1cleg_oa.indd

Glucosamine, Chondroitin Sulfate, and the Two in Combination Daniel O. Clegg, M.D., Domenic J. Reda, Ph.D., Crystal L. Harris, Pharm.D., Marguerite A. Klein, M.S., James R. O’Dell, M.D., Michele M. Hooper, M.D., John D. Bradley, M.D., Clifton O. Bingham III, M.D., Michael H. Weisman, M.D., Christopher G. Jackson, M.D., Nancy E. Lane, M.D., John J. Cush, M.D., Larry W. Moreland, M.D., H. Ralph Schumacher, Jr., M.D., Chester V. Oddis, M.D., Frederick Wolfe, M.D., Jerry A. Molitor, M.D., David E. Yocum, M.D., Thomas J. Schnitzer, M.D., Daniel E. Furst, M.D., Allen D. Sawitzke, M.D., Helen Shi, M.S., Kenneth D. Brandt, M.D., Roland W. Moskowitz, M.D., and H. James Williams, M.D.
Background
Glucosamine and chondroitin sulfate are used to treat osteoarthritis. The multi- From the University of Utah School of Medicine, Salt Lake City (D.O.C., C.G.J., center, double-blind, placebo- and celecoxib-controlled Glucosamine/chondroitin A.D.S., H.J.W.); the Hines Veterans Arthritis Intervention Trial (GAIT) evaluated their efficacy and safety as a treatment Affairs Cooperative Studies Program for knee pain from osteoarthritis.
Coordinating Center, Hines, Ill. (D.J.R., H.S.); the Clinical Research Pharmacy Co ordinating Center, Albuquerque, N.M. (C.L.H.); the National Center for Com- We randomly assigned 1583 patients with symptomatic knee osteoarthritis to re- plementary and Alternative Medicine, ceive 1500 mg of glucosamine daily, 1200 mg of chondroitin sulfate daily, both National Institutes of Health, Bethesda, glucosa mine and chondroitin sulfate, 200 mg of celecoxib daily, or placebo for 24 Md. (M.A.K.); the University of Nebraska weeks. Up to 4000 mg of acetaminophen daily was allowed as rescue analgesia. Western Reserve University, Cleveland Assignment was stratified according to the severity of knee pain (mild [N = 1229] (M.M.H., R.W.M.); the Indiana University vs. moderate to severe [N = 354]). The primary outcome measure was a 20 percent School of Medicine, Indianapolis (J.D.B., K.D.B.); the Hospital for Joint Diseases, decrease in knee pain from baseline to week 24.
(C.O.B.); Cedars-Sinai Medical Center, Los Angeles (M.H.W.); the University The mean age of the patients was 59 years, and 64 percent were women. Overall, glu- of California, San Francisco (N.E.L.); cosamine and chondroitin sulfate were not significantly better than placebo in re- the Presbyterian Hospital of Dallas, Dallas (J.J.C.); the University of Alabama, ducing knee pain by 20 percent. As compared with the rate of response to placebo Birmingham (L.W.M.); the Hospital of the (60.1 percent), the rate of response to glucosamine was 3.9 percentage points higher University of Pennsylvania, Philadelphia (P = 0.30), the rate of response to chondroitin sulfate was 5.3 percentage points high- (H.R.S.); the University of Pittsburgh er (P = 0.17), and the rate of response to combined treatment was 6.5 percentage points Clinical Centers, Wichita, Kans. (F.W.); higher (P = 0.09). The rate of response in the celecoxib control group was 10.0 percent- the Virginia Mason Medical Center, age points higher than that in the placebo control group (P = 0.008). For patients Seattle (J.A.M.); the University of Arizona, with moderate-to-severe pain at baseline, the rate of response was significantly high- Northwestern University, Chicago (T.J.S.); er with combined therapy than with placebo (79.2 percent vs. 54.3 percent, P = 0.002). and the Uni versity of California, Los Adverse events were mild, infrequent, and evenly distributed among the groups.
Angeles (D.E.F.). Address reprint requests to Dr. Clegg at the University of Utah, Conclusions
Division of Rheumatology, 4B200 School of Medicine, 50 N. Medical Dr., Salt Lake Glucosamine and chondroitin sulfate alone or in combination did not reduce pain City, UT 84132, or at gait.study@hsc.
effectively in the overall group of patients with osteoarthritis of the knee. Explor- utah.edu.
atory analyses suggest that the combination of glucosamine and chondroitin sul- fate may be effective in the subgroup of patients with moderate-to-severe knee pain. Copyright 2006 Massachusetts Medical Society. (ClinicalTrials.gov number, NCT00032890.) n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Osteoarthritis is the most common The institutional review board of each participat- of the arthritides, affecting at least 20 mil- ing center approved the study, and all patients gave lion Americans, a number that is expect- written informed consent. Patients’ race or ethnic ed to double over the next two decades.1,2 Cur- group was self-reported.
rently available medical therapies primarily address
the treatment of joint pain in patients with osteo- Treatment Regimens
arthritis.3 Analgesics as well as traditional and Eligible patients were randomly assigned with the
cyclooxygenase-2–selective nonsteroidal antiin- use of a double-dummy scheme to one of five
flammatory drugs (NSAIDs) have suboptimal ef- orally administered treatments: 500 mg of glucos-
fectiveness,4,5 and there is some question about amine hydrochloride three times daily, 400 mg
their safety, especially in the light of recent reports of sodium chondroitin sulfate three times daily,
of increased cardiovascular risk.6-8
500 mg of glucosamine plus 400 mg of chondroi- The dietary supplements glucosamine and tin sulfate three times daily, 200 mg of celecoxib chondroitin sulfate have been advocated, especial- (Celebrex, Pfizer) daily, or placebo. Permuted-block ly in the lay media, as safe and effective options randomization was used with random block sizes, for the management of symptoms of osteoarthri- stratified according to the 16 clinical centers and tis. A meta-analysis of studies evaluating the ef- baseline WOMAC pain stratum (mild, defined as ficacy of these supplements for osteoarthritis9 a score of 125 to 300, or moderate to severe, de-suggested potential benefit from these agents but fined as a score of 301 to 400; scores on this raised questions about the scientific quality of the scale can range from 0 to 500). The randomiza-studies. We conducted the Glucosamine/chondroi- tion code list was developed by the Veterans Af-tin Arthritis Intervention Trial (GAIT), a 24-week, fairs Cooperative Studies Program Data Coordi-randomized, double-blind, placebo- and celecoxib- nating Center in Hines, Illinois. During data controlled, multicenter trial sponsored by the collection, neither the clinical centers nor the National Institutes of Health, to evaluate rigor- coordinating center at the University of Utah had ously the efficacy and safety of glucosamine, access to the randomization codes or statistical chondroitin sulfate, and the two in combination summaries of follow-up data. Patients were al-in the treatment of pain due to osteoarthritis of lowed to take up to 4000 mg of acetaminophen the knee.
(Tylenol, McNeil) daily, except during the 24 hours before a clinical evaluation for joint pain. Other analgesics, including narcotics and NSAIDs, were not permitted. Patients were evaluated at baseline Patients
and 4, 8, 16, and 24 weeks after randomization.
Eligible patients were at least 40 years of age and
had clinical evidence (knee pain for at least six Outcome Measures
months and on the majority of days during the The primary outcome measure was a response to
preceding month) and radiographic evidence treatment, defined a priori by expert consensus
(tibiofemoral osteophytes of at least 1 mm [Kell- as a 20 percent decrease in the summed score for
gren and Lawrence grade 2 or 3]) of osteoarthri- the WOMAC pain subscale from baseline to week
tis.10 Patients had to have a summed pain score of 24. Secondary outcome measures, selected a pri-
125 to 400 on the index (more symptomatic) knee ori in accordance with the preliminary recommen-
according to the Western Ontario and McMaster dations of the Osteoarthritis Research Society
Universities Osteoarthritis Index (WOMAC)11,12 International (OARSI) task force,14 included the
and to be in American Rheumatism Association following: scores for the stiffness and function
functional class I, II, or III.13 Patients were ex- subscales of WOMAC; the patient’s global assess-
cluded if they had concurrent medical or arthritic ments of disease status and response to therapy,
conditions that could confound evaluation of the obtained with the use of a 100-mm visual-ana-
index joint, predominant patellofemoral disease, logue scale on which higher scores indicate more
a history of clinically significant trauma or sur- severe disease; the investigator’s global assess-
gery to the index knee, or coexisting disease that ment of disease status, obtained with the use of
could preclude successful completion of the trial. a 100-mm visual-analogue scale; the presence or
n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is absence of soft-tissue swelling, effusion, or both were purchased and overencapsulated (for mask-
in the index knee; scores on the Medical Outcomes ing) and a matching placebo was prepared.
Study 36-item Short-Form General Health Survey
(SF-36), which reflect the health-related quality Adverse Events
of life15; scores on the Health Assessment Ques- Adverse events and serious adverse events were as-
tionnaire, which reflect physical function16; and sessed by the investigator at each study visit and
acetaminophen use, according to diary entries and followed until resolution. Safety monitoring in-
tablet counts. All outcome measures were assessed cluded complete blood counts; measurement of
at each study visit, except for the patient’s global serum aspartate aminotransferase, alanine ami-
assessment of response to therapy, which was as- notransferase, glucose, creatinine, and partial-
sessed only after randomization.
thromboplastin time; and urinalysis at each study In May 2004, the Outcome Measures in Rheu- visit. Specific cardiovascular monitoring for adverse matology Clinical Trials (OMERACT) and OARSI events was not done. Patients with abnormal blood
published their criteria for a response to treat- glucose results had blood glucose levels measured
ment for osteoarthritis.17 A response was classi- after an overnight fast. In patients with diabetes
fied as an improvement in pain or function of at at enrollment, fasting blood glucose and glycosyl-
least 50 percent and a decrease of at least 20 mm ated hemoglobin levels were monitored. A test for
on the visual-analogue scale for pain or function fecal occult blood (Hemoccult, Beckman Coulter)
or the occurrence of at least two of the follow- was performed at the visit at week 24. Medication
ing: a decrease in pain of at least 20 percent and was withdrawn from patients in whom diabetes
at least 10 mm on the visual-analogue scale; an or gastrointestinal bleeding developed, and the
improvement in function of at least 20 percent patients were referred for further evaluation.
and a decrease of at least 10 mm on the visual-
analogue scale; and an increase in the patient’s Statistical Analysis
global assessment score by at least 20 percent and An absolute increase in the response rate of 15
at least 10 mm on the visual-analogue scale. Since percent, as compared with the rate in the placebo
we prospectively collected data on each compo- group, was considered to indicate a clinically
nent, the OMERACT–OARSI response rate is also meaningful treatment effect. We estimated that
reported.
1588 patients would need to be enrolled to pro-vide the study with a statistical power of 85 per- Product Selection
cent to detect one or more clinically meaningful Our study was conducted under an investigational differences between the placebo group and the new drug application, and the study agents were glucosamine group, the chondroitin sulfate group, subject to pharmaceutical regulation by the Food and the combined-treatment group, assuming a and Drug Administration (FDA). The Cooperative rate of response of 35 percent in the placebo group Studies Program Clinical Research Pharmacy Co- and a withdrawal rate of 20 percent. Pairwise ordinating Center, a facility licensed by the FDA, comparisons of the glucosamine group, the chon-used a vendor-certification program to evaluate droitin sulfate group, and the combined-treatment available commercial products and raw materials group with the placebo group were made with in order to select the suppliers of glucosamine the use of a two-sided chi-square test with an and chondroitin sulfate. Donated or purchased α value of 0.017 for each comparison (overall ingredients were tested for purity, potency, and α value, 0.05). A side comparison between cele-quality. Certificates of analysis were obtained for coxib and placebo also used an α value of 0.017. the agents, and Drug Master Files were on file The data and safety monitoring board reviewed with the FDA. Capsules containing 250 mg of glu- study performance and safety data annually but did cosamine hydrochloride, 200 mg of sodium chon- not conduct interim monitoring of the primary droitin sulfate, the two in combination, and match- outcome. Analysis of the primary outcome mea-ing placebo were manufactured, distributed, and sure was conducted according to the intention placed on a shelf-life–stability program through- to treat.
out the study at the Pharmacy Coordinating Cen- Analyses of the secondary outcome measures ter. In addition, 200-mg capsules of celecoxib followed the pairwise-comparison plan described n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e above. The chi-square test was used to compare tests were two-sided. We used SAS software (ver-categorical data. The t-test for independent groups sion 8) for all statistical analyses.18was used to compare changes between groups in quantitative data from baseline to the end of follow-up. A total of 71 patients who did not at-
tend any follow-up visits were classified as hav- Characteristics of the Patients
ing no response for the primary outcome measure, Recruitment began November 29, 2000, at 13 clin-
the OMERACT–OARSI response, and a response ical centers, and 3 centers were added to the study
based on a 50 percent reduction in the score for in February 2003 to ensure timely recruitment.
the WOMAC pain subscale. These patients were The study was completed on July 8, 2004. A total
excluded from the analyses of all other secondary of 3238 patients were screened, and 1583 under-
outcomes. We used the last-observation-carried- went randomization (Fig. 1). The most common
forward method in the analysis of all outcomes reasons for exclusion were an inability to meet
among patients who made at least one follow-up radiographic criteria (in 1089 patients) and a
visit but who did not complete follow-up.
WOMAC pain score of less than 125 or more than We also analyzed the results according to the 400 (in 321 patients). The majority of patients WOMAC pain stratum, since logistic-regression were women (64.1 percent), with a mean age of analysis showed a significant (P = 0.008) interac- 58.6 years and a mean body-mass index (the weight tion between treatment and pain stratum in the in kilograms divided by the square of the height comparison of combined treatment with placebo in meters) of 31.7. The groups were well balanced for the primary outcome measure. All statistical at baseline (Table 1). The withdrawal rate of 20.5 Figure 1. Enrollment and Outcomes.
n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is Table 1. Baseline Characteristics of the Patients.*
Chondroitin
Glucosamine +
Glucosamine
Chondroitin Sulfate
Celecoxib
Characteristic
Duration of osteoarthritis symptoms — yr Time since diagnosis of osteoarthritis — yr Kellgren and Lawrence radiographic reading Global assessment of disease status score Joint swelling, effusion, or both on clinical * Plus–minus values are means ±SD. Race or ethnic group was self-determined. The body-mass index is the weight in kilograms divided by the square of the height in meters. Patient’s and physician’s global assessment scores can range from 0 to 100. Scores for the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) can range from 0 to 500 for the pain subscale, 0 to 200 for the stiffness subscale, and 0 to 1700 for the function subscale; normalized scores can range from 0 to 300. Scores for the Health Assessment Question-naire can range from 0 to 3 for the Alternative Disability portion and from 0 to 100 for the Pain portion. Scores for the Medical Outcomes Study 36-item Short-Form General Health Survey (SF-36) can range from 4 to 71 for the physical component and from 2 to 74 for the mental component. For all instruments, higher scores indicate more severe disease. ARA denotes American Rheumatism Association (now called the American College of Rheumatology).
n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e percent did not differ significantly among the the primary outcome measure revealed that the groups and was similar to the predicted rate of 20 rate of response to glucosamine and chondroitin percent. Adherence to the assigned treatment reg- sulfate, either alone or in combination, was not imen, measured by capsule count at each visit, significantly higher than the rate of response to ranged from 88.8 percent to 97.0 percent.
placebo. As compared with the rate of response to placebo, the rate of response to chondroitin Clinical Outcomes
sulfate was 5.3 percentage points higher (P = 0.17), Figure 2 shows the relative likelihood of a response the rate of response to glucosamine was 3.9 per-and 98.3 percent confidence intervals (correspond- centage points higher (P = 0.30), and the rate of ing to the use of criteria in which a P value of less response to the combination of glucosamine and than 0.017, rather than less than 0.05, indicated chondroitin sulfate was 6.5 percentage points statistical significance, owing to multiple com- higher (P = 0.09). The rate of response to the cele-parisons) for the total study population as well as coxib control was 10.0 percentage points higher both pain strata for each group, as compared than that for the placebo control (P = 0.008). The with the placebo group. Results of primary and OMERACT–OARSI response rates showed a sim-secondary outcome measures for the entire study ilar pattern, with differences between the placebo population and each WOMAC pain stratum are group and the glucosamine, chondroitin sulfate, given in Table 2.
and combined-treatment groups not reaching sig- Overall, differences between placebo and the nificance. As compared with the rate of response various agents were relatively small. Analysis of to placebo, the rate of response to chondroitin Variable
Likelihood (98.3% confidence interval)
All randomized patientsPatients with mild pain to-severe pain (WOMACpain score, 301–400) All randomized patientsPatients with mild pain to-severe pain (WOMACpain score, 301–400) Glucosamine
Chondroitin
Glucosamine+
Celecoxib Better
Sulfate Better
Chondroitin
Sulfate Better
Figure 2. Pairwise Comparisons of the Overall Likelihood of a Response.
Scores for the pain subscale of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) can range from 0 to
500, with higher scores indicating more pain. A response according to the guidelines of the Outcome Measures in Rheumatology Clini-
cal Trials and Osteoarthritis Research Society International (OMERACT–OARSI) was classified as an improvement in function or pain of
at least 50 percent and a decrease of at least 20 mm on the visual-analogue scale for pain or function or the occurrence of at least two
of the following: a decrease in pain of at least 20 percent and at least 10 mm on the visual-analogue scale; an improvement in function
of at least 20 percent and a decrease of at least 10 mm on the visual-analogue scale; and an increase in the patient’s global assessment
score by at least 20 percent and at least 10 mm on the visual-analogue scale.
n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is Table 2. Primary and Secondary Outcomes.*
Glucosamine +
Chondroitin
Chondroitin
Glucosamine
Celecoxib
All randomized patients
Primary outcome: 20% decrease in WOMAC pain score Patient’s global assessment of response to therapy score n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Table 2. (Continued.)
Glucosamine +
Chondroitin
Chondroitin
Glucosamine
Celecoxib
Patient’s global assessment of disease status score Physician’s global assessment of disease status Joint swelling, effusion, or both on clinical examination At end of follow-up — no./total no. (%) Patients with moderate-to-severe pain (WOMAC pain score, 301–400)
Primary outcome: 20% decrease in WOMAC pain score n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is Table 2. (Continued.)
Glucosamine +
Chondroitin
Chondroitin
Glucosamine
Celecoxib
Patient’s global assessment of response to therapy Patient’s global assessment of disease status score Physician’s global assessment of disease status score Joint swelling, effusion, or both on clinical examination At end of follow-up — no./total no. (%) Patients with mild pain (WOMAC pain score, 125–300)
Primary outcome: 20% decrease in WOMAC pain score n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e Table 2. (Continued.)
Glucosamine +
Chondroitin
Chondroitin
Glucosamine
Celecoxib
sulfate was 6.6 percentage points higher (P = 0.09), cebo group. None of the differences were signifi-the rate of response to glucosamine was 3.7 per- cant. Treatment effects in the moderate-to-severe centage points higher (P = 0.35), and the rate of pain stratum were more substantial. Results for response to combined treatment was 8.7 percent- the primary outcome in this stratum, which in-age points higher (P = 0.02). With the exception cluded 22 percent of the patients in the trial, of the incidence of joint swelling, effusion, or indicated that combined treatment was signifi-both, for the secondary outcome measures, there cantly more effective than placebo (24.9 percent-were no significant differences between the pla- age points higher, P = 0.002). As compared with cebo group and the glucosamine, chondroitin placebo, however, celecoxib (difference, 15.1 per-sulfate, or combined-treatment groups.
centage points; P = 0.06), glucosamine (difference, Analysis of the primary outcome in the sub- 11.4 percentage points; P = 0.17), and chondroitin group of patients with mild pain showed even sulfate (difference, 7.1 percentage points; P = 0.39) smaller treatment effects, with the rate of re- were not significantly better. Similarly, the OMER-sponse ranging from 8.6 percentage points high- ACT–OARSI response rate ranged from 26.4 per-er in the celecoxib group to 1.9 percentage points centage points higher with combined treatment higher in the glucosamine group than in the pla- (P = 0.001) to 10.0 percentage points higher with n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is Table 2. (Continued.)
Glucosamine +
Chondroitin
Chondroitin
Glucosamine
Celecoxib
Patient’s global assessment of response to therapy score Patient’s global assessment of disease status score Physician’s global assessment of disease status score Joint swelling, effusion, or both on clinical examination At end of follow-up — no./total no. (%) * Plus–minus values are means ±SD. All P values are for the comparison with the placebo group. WOMAC denotes Western Ontario and McMaster Universities Osteoarthritis Index, OMERACT–OARSI Outcome Measures in Rheumatology Clinical Trials–Osteoarthritis Research Society, and HAQ Health Assessment Questionnaire.
† P≤0.017 for the comparison with placebo.
‡ P≤0.05 for the comparison with placebo.
chondroitin sulfate (P = 0.24), as compared with the placebo group, the celecoxib group had a placebo.
significantly lower incidence of headache and Overall, the rate of use of rescue acetamino- nausea but had a nonsignificant but higher inci- phen was low (1.6 to 1.9 tablets per day) (Table 2). dence of increased blood pressure. Patients who There were no significant differences in the use received chondroitin sulfate had the highest inci-of acetaminophen among the groups for all ran- dence of “musculoskeletal and connective-tissue” domized patients or within each pain stratum.
events and the lowest incidence of vomiting. Because of concern about the possibility of ische- Adverse Events
mic cardiovascular events with the use of selec- Seventy-seven serious adverse events were reported tive cyclooxygenase-2 inhibitors, the data and in 61 patients. Three serious adverse events were safety monitoring board requested an interim re-judged by the investigator to be related to study view of adverse events. Although the celecoxib treatment: congestive heart failure (in a patient re- group had a nonsignificant but higher incidence ceiving combined treatment), stroke (in a patient of “cardiac” events than the other four groups, receiving celecoxib), and chest pain (in a patient these events were predominantly arrhythmias receiving glucosamine). There were no serious gas- (palpitations and atrial fibrillation), rather than trointestinal adverse events or deaths. The num- ischemic events.
ber of patients who withdrew because of adverse events was similar among the groups (Fig. 1).
Time to Response
Adverse events were generally mild and evenly Figure 3 shows the percentage of patients with a distributed among the groups. As compared with primary response in each group at weeks 4 and 24. n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. T h e n e w e n g l a n d j o u r n a l o f m e d i c i n e tive effect in the subgroup of patients with mod- erate-to-severe pain cannot be excluded, since the difference from placebo in the OMERACT–OARSI response rate approached significance in this group. Treatment with chondroitin sulfate was associated with a significant decrease in the inci- dence of joint swelling, effusion, or both. We did not find an increased risk of ischemic cardiovas- cular events among patients who received cele- coxib or among patients with diabetes who received glucosamine, but this study was not powered to Rate of Primary Response (%)
Our study has a number of limitations. First, the high rate of response to placebo (60.1 per- cent) and the relatively mild degree of pain from osteoarthritis among the participants may have Figure 3. Rates of a Primary Response in the Five Groups at 4 and 24
limited our ability to detect benefits of the treat- ments. Elevated rates of response to placebo have A primary response was defined as a 20 percent decrease in the summed been reported in other osteoarthritis trials20,21 score for the pain subscale of the Western Ontario and McMaster Universi- and may relate, in part, to patients’ biases and expectations and to the enrollment of patients with relatively mild symptoms of osteoarthritis. The onset of pain relief was fastest in the cele- In addition, our patients had relatively mild knee coxib group, with substantial improvement at four pain at baseline, as compared with that in classic weeks. The other four groups had more gradual studies of osteoarthritis, in which a criterion for improvement.
entry was a disease flare after the discontinuation of NSAIDs.22,23 Widely used outcome measures for osteoarthritis treatments may be insensitive in identifying improvement, making it difficult Osteoarthritis is the most common form of ar- to identify improvement in patients with mild thritis in the United States and has a major effect symptoms. The OMERACT–OARSI response out-on the health-related quality of life. In 2004, the come seemed to perform best in the face of these estimated direct and indirect medical costs as- challenges (as it was designed to do). Thus, it is sociated with all forms of arthritis exceeded $86 not clear whether the small but consistently posi-billion.2 Glucosamine and chondroitin sulfate are tive differences between groups in some of the the most widely used dietary supplements for analyses represent clinically meaningful effects osteoarthritis, with estimated sales in 2004 ap- obscured by the factors outlined above or effects proaching $730 million.19 GAIT was designed to of marginal clinical value. However, even the ef-evaluate rigorously the efficacy of glucosamine, fects of celecoxib were smaller than those seen chondroitin sulfate, and the two in combination in other studies.24in treating knee pain related to osteoarthritis. Treatment effects were more substantial in the The analysis of the primary outcome measure did subgroup of patients with moderate-to-severe not show that either supplement, alone or in com- pain, but the relatively small numbers of patients bination, was efficacious. Analysis of the prespec- in this subgroup may have limited the study’s ified subgroup of patients with moderate-to- power to demonstrate significant benefits in the severe pain demonstrated that combination therapy glucosamine, chondroitin sulfate, and celecoxib significantly decreased knee pain related to osteo- groups. For example, as compared with placebo, arthritis, as measured by the primary outcome or celecoxib therapy was associated with a clinically by the OMERACT–OARSI response rate. We did meaningful difference in the primary outcome not identify significant benefits associated with measure of 15 percentage points, but the differ-the use of glucosamine or chondroitin sulfate ence did not reach statistical significance.
alone. Although the results for glucosamine did Several studies have evaluated the efficacy of not reach significance, the possibility of a posi- glucosamine25-29 and chondroitin sulfate.30,31 Some n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is have demonstrated efficacy but have been criti- and chondroitin sulfate may have some efficacy cized as having flaws, which were addressed in in patients with moderate-to-severe symptoms is the design of GAIT, such as the failure to adhere interesting but must be confirmed by another to the intention-to-treat principle, the enrollment trial. In making therapeutic decisions, physicians of small numbers of patients, potential bias re- and patients alike should be aware of our data lated to sponsorship of the study by the manu- suggesting that celecoxib has a much faster time facturers of the dietary supplements, and inad- to response than glucosamine, chondroitin sulfate, equate masking of the study agent. In general, or the two in combination. Continuing research is these studies have recruited patients with lower needed to establish the potential efficacy and in-levels of knee pain26-29 and failed to show im- crease our understanding of the biology, pharma-provement in WOMAC pain scores.32 However, cology, and pharmacokinetics of these agents.
in some instances,26,27 benefits of glucosamine Supported by a contract from the National Center for Comple- have been demonstrated with the use of other mentary and Alternative Medicine and the National Institute of outcome measures.
Arthritis and Musculoskeletal and Skin Diseases (NO1AR92236).
Drs. Bingham, Brandt, Clegg, Hooper, and Schnitzer report In the United States, glucosamine and chon- having received consulting fees or having served on advisory droitin sulfate are considered dietary supplements boards for McNeil Consumer and Specialty Pharmaceuticals. and are not held to the stringent standards of Drs. Brandt, Moskowitz, Schnitzer, and Schumacher report hav- ing received consulting fees or having served on advisory boards pharmaceutical manufacture. If these agents are for Pfizer. Dr. Brandt reports having equity interests in Pfizer. to be widely used for the treatment of osteoar- Drs. Moskowitz and Weisman report having received lecture thritis, serious consideration must be given to fees from Pfizer; Dr. Brandt, lecture fees from McNeil Consumer and Specialty Pharmaceuticals; Drs. Bingham, Clegg, Hooper, their current regulatory status in order to ensure Jackson, Molitor, Sawitzke, and Schnitzer, grant support from potency and purity. Studies have demonstrated Pfizer; and Dr. Bingham, grant support from McNeil Consumer substantial variation between the content listed and Specialty Pharmaceuticals. Dr. Brandt reports having received royalties from books related to osteoarthritis. Dr. Moskowitz on the labels of these products and the actual reports having served as an expert consultant for Pfizer. No content.33-35 Because our study was conducted other potential conflict of interest relevant to this article was under pharmaceutical rather than dietary-supple- reported.
We are indebted to McNeil Consumer and Specialty Pharma- ment regulations, agents identical to the ones we ceuticals, Fort Washington, Pa., for donating acetaminophen; to used may not be commercially available.
Bioiberica, S.A., Barcelona, for donating sodium chondroitin How should our results affect the treatment sulfate; and to Ferro Pfanstiehl Laboratories, Waukegan, Ill., for donating a portion of the glucosamine hydrochloride through of symptomatic osteoarthritis of the knee? Our Wilke Resources.
finding that the combination of glucosamine appendix
In addition to the authors, the following investigators participated in GAIT: National Center for Complementary and Alternative Medi-cine/National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Md. — S. Serrate-Sztein, E. Webster-Cissel; Steering Committee — N. Bellamy, G. Cannon, D. McClain, J. Sandy, A. Sawitzke, J. Theodosakis; Univer-sity of Utah Coordinating Center, Salt Lake City — G. Cannon, J. Christensen, M. Finco, D. Kumeroski, M. Lewandowski, K. Mara, J. Mendivil, A. Sawitzke; Veterans Affairs (VA) Cooperative Studies Program Coordinating Center, Hines, Ill. — M. Abdellatif, D. Motyka, J. Motyka, T. Nydam, M. Reinhard; VA Cooperative Studies Program Pharmacy Coordinating Center, Albuquerque, N.M. — J. Barnhill, C. Fye, W. Gagne; University of Nebraska, Omaha — A. Burch, D. O’Grady; Case Western Reserve University, Cleveland — M. Hooper, M. Lesko; Indiana University, Indianapolis — D. Ang, R. Grau, H. O’Brien, B. Shultz, A. Schaffter; Hospital for Joint Diseases, New York — V. Abellana, P. Rosenthal; Cedars-Sinai Medical Center, Los Angeles — D. Silver, S. Choudry, K. Gilley, A. Guieliano, C. Joseph, A. Tiwari; University of Utah, Salt Lake City — K. Cooper, K. Fredley, A. Kim, A. Portmann, S. Smith; University of California, San Fran-cisco, San Francisco — T. Munoz, N. Palmetto, R. Stuart-Thiessen, T. Solomon; Presbyterian Hospital, Dallas — E. Barnboym, J. Perla; University of Alabama, Birmingham, Birmingham — L. Heck, E. Coffey, M. Robertson, K. Woods; University of Pennsylvania, Philadel-phia — J. Dinnella, H. Hannah, C. Shaw, C. Smith; University of Pittsburgh, Pittsburgh — S. Manzi, M. Graham, J. Jablon; Arthritis Research and Clinical Centers, Wichita, Kans. — T. Pryor, R. Schumacher, E. Sparr, K. Urbansky; Virginia Mason Medical Center, Seattle — C. Edwards, A. Mondt, E. Voon, K. Yim; University of Arizona, Tucson — C. Bush, J. Fuessler, J. Sonder; Northwestern University, Chicago — D. Czech, L. Robinson; University of California, Los Angeles, Los Angeles — D. Khanna, W. Chen, E. Hasan, L. Woolcock.
Refe renc e s
1. Lawrence RC, Helmick CG, Arnett FC, 2. United States Senate Committee on partment of Health and Human Services,
et al. Estimates of the prevalence of ar-
Health, Education, Labor and Pensions, 2004.
thritis and selected musculoskeletal dis- 3. Hochberg MC, Altman RD, Brandt
orders in the United States. Arthritis ease Control’s role in combating the bur- n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved. gluc o s a mine a nd chondroi t in sul fat e f or k nee o s t e oa r t hr i t is arthritis of the knee. Arthritis Rheum inary recommendations from a task force 26. Reginster JY, Deroisy R, Rovati LC, et al.
1995;38:1541-6.
of the Osteoarthritis Research Society. Long-term effects of glucosamine sul- 4. Williams HJ, Ward JR, Egger MJ, et al.
phate on osteoarthritis progression: a ran- 15. Ware JE Jr, Sherbourne CD. The MOS domised, placebo-controlled clinical trial.
phen in a two-year study of treatment of 36-item short-form health survey (SF-36). Lancet 2001;357:251-6.
osteoarthritis of the knee. Arthritis Rheum 27. Pavelka K, Gatterova J, Olejarova M,
Machacek S, Giacovelli G, Rovati LC. Glu- 5. Bradley JD, Brandt KD, Katz BP, Kala-
16. Fries JF, Spitz P, Kraines RG, Holman cosamine sulfate use and delay of pro-
sinski LA, Ryan SI. Comparison of an HR. Measurement of patient outcome in gression of knee osteoarthritis: a 3-year, antiinflammatory dose of ibuprofen, an arthritis. Arthritis Rheum 1980;23:137-45.
analgesic dose of ibuprofen, and acetamin- 17. Pham T, van der Heijde D, Altman RD, blind study. Arch Intern Med 2002;162:
ophen in the treatment of patients with et al. OMERACT-OARSI initiative: Osteo- osteoarthritis of the knee. N Engl J Med arthritis Research Society International 28. Cibere J, Kopec JA, Thorne A, et al.
1991;325:87-91.
set of responder criteria for osteoarthritis Randomized, double-blind, placebo-con- 6. Ray WA, Griffin MR, Stein CM. Car-
clinical trials revisited. Osteoarthritis Car- trolled glucosamine discontinuation trial diovascular toxicity of valdecoxib. N Engl tilage 2004;12:389-99.
18. SAS user’s guide, version 8. Cary, N.C.: 2004;51:738-45.
7. Bresalier RS, Sandler RS, Quan H, et al.
29. McAlindon T, Formica M, LaValley M,
Cardiovascular events associated with ro- 19. Annual nutrition industry overview. Lehmer M, Kabbara K. Effectiveness of
20. Altman RD, Moskowitz R. Intraartic-
arthritis: results from an Internet-based ular sodium hyaluronate (Hyalgan) in the randomized double-blind controlled trial. 8. Nussmeier NA, Whelton AA, Brown treatment of patients with osteoarthritis Am J Med 2004;117:643-9.
MT, et al. Complications of the COX-2 in-
of the knee: a randomized clinical trial. 30. Mazieres B, Loyau G, Menkes CJ, et al.
hibitors parecoxib and valdecoxib after J Rheumatol 1998;25:2203-12. [Erratum, Chondroitin sulfate in the treatment of cardiac surgery. N Engl J Med 2005;352: 21. Lehmann R, Brzosko M, Kopsa P, et al. result of a multicenter double-blind con-
9. McAlindon TE, LaValley MP, Gulin JP, Efficacy and tolerability of lumiracoxib trolled prospective study using placebo.
Felson DT. Glucosamine and chondroitin 100 mg once daily in knee osteoarthritis: Rev Rhum Mal Osteoartic 1992;59:466-72.
for treatment of osteoarthritis: a system-
a 13-week, randomized, double-blind study atic quality assessment and meta-analysis. vs. placebo and celecoxib. Curr Med Res 31. Conrozier T. Anti-arthrosis treatments:
JAMA 2000;283:1469-75.
efficacy and tolerance of chondroitin sul- 10. Kellgren JH, Lawrence JS. Radiologi-
22. Cannon GW, Caldwell JR, Holt P, et al. fates (CS 4&6). Presse Med 1998;27:1862-
cal assessment of osteo-arthrosis. Ann Rofecoxib, a specific inhibitor of cyclo- 32. Towheed TE, Maxwell L, Anastassiades
11. Bellamy N. Pain assessment in osteo-
parable with that of diclofenac sodium: TP, et al. Glucosamine therapy for treat- arthritis: experience with the WOMAC results of a one-year, randomized, clinical ing osteoarthritis. Cochrane Database Syst osteoarthritis index. Semin Arthritis trial in patients with osteoarthritis of the 33. Adebowale AO Cox DS, Liang Z, Ed-
12. Bellamy N, Buchanan WW, Goldsmith
CH, Campbell J, Stitt LW. Validation study 23. Schnitzer TJ, Weaver AL, Polis AB, et al.
of WOMAC: a health status instrument for Efficacy of rofecoxib, celecoxib, and aceta- measuring clinically important patient minophen in patients with osteoarthritis of relevant outcomes to antirheumatic drug the knee: a combined analysis of the VACT therapy in patients with osteoarthritis of studies. J Rheumatol 2005;32:1093-105.
34. Joint remedies — glucosamine and
the hip or knee. J Rheumatol 1988;15:1833- 24. Deeks JJ, Smith LA, Bradley MD. Ef-
ficacy, tolerability, and upper gastrointes- 13. Hochberg MC, Chang RW, Dwosh I, tinal safety of celecoxib for treatment of the disease itself: we tested 19 national
Lindsey S, Pincus T, Wolfe F. The Ameri-
osteoarthritis and rheumatoid arthritis: products. Consum Rep 2002;67:18-21.
can College of Rheumatology 1991 revised 35. Russell AS, Aghazadeh-Habashi A,
criteria for the classification of global trolled trials. BMJ 2002;325:619.
Jamali F. Active ingredient consistency of functional status in rheumatoid arthritis. 25. Houpt JB, McMillan R, Wein C, Paget-
14. Hochberg MC, Altman RD, Brandt chloride in the treatment of pain of osteo-
KD, Moskowitz RW. Design and conduct arthritis of the knee. J Rheumatol 1999; Copyright 2006 Massachusetts Medical Society. of clinical trials in osteoarthritis: prelim- RECEIVE IMMEDIATE NOTIFICATION WHEN
A JOURNAL ARTICLE IS RELEASED EARLY
To be notified when an article is released early on the Web and to receive the table of contents of the Journal by e-mail every Wednesday evening, www.nejm.org
n engl j med 354;8 www.nejm.org february Downloaded from www.nejm.org on April 8, 2008 . For personal use only. No other uses without permission. Copyright 2006 Massachusetts Medical Society. All rights reserved.

Source: http://www.porckepzes.hu/sites/default/files/upload/GAIT_Study_Abstract.pdf