Glucosamine and chondroitin sulfate have been available for almost 2 decades in Europe and other countries as prescription agents for treating osteoarthritis. To date, there are 40 clinical trials on this natural combination therapy, almost all showing efficacy. Despite wide use in other parts of the world, glucosamine and chondroitin were practically unknown in the United States until the late 1990s.

Since then, their popularity has grown quickly, vaulting the combination to the number-three position among best-selling dietary supplement categories, just behind multivitamins and calcium/mineral products.

I was an early adopter of glucosamine/chondroitin in my clinical practice. The Arthritis Cure, a book I published in 1997, places it as a cornerstone of a 9-step integrative treatment protocol for patients with osteoarthritis (OA). I believe the combination can provide effective treatment, often obviating the need for over-the-counter or prescription pain relievers and anti-inflammatory drugs, many of which have significant long-term side effects.

Consider that roughly 45 people die each day from side effects of NSAIDs—the mainstay conventional treatment for OA (Singh G. Am J Med. 1998; 105(1): 31S–38S). No deaths or hospitalizations have ever been attributed to glucosamine/chondroitin, as monitored by European pharmacovigilance programs. Finding an effective alternative to conventional drugs has become a major public health concern, prompting the National Institute of Arthritis, Musculoskeletal and Skin Diseases (NIAMS) to spearhead a large 14-center, placebo-controlled study of glucosamine and chondroitin.

In 1999, a team led by Daniel Clegg, MD, a rheumatologist at the University of Utah, convened the Glucosamine/Chondroitin Arthritis Intervention Trial (GAIT). The study involved more than 1,500 patients, making it one of the largest supplement studies ever undertaken. I served on the GAIT steering oversight committee. The long-awaited data were initially presented late last year at the annual meeting of the American College of Rheumatology, and were just published in the New England Journal of Medicine.

The data show clearly that for patients with severe baseline symptoms, glucosamine/chondroitin performs as well as or even slightly better than celecoxib (Celebrex), the popular anti-inflammatory drug that served as a positive control in the GAIT protocol.

Designing GAIT

GAIT is a two-part study. The initial six months looked for differences in pain and function in patients suffering from knee OA. The continuation phase, which lasted two years but included only half of the initial population, evaluates possible structure-modifying properties of glucosamine and chondroitin.

GAIT had a long list of inclusion criteria, which were important in keeping the study well controlled. Patients with x-ray proven, Kellgren and Lawrence grades 2 or 3 OA were included. Subjects were symptomatic for at least six months.

Initially, the study compared glucosamine HCl alone (500 mg, TID) versus the combination of glucosamine HCl and chondroitin sulfate (500 mg and 400 mg, TID) versus placebo. Soon afterwards, a chondroitin-alone arm (400 mg TID) was added, as was the positive control arm (celecoxib 200 mg, once daily). The glucosamine HCl used in GAIT is manufactured by Ferro-Pfansteihl, a German chemical company. The chondroitin sulfate was the CSb Bio-Active compound from Bioiberica, a Spanish company.

The primary outcome measure was a 20% or greater reduction in WOMAC (Western Ontario McMaster University Osteoarthritis Index) score; several key secondary outcomes were also tracked. An important facet of the study was the baseline stratification of subjects based on symptom severity: the low/moderate pain group had WOMAC scores between 125 and 300; the moderate-to-severe cohort had scores between 301 and 400.

The stratification was essential because response to any given treatment can differ greatly depending on initial symptom severity. From a structural standpoint, this makes sense. Early in OA, there is a breakdown of the collagen framework, but chondrocytes are able to keep up. Later, when production and degradation of matrix components becomes “uncoupled,” the chondrocytes become progressively less able to keep up with the rapid breakdown of large cartilage macromolecules (mostly proteoglycans). The result is a net loss of cartilage matrix volume and a decrease in joint space, visible in weight-bearing x-rays.

Since it is unethical to study patients in pain without offering rescue medicine, the GAIT protocol provided patients with acetaminophen in 500 mg tablets. Use of rescue medicine was compared from group to group.

Study Outcomes: Reasons to Be Cheerful

	Arthritis Rheum. 2005; 9(Suppl.): 622.

One of the most striking features of the GAIT data was the huge placebo effect of 60%. For the primary outcome measure (reduction of WOMAC score) each of the supplement groups showed response rates better than placebo, but the differences were not statistically significant. The positive control, celecoxib, did show a statistically significant difference. However, bear in mind that only 10 more patients out of each 100 responded to celecoxib compared to placebo.

The most interesting finding, and perhaps the most clinically relevant, was the outcome in the high WOMAC group. In clinical practice, patients with greatest pain and most restricted overall function are most likely to require drug intervention or surgery. From a public health standpoint, these patients utilize the most healthcare resources and have the greatest disability.

Surprisingly, among the 348 high WOMAC patients, there was no difference between celecoxib and placebo. In contrast, glucosamine and chondroitin showed a highly significant advantage over placebo. Seventy-nine percent of high WOMAC patients had a meaningful clinical response to the supplements.

In the study cohort as a whole, patients on glucosamine/chondroitin showed the lowest mean use of rescue medication (1.3 tablets per day versus 2.6 in the placebo group). This is clinically significant in light of the recent Nurses Health Study showing that women aged 34 to 77 who took an average of 500 mg of acetaminophen per day had double the risk of hypertension. The glucosamine/chondroitin group in GAIT utilized 650 mg less acetaminophen per patient, on average. Given that OA affects over 20 million people, there is a great potential for cutting hypertension risk by reducing the need for acetaminophen.

Is the Research Deck Stacked Against Supplements?

Champions of dietary supplements are often discouraged by large federally funded or university sponsored supplement studies that seem destined to yield negative results. The most notable example was the study of St. John’s Wort for moderate to severe depression (typically not the patient group for which this herb is intended). This NIH-sponsored trial was widely cited as definitive proof that “the herb doesn’t work.” Few media outlets reported the equally important finding that Zoloft, the comparator drug, was also no better than placebo.

Was GAIT another study designed to discredit supplements? It did seem odd to many people that the celecoxib arm was added seemingly late in the game. Be that as it may, GAIT’s lead investigator, Dr. Clegg, though having financial ties to both drugs used in the study, is a well-respected, experienced rheumatologist, and is considered one of the most honest and fair academic practitioners in the field.

To my mind, the most striking finding from GAIT is the fact that celecoxib was no better than placebo in patients with the most severe OA. This is in sharp contrast to other studies showing benefit from celecoxib—that’s how the drug won FDA approval in the first place. In a study like this, when the positive control (celecoxib) is negative and the negative control (placebo) is relatively positive, one must question the results. Should GAIT simply be called inconclusive?

I’m not sure we should go that far, but there’s clearly a question of why the drug did not help those with the most severe pain. Obviously, something about the study blunted the results, not only for Celebrex®, but for all treatment groups. As impressive as the data may appear in favor of glucosamine/chondroitin in the high WOMAC group, there were many negative biases in the study that would likely have led to false negative results for GAIT:

1. GAIT permitted enrollment of patients who had taken glucosamine and chondroitin in the past. Since we don’t know how long the washout period is for these substances, this could have blunted the net effects. An eight-year study on glucosamine (Osteoarthritis and Cartilage. 2004; 12(Suppl. B): 180) showed that even people who took crystalline glucosamine sulfate alone for 3 years had 75% fewer joint replacements, suggesting glucosamine has large long-term effect.
2. Most OA studies include only patients with primary OA (of unknown origin). Yet, many clinicians report that the most pronounced effects of glucosamine and chondroitin are in secondary OA, especially that caused by gout or pseudogout. GAIT’s population may not have been ideal for evaluating these supplements.
3. Split dosing, thrice daily for the supplements, might have lowered the average concentration in the blood and synovial fluid to levels below the threshold for modulating gene activation in chondrocytes. We can’t blame the investigators for this—GAIT was designed when the thinking was that split dosing was okay. Subsequent pharmacokinetics data have challenged this. We now know once daily dosing not only increases compliance, it also increases peak blood levels.
4. The high placebo rate obviously could have washed out evidence of efficacy in all groups. It is not surprising that the placebo rate was so high. Subjects knew they had an 80% chance of receiving active treatment. In addition, each study subject took seven capsules daily. Research staff called subjects almost weekly to help maintain compliance. It is difficult to quantify the psychological impact of these factors, but equally difficult to ignore it. Also, compliance with once-daily Celebrex was, not surprisingly, higher than for the thrice daily supplements.
5. There may be concealment issues related to Celebrex. Standard Celebrex capsules were used, and an outer capsule was fitted over them. I don’t think the concealment capsules inhibited digestion or disintegration, but if someone opened the outer capsules, they could easily see what they had, thus compromising the blinding.

I am not cynical enough to believe that the symptom/function part of GAIT was designed to fail, but we have certainly learned a lot since the study’s origins in 1999. No doubt there are many things we would do differently today.

I am more skeptical about the structure part of the study, due to be published next year. Other studies on glucosamine/chondroitin show that the natural progression of joint space loss is about 0.1 mm per year. The study protocol requires a difference of greater than 0.2 mm per year to show significance, meaning it will be nearly impossible to see a positive effect on cartilage structure.

Obviously, if the supplements showed a structure-modifying benefit, it would have a catastrophic influence on the pharmaceutical industry’s arthritis sector. I have no doubt drug industry representatives were pleased with the design of the structure component of GAIT.

Should the structure findings prove negative for the supplements, which is likely, I hope people will understand that this portion of the study was doomed from the start. GAIT data should not outweigh all of the other studies showing positive cartilage structure modification for both glucosamine and chondroitin.

A First Line Therapy

Bearing in mind the limitations of the study, I believe we can still learn a lot from GAIT. First, since glucosamine/chondroitin was shown to be more effective than Celebrex for moderate to severe OA, and since about 40 other studies show benefit, these supplements should be considered first-line therapy for OA.

Second, the study sheds light on risk/benefit and cost questions. Celebrex costs approximately $2–$3 per day; in contrast, a high-quality glucosamine/chondroitin product costs between $0.50 to $1 per day. Multiplied by the millions of people with OA, that’s a significant cost differential. Add to this the hidden costs of a drug like Celebrex. To obtain Celebrex, a patient must see a physician, which carries a cost. Then there are costs of co-medication to keep the drug, in this case Celebrex, from causing problems like bleeding ulcers.

The 2004 GI guidelines from Pharmacy & Therapeutics call for all people at high risk for ulcers—and this includes those just taking low-dose aspirin, let alone an NSAID—to also take a proton pump inhibitor as a preventive therapy. The cost for PPIs ranges from about $1 to $4 per day. If even a small portion of the OA population follows the guidelines, the cost would be massive. PPIs are not without side effects. Co-medication is unnecessary with glucosamine/ chondroitin.

Adverse effects in all GAIT groups were generally mild. Nevertheless, there were more cardiovascular events in the Celebrex group (dysrhythmias and BP elevations). Bear in mind, the data reported so far were only for six months in treatment. It took 18 months of treatment to see the adverse CV effects from COX-2 inhibitors, so we shouldn’t give Celebrex a free pass based on GAIT.

In my experience, the main problem with glucosamine/chondroitin as a first-call therapy is that it takes a few weeks to start working, and may not reach full effect until six months or more. This is important not only for clinical practice but also for trial interpretation. Any study of glucosamine and chondroitin should be at least six months long. Shorter trials are probably invalid, including negative ones.

This slow onset of effect is, in part, why I recommend that patients with significant pain or inflammation be best treated with glucosamine/chondroitin along with an anti-inflammatory drug for pain relief. My goal is always to stop the drug as soon as possible. When it comes to treating arthritis, an integrative approach is best; supplements and pharmaceuticals can and should be used together when in the best interests of our patients.

Product Quality and Reliability

Many products purported to contain glucosamine and chondroitin are deficient in dose or activity level of the chondroitin. Too many products are utilizing less than 800 mg of pure chondroitin sulfate, or are using inactive agents such as avian chondroitin. A study reported at the International Cartilage Repair Society meeting showed 5 out of 10 raw material sources of chondroitin failed to show anti-inflammatory efficacy. Chondroitin from avian sources, unlike bovine, porcine, or fish sources, have no clinical effect.

On the glucosamine side, there is a great body of evidence supporting the crystalline form of glucosamine sulfate (sold in the US only under the brand name Dona®) for both symptoms and structure modification. I counsel my patients against other brands of glucosamine sulfate since they have a different chemical composition and have not been tested for efficacy. GAIT used glucosamine HCl, which is in about 90% of the products in the US. However, since glucosamine HCl hasn’t been studied clinically against Dona®, we can’t say if it’s equivalent or not.

Many companies looking for competitive advantage when marketing glucosamine/chondroitin will cite data from very brief trials, yet no meaningful information can come from an 8 or 12-week study. I believe these companies should be fined by the FDA or FTC. Junk science not only misleads our patients, it hurts the entire natural medicine movement. Adding subpotent or sub-therapeutic doses of ancillary ingredients to glucosamine/chondroitin is another common tactic. Unless there is science to support such a strategy, this also harms the credibility of legitimate products.

Vegetarians and people who keep Kosher or Halal diets cannot take most common forms of glucosamine, since they come from shellfish chitin. There is a vegetarian form of glucosamine HCl derived by fermentation (from corn), but currently there is no vegetarian source for chondroitin.

Jason Theodosakis, MD, MS, MPH, FACPM is board certified in Preventive Medicine and Sports Medicine. He is on the medical staff at Canyon Ranch resort in Tucson, AZ, and served as the director of the University of Arizona College of Medicine’s preventive medicine residency training program. His 1997 book, The Arthritis Cure (revised 2004), greatly influenced broader acceptance of glucosamine and chondroitin as an arthritis therapy in the US. Most recently, he served on the oversight steering committee for the NIH-sponsored GAIT trial.