Michel BA, Tyndall AG, Dick W, et al. Evaluation of a German version of WOMAC (Western Ontario and McMaster Universities) Arthrosis Index. Z Rheumatol 1996;55(1):40—9. 22. Scharf HP, Mansmann U, Streitberger K, Witte S, Krämer J, Maier C, et al. Acupuncture and knee osteoarthritis: a three-armed randomized trial. Ann Intern Med 2006;145(1): 12—20. 23. Witt C, Brinkhaus B, Jena S, Linde K, Streng A, Wagenpfeil S, et al. Acupuncture in patients with osteoarthritis of the knee: a randomised trial. Lancet 2005;366(9480):136—43. 24. Tannenbaum H, Berenbaum F, Reginster JY, Zacher J, Robinson J, Poor G, et al. Lumiracoxib is effective in the treatment of osteoarthritis of the knee: a 13 week, randomised, double blind study versus placebo and celecoxib. Ann Rheum Dis 2004;63(11):1419—26. 25. Pincus T, Koch G, Lei H, Mangal B, Sokka T, Moskowitz R, et al. Patient preference for placebo. Acetaminophen (paracetamol) or Celecoxib Efficacy Studies (PACES): two randomised, double blind, placebo controlled, crossover clinical trials in patients with knee or hip osteoarthritis. Ann Rheum Dis 2004;63(8):931—9. 26. Raynauld JP, Buckland-Wright C, Ward R, Choquette D, Haraoui B, Martel-Pelletier J, et al. Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2003;48(2):370—7. 27. Karlsson J, Sjögren LS, Lohmander LS. Comparison of two hyaluronan drugs and placebo in patients with knee osteoarthritis. A controlled, randomized, double-blind, parallel-design multicentre study. Rheumatology (Oxford) 2002;41(11):1240—8. 28. de Craen AJ, Tijssen JG, de Gans J, Kleijnen J. Placebo effect in the acute treatment of migraine: subcutaneous placebos are better than oral placebos. J Neurol 2000;247(3):183—8.
available at www.sciencedirect.com
journal homepage: www.elsevierhealth.com/journals/ctim
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee—–A randomized, placebo controlled phase II study Roman Huber a,∗, Ute Prestel a, Isabel Bloss a, Ulrich Meyer b, Rainer Lüdtke c a
Department of Environmental Health Sciences, University Hospital Freiburg, Breisacher Str. 115b, D-79106 Freiburg, Germany WALA Heilmittel GmbH, Bad Boll, Germany c Karl and Veronica Carstens Foundation, Essen, Germany Available online 1 July 2010 b
KEYWORDS Osteoarthritis; Organ preparation; Anthroposophic medicine; WOMAC; Randomized controlled trial
Summary Objectives and methods: Articulatio Genus D5 (AG5), a cartilage preparation, is widely used within anthroposophic medicine to treat osteoarthritis of the knee (OA). To evaluate its efficacy 91 OA patients were randomized to receive either AG5 (46) or placebo (45) following a fully sequential study design. For a period of 4 weeks AG5 or placebo were injected subcutaneously 3 times a week in the predominantly affected knee. Outcome parameters were the Western Ontario and Mc. Masters University Arthritis Index (WOMAC) scores and the use of concomitant pain medication during a follow up period of 2 months. Results: 89 patients completed the study. After 4 weeks of treatment WOMAC pain scores were significantly reduced compared to baseline (mean from 5.1 ± 1.8 to 3.2 ± 2.2 with AG5 and from 4.3 ± 1.5 to 3.0 ± 1.8 with placebo, p < 0.001) in both groups. This effect was maintained until follow up (mean 3.3 ± 2.6 with AG5 and 3.2 ± 2.4 with placebo). WOMAC function- and stiffness scores and use of concomitant pain medication were also significantly reduced. There was, however, no significant difference between AG5 and placebo (group difference after 4 weeks: −3.3, CI: −11.8 to 5.3, p = 0.46, triangular test). After follow up, WOMAC scores were also not different between AG5 and placebo but the use of pain medication dropped significantly more in the AG5 group (p = 0.036). Severe side effects did not occur, minor side effects in 2 patients (1 in each group) were related to the subcutaneous injection. Conclusions: For the primary outcome parameter pain AG5 was not superior to placebo after 4 weeks of treatment but the equal reduction of pain in combination with higher reduction of pain medication suggests an effect of the preparation. The long term effect of AG5 on sparing pain medication should be further investigated. © 2010 Elsevier Ltd. All rights reserved.
Introduction ∗ Corresponding author. Tel.: +49 761 270 8201; fax: +49 761 270 8323. E-mail address: [email protected] (R. Huber).
Osteoarthritis of the knee (OA) is the most common form of OA,1 associated with a distinct reduction of quality of life2 and enormous costs for the health systems in industrialized
0965-2299/$ — see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ctim.2010.06.001
114 countries. In Germany the direct and indirect costs of OA amount to more than 8 billion Euro.3 Oral NSAIDs are frequently used to treat OA symptoms, but bear the risk of gastrointestinal bleeding and disturbed kidney function especially in elderly patients.4,5 Therefore, there is a need for well tolerable, effective treatments of OA related complaints which consist mainly in pain, associated to physical strain and later in function deficit of the knee. The causes of pain in patients with OA knee are complex, involve central and peripheral mechanisms and are yet not fully understood. Because the joint cartilage is not directly supplied by nerves, surrounding tissues play a major role, including the joint capsule, ligaments, synovium, bone and the outer edge of the menisci.6,7 Changes of the subchondral bone,8 stiffness of ligaments,9 enthesopathy,10 altered proprioception11 and muscle hyperalgesia12 have been found as potential factors generating or exaggerating OA pain. It has therefore been hypothesized that subcutaneous injection of analgetic or anti-inflammatory substances into trigger points at the knee may improve OA symptoms. Local anaesthetics, different plant extracts and organ extracts have been used.13—16 Their effectiveness has, however, not yet been evaluated in larger RCT’s. In this study we investigated a diluted cartilage extract which has been used since decades in anthroposophical medicine. Numerous case reports have suggested its effectiveness in OA claiming that the extract stimulates the self-healing properties in the knee.17—19 According to these reports subcutaneous injections to the knee with extracts from bovine knee cartilage have significantly and sustainably improved OA symptoms after 3—4 weeks of treatment.
Patients and methods The study was prospective, randomized, double blind and placebo controlled with two parallel arms. Patients were sequentially allocated to treatments in incoming order. Allocation was concealed and based on an unstratified block randomization with a block length unknown by the treating physicians. The study medication was provided from the manufacturer (WALA Heilmittel GmbH, Bad Boll, Germany) in sequentially numbered ampoules, identical in appearance. Patients and all investigators were completely blind to treatment allocation throughout the whole study and the code was opened only after all data had been collected. The study was approved by the Ethical Committee of University Hospital Freiburg and registered in the clinical trials register of University Hospital Freiburg, Germany. Patients of both sexes were recruited by an announcement in the local newspaper and were screened when suffering from primary knee OA. Further inclusion criteria were Kellgren and Lawrence osteoarthritis stage 2—3, age between 40 and 85 years, body mass index <35 kg/m2 , pain in the knee on more than 50% of the days of the last 3 months before study onset, and initial pain scores >4 cm in at least one of the Western Ontario and Mc. Masters University Arthritis Index (WOMAC) pain scales. Diagnosis of OA was documented by conventional X-ray of the knee prior to inclusion in the study and physical examination. Exclusion criteria were:
R. Huber et al. - coagulation disorders with INR >2.5, intake of anticoagulants from Cumarine-type or haemophilia, - systemic intake of steroids or other immunosuppressants, - injections into the joint or arthroscopy within the last 3 months before onset of the study, - secondary arthritis as suggested from X-ray, blood sedimentation rate >40 cm in 1 h or from the patients history, - known fibromyalgia or least 11 from 18 positive fibromyalgia trigger points in physical examination, - knee injury in the last 6 weeks before onset of the study, - planned surgical intervention within the next 3 months, - severe concomitant diseases (e.g. decompensated heart, kidney- or liver disease, malignancies, HIV, severe psychiatric disorders), - pregnancy.
Interventions and outcome parameters AG5 is a glycerol extract from bovine knee cartilage which has been diluted 5 times 1:10. AG5 is licensed by WALA Heilmittel GmbH, Bad Boll, Germany, for supportive treatment of degenerative knee disorders according to principles of Anthroposophical Medicine. Placebo was the solvent of AG5, an isotonic solution of sodium hydrogen carbonate and sodium chloride. Each ampoule contained 1 ml. According to recommendations of the manufacturer one ampoule was injected 3 times a week (Monday, Wednesday, Friday) subcutaneously for a period of 4 weeks. The injections were performed at two trigger points (points painful at palpation) if available, otherwise above the medial and lateral ligaments of the knee. 1/2 ampoule was injected at each trigger point. The injections were given by physicians of the study centre or by the general practitioners of the patients. The patients were asked to reduce their pain medication to an on demand non-steroidal anti-inflammatory drug (NSAID) only. Primary outcome parameter was the WOMAC pain score after 4 weeks of treatment.20,21 Secondary outcome parameters were WOMAC pain score at day 84 (follow up), WOMAC stiffness and function scores, concomitant pain medication, which was recorded before onset of the study, day 28 and day 84, tolerability (excellent, good, moderate, bad, very bad), which was assessed day 28, safety parameters (C-reactive protein, prothrombin time, urea, creatinine, bilirubin, alanine aminotransferase, aspartate-aminotransferase, gamma-glutamyl-transferase) which were investigated before onset of the study and at day 28 and side effects, assessed at every visit. Furthermore, the patients were asked at day 84, which medication they believed to have received.
Statistics This study was conducted in a fully sequential design, meaning that group differences were evaluated immediately each time a patient contributes outcome data. Consequently, sequential trials are terminated as early as there is enough evidence for or against the primary study hypothesis. In a sequential trial, the number of patients cannot be determined a priori: even when the anticipated effect sizes and error probabilities are identical, the actual number
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee of patients differs from trial to trial. Instead, sample size calculations are usually based on an average trial, thus estimating the expected number of patients to be included. In this study it was a priori assumed that the patients in the AG5 group improved by 1.8 ± 1.6 cm in the WOMAC pain scale compared to 1.0 ± 1.6 cm in the placebo group. Allowing for a type 1 error of ˛ = 2.5% (one-sided) and a statistical power of ˇ = 80%, we expected to enrol 84 patients (42 per group) on average (calculated with the PEST® software package 4.1, MPS Research Unit, Reading, UK). Effectiveness analyses were based on an intention-totreat (ITT) population, meaning that each patient was evaluated in that group he or she was randomized to regardless of any protocol violations. Missing values were replaced, if necessary, by last-observation-carried-forward. According to its sequential nature the primary analysis in our study was based on a one-sided triangle test (˛ = 2.5%), which included treatment group and WOMAC pain score at baseline as covariables. Secondary parameters were evaluated by analyses of covariance (ANCOVA) again including treatment group and the respective baseline parameter as covariables.
Results Due to the sequential nature of the trial, it was stopped after inclusion of 91 patients. 89 patients finished the study regularly; two (one in each group) dropped out due to reasons not related to the study medication (Fig. 1). Despite randomization the groups differed significantly in WOMAC pain and function scores with higher scores in the AG5 group. All other baseline characteristics were balanced between the groups (Table 1). From baseline to day 28 WOMAC pain, stiffness and function scores improved significantly in both groups (each p ≤ 0.001, Table 2). Mean improvements of the WOMAC pain
Table 1
Figure 1
115
Flow of the study.
score at day 28 and 84 were −1.9 ± 2.1 and −1.9 ± 2.4 cm with AG5 and −1.3 ± 1.8 and −1.1 ± 2.3 cm with placebo. This resulted in a small and insignificant group difference of d = 0.33 cm in favour of AG5 (95% confidence interval (CI): −11.8 to 5.3, p = 0.459) at day 28. Similar results were obtained for day 84 (d = 0.47, CI: −13.3 to 3.9, p = 0.283). Concomitant use of pain medication, mostly oral NSAID could be reduced in both groups and was at day 84 even lower than after treatment at day 28 (Fig. 2). While at day 28 the difference was not yet significant (p = 0.941), the number of patients who stopped pain medication at follow up was significantly higher in the AG5 group (n = 22) than in the placebo group (n = 11, p = 0.036). At day 84, 40% of the placebo but only 15% of the AG5 patients suspected to have received placebo. In contrast 27%
Baseline characteristics of the patients (absolute numbers and percentage or means ± standard deviations). Placebo (n = 45)
AG5 (n = 46)
p-Value
Male/female Age in years Body mass index (kg/m2 )
12/33 64 ± 10 26.9 ± 3.7
17/29 68 ± 8 26.1 ± 4.1
0.292 0.083 0.674
Arthroscopy in the past Clinical signs of effusion Pain medication (NSAID) at baseline Other therapies at baseline
16 (35.6%) 9 (20.0%) 24 (53.3%) 12 (26.7%)
13 (28.3%) 9 (19.6%) 29 (63.0%) 13 (28.3%)
0.455 0.958 0.348 0.856
Kellgren stage 1 2 3 4
2 (4.4%) 23 (51.1%) 20 (44.4%) 0 (0.0%)
0 (0.0%) 28 (60.9%) 15 (32.6%) 3 (6.5%)
0.978
Duration of disease (years)
9.7 ± 9.8
9.2 ± 8.1
0.855
WOMAC pain score WOMAC stiffness score WOMAC function score
4.3 ± 1.5 4.2 ± 2.2 4.3 ± 1.5
5.1 ± 1.8 4.6 ± 2.1 5.1 ± 1.7
0.016 0.472 0.022
116 Table 2
R. Huber et al. WOMAC scores before, during and after treatment with AG5 or placebo: median and mean ± standard deviations. Baseline
WOMAC pain score WOMAC stiffness score WOMAC function score
4 weeks treatment
8 weeks follow up
AG5
Placebo
AG5
Placebo
AG5
Placebo
5.2 5.1 ± 1.8 4.5 4.6 ± 2.1 5.1 5.1 ± 1.7
4.2 4.3 ± 1.5 4.5 4.2 ± 2.2 4.3 4.3 ± 1.5
2.7 3.2 ± 2.2 2.9 3.3 ± 2.5 2.9 3.5 ± 2.4
2.7 3.0 ± 1.8 2.3 2.9 ± 2.1 2.9 3.2 ± 1.9
2.9 3.3 ± 2.6 2.8 3.3 ± 2.7 2.7 3.4 ± 2.6
2.8 3.2 ± 2.4 2.3 3.3 ± 2.6 2.6 3.4 ± 2.5
Figure 2 Percentage of patients with use of concomitant pain medication.
of the placebo but 37% of the verum group suspected to have received verum (p = 0.073). Tolerability was rated by the patients as good or excellent by 98% of both, placebo and AG5 (rest did not answer the question). Mild local irritations occurred after single injections in 1 patient of the AG5 and 1 patient of the placebo group. Severe side effects did not occur. There was no systematic deviation and no significant difference between AG5 and placebo regarding the safety laboratory parameters.
Discussion AG5 is claimed to stimulate self-healing properties in patients with OA knee. We found, that in patients with moderate complaints from osteoarthritis of the knee injections of AG5 or placebo into painful trigger points of the knee resulted in clinically relevant and statistically significant improvements in all WOMAC scales. Superiority of AG5 beyond placebo, however, could not be demonstrated for the primary outcome parameter after 4 weeks of treatment. At follow up AG5 had an effect, because stop of pain medication was more frequent in the AG5 than in the placebo group and an improvement of the WOMAC scores could therefore have been masked. Regarding the major health risks
of NSAID the higher rate of stopping pain medication is a relevant outcome and deserves further investigations. Also the trend to a higher number of patients guessing correctly to have verum (p = 0.073) speaks for an effect of AG5. The ampoules were absolutely indistinguishable from appearance and injection did not cause local reactions which could deblind the patients. So this result could be explained either by chance or because the verum had an effect which is not covered by the questionnaires. Case reports with AG5 in OA patients18,19 suggest a rapid effect after 4 and 3 weeks, respectively which correlates to our results. Therefore, and because it was a recommendation of the manufacturer, we decided to treat for 4 weeks only. Stimulation of selfhealing properties might, however, need longer treatment for maximal effects. In studies with successful acupuncture treatment, also a stimulatory principle, and similar baseline pain scores as in our study22,23 the treatment duration was longer (8—13 weeks) than in our study. It was, however, not longer than in studies with successful symptomatic NSAID treatment and similar baseline pain scores (6—13 weeks, Table 3).24,25 From these data it is unlikely that stimulatory therapeutic principles need much longer time than symptomatic therapies. It cannot be excluded, however, that prolonged treatment would have shown better effects in our study but it is unlikely that differences between placebo and verum were missed after 4 weeks treatment and 8 weeks follow up. Despite randomization the treatment groups were not well balanced in baseline WOMAC scores. It should, however, been noted, that the estimated group differences were not affected by this imbalance because baseline scores entered all statistical models as a covariate. Irrespective of the medication the effects of subcutaneous injections were distinct and comparable to or even better than effects of NSAID in other studies with similar baseline pain scores24,25 and a little less than improvements with acupuncture (Table 3).22,23 Also the excellent tolerability and safety of the subcutaneous injections suggests their relevance as complementary or alternative symptom modifying treatment in OA patients. Numerous study patients of both groups asked because of their good experiences for repetition of the treatment in the months after the study was finished. Most improvements in this study were probably caused by strong unspecific effects. In studies on acupuncture or intraarticular injections ‘‘placebo’’ effects were considerably stronger than the effects of oral placebos (Table 3).26,27
Randomized, placebo controlled studies with primary osteoarthritis of the knee and similar baseline characteristics to our study.
Study
Groups
n
End of treatment
WOMAC baseline score
WOMAC score at end of treatment
Difference baseline-end of treatment
p-Value
Pincus et al.25
Celecoxib 200 mg daily vs. paracetamol 4× 1000 mg daily vs. placebo
524
6 weeks
1. Celecoxib: 4.9 ± 1.5 2. Paracetamol: 5.3 ± 1.5 3. Placebo: 5.0 ± 1.6
3.8 ± 1.5
−1.0
1. vs. 2. n.s.
4.4 ± 1.5
−0.8
2. vs. 3. n.s.
4.6 ± 1.7
−0.5
1. vs. 3. =0.002
3.5 ± 1.7
−1.4
1. vs. 4. <0.01
3.4 ± 1.7
−1.4
2. vs. 4. <0.01
3.5 ± 1.6
−1.3
3. vs. 4. <0.01
4.0 ± 1.6
−0.9
Tannenbaum et al.24
Lumiracoxib 200 mg vs. lumiracoxib 400 mg vs. celecoxib 200 mg vs. placebo
1702
13 weeks
1. Lumiracoxib 200: 9 ± 1.4 2. Lumiracoxib 400: 4.8 ± 1.5 3. Celecoxib: 4.9 ± 1.5 4. Placebo: 4.9 ± 1.4
Raynauld et al.26
Steroids vs. saline intraarticular every 3 months
68
1 year
1. Steroids: 4.0 ± 2.6 2. Placebo: 4.8 ± 2.8
2.6 ± 2.0 3.5 ± 1.9
−1.4 −1.3
0.87
Karlsson et al.27
Hyaluronic acid A vs. hyaluronic acid B. vs. saline intraarticular thrice per week
246
12 weeks
1. Hyaluronic acid A: 4.9 ± 1.3 2. Hyaluronic acid B: 4.9 ± 1.1 3. Placebo: 4.9 ± 1.3
3.5
−1.4
n.s.
3.2
−1.7
3.1
−1.8
1. Acupuncture: 5.1 ± 1.9 2. Minimal acupuncture: 5.3 ± 1.9 3. Waiting group: 5.2 ± 1.9 1. Acupuncture: 5.4 ± 2.0 2. Sham acupuncture: 5.5 ± 1.9 3. Standard therapy: 5.5 ± 2.0
2.7 ± 1
−2.4
1. vs. 2. <0.001
3.6 ± 2
−1.7
1. vs. 3. <0.001
5.0 ± 2
−0.2
Witt et al.23
Scharf et al.22
12 sessions acupuncture vs. minimal acupuncture vs. waiting group
10 sessions acupuncture vs. sham acupuncture vs. standard therapy with NSAID
300
1039
8 weeks
13 weeks
−2.1
1. vs. 3. <0.001
−1.9
2. vs. 3 <0.001
−0.9
1. vs. 2. =0.51
Effectiveness of subcutaneous injections of a cartilage preparation in osteoarthritis of the knee
Table 3
vs.: versus and n.s.: not significant.
117
118 This corroborates the hypothesis that injections have stronger placebo effects than oral medications.28 Effects of injections to the knee in osteoarthritis patients may however not merely be attributed to psychological placebo effects. Probably more appropriate they can be classified as unspecific effects, because injections to the knee can affect pain mediating structures as pointed out in introduction. Such unspecific effects could have occurred in this study by penetrating the skin, counter irritation, application of 1 ml fluid subcutaneously and hereby neuronal or metabolic changes at trigger or acupuncture points or other pain mediation structures of the knee. This is also supported by the acupuncture studies which suggested that sham acupuncture at the knee was more effective than sham acupuncture at points distant from the knee.22,23 In conclusion, injections to the knee were accompanied by strong and clinical relevant improvements of the complaints of knee osteoarthritis in both, AG5 and verum group. For the primary outcome parameter pain AG5 was not superior to placebo after 4 weeks of treatment but the equal reduction of pain in combination with higher reduction of pain medication suggests an effect of the preparation. Because subcutaneous injections to the knee are a safe and cheap procedure, their clinical relevance should be further investigated in comparison to standard therapies and acupuncture. AG5 should be further investigated in studies with longer treatment duration.
Conflict of interest The authors declare that there is no conflict of interest.
Acknowledgment Medication and financial compensation for the study was kindly provided by WALA Heilmittel GmbH, Bad Boll, Germany.
References 1. Engelhardt M. Epidemiology of osteoarthritis in Western Europe. Dtsch Z Sportmed 2003;54(6):171—3. 2. Sprangers MA, de Regt EB, Andries F, van Agt HM, Bijl RV, de Boer JB, et al. Which chronic conditions are associated with better or poorer quality of life? J Clin Epidemiol 2000;53(9):895—907. 3. Kessler S. Behandlungsstrategien der Gonarthrose. Med Welt 2003;54:12—7. 4. Gurwitz JH, Avorn J. The ambiguous relation between aging and adverse drug reactions. Ann Intern Med 1991;114(11):956—66. 5. Tramèr MR, Moore RA, Reynolds DJ, McQuay HJ. Quantitative estimation of rare adverse events which follow a biological progression: a new model applied to chronic NSAID use. Pain 2000;85(1—2):169—82. 6. Felson DT. The sources of pain in knee osteoarthritis. Curr Opin Rheumatol 2005;17(5):624—8. 7. Felson DT. Developments in the clinical understanding of osteoarthritis. Arthritis Res Ther 2009;11(1):203. 8. Sulzbacher I. Arthrosis-histology and pathogenetic approaches. Radiologe 2000;40(12):1126—33. 9. Fishkin Z, Miller D, Ritter C, Ziv I. Changes in human knee ligament stiffness secondary to osteoarthritis. J Orthop Res 2002;20(2):204—7.
R. Huber et al. 10. Slobodin G, Rozenbaum M, Boulman N, Rosner I. Varied presentations of enthesopathy. Semin Arthritis Rheum 2007;37(2):119—26. 11. Weiler HT, Pap G, Awiszus F. The role of joint afferents in sensory processing in osteoarthritic knees. Rheumatology (Oxford) 2000;39(8):850—6. 12. Bajaj P, Bajaj P, Graven-Nielsen T, Arendt-Nielsen L. Osteoarthritis and its association with muscle hyperalgesia: an experimental controlled study. Pain 2001;93(2):107—14. 13. Scott NA, Guo B, Barton PM, Gerwin RD. Trigger point injections for chronic non-malignant musculoskeletal pain: a systematic review. Pain Med 2009;10(1):54—69. 14. Sickel K. Plenosol in the treatment of arthrotic changes. Z Allgemeinmed 1971;47(31):1667—9. 15. Roemer F, Roggatz M, Soldner G, Sommer M, Vogel HH. Anthroposophische Medizin in der Praxis 2. 1st ed. Natur Mensch Medizin Verlags GmbH Bad Boll; 2002. 16. Barron MC, Rubin BR. Managing osteoarthritic knee pain. J Am Osteopath Assoc 2007;107(10 Suppl. 6):ES21—7. 17. Roemer F, Sommer M. Zur Bedeutung der potenzierten Organpräparate in der anthroposophischen Therapierichtung. Der Merkurstab; 1999. Sonderheft: 2—5. 18. Roemer F. Kasuistik fortgeschrittene Valgusgonarthrose. Der Merkurstab; 1999. Sonderheft: 33—34. 19. Roemer F. Kasuistik entzündlicher Reizzustand mit massivem Kniegelenkserguss auf der Grundlage einer Gonarthrose. Der Merkurstab; 1999. Sonderheft: 35. 20. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. 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 1988;15(12): 1833—40. 21. Stucki G, Meier D, Stucki S, Michel BA, Tyndall AG, Dick W, et al. Evaluation of a German version of WOMAC (Western Ontario and McMaster Universities) Arthrosis Index. Z Rheumatol 1996;55(1):40—9. 22. Scharf HP, Mansmann U, Streitberger K, Witte S, Krämer J, Maier C, et al. Acupuncture and knee osteoarthritis: a three-armed randomized trial. Ann Intern Med 2006;145(1): 12—20. 23. Witt C, Brinkhaus B, Jena S, Linde K, Streng A, Wagenpfeil S, et al. Acupuncture in patients with osteoarthritis of the knee: a randomised trial. Lancet 2005;366(9480):136—43. 24. Tannenbaum H, Berenbaum F, Reginster JY, Zacher J, Robinson J, Poor G, et al. Lumiracoxib is effective in the treatment of osteoarthritis of the knee: a 13 week, randomised, double blind study versus placebo and celecoxib. Ann Rheum Dis 2004;63(11):1419—26. 25. Pincus T, Koch G, Lei H, Mangal B, Sokka T, Moskowitz R, et al. Patient preference for placebo. Acetaminophen (paracetamol) or Celecoxib Efficacy Studies (PACES): two randomised, double blind, placebo controlled, crossover clinical trials in patients with knee or hip osteoarthritis. Ann Rheum Dis 2004;63(8):931—9. 26. Raynauld JP, Buckland-Wright C, Ward R, Choquette D, Haraoui B, Martel-Pelletier J, et al. Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial. Arthritis Rheum 2003;48(2):370—7. 27. Karlsson J, Sjögren LS, Lohmander LS. Comparison of two hyaluronan drugs and placebo in patients with knee osteoarthritis. A controlled, randomized, double-blind, parallel-design multicentre study. Rheumatology (Oxford) 2002;41(11):1240—8. 28. de Craen AJ, Tijssen JG, de Gans J, Kleijnen J. Placebo effect in the acute treatment of migraine: subcutaneous placebos are better than oral placebos. J Neurol 2000;247(3):183—8.