Glucosamine sulfate and cartilage type II collagen degradation in patients with knee osteoarthritis: randomized discontinuation trial results employing biomarkers

OBJECTIVE: To determine whether glucosamine sulfate has an effect on cartilage type II collagen degradation in patients with knee osteoarthritis (OA). METHODS: A randomized, double blind, placebo controlled glucosamine discontinuation trial was conducted in 137 subjects with knee OA, who had had at least moderate relief of knee pain after starting glucosamine. Subjects were randomized to glucosamine at prestudy dose or placebo at an equivalent dose. Treatment was continued to Week 24 or disease flare, whichever occurred first. Serum and urine samples were collected at Weeks 0, 4, 12, and 24 or flare visit. Samples were analyzed in triplicate for 2 type II collagen degradation biomarkers: C2C epitope (COL2-3/4C(long)) and C1,2C epitope (COL2-3/4C(short)). The primary outcome was the mean change in serum and urine C1,2C/C2C ratio in the glucosamine and placebo groups from baseline to final (flare or Week 24) visit. Linear regression analyses were conducted to adjust for potential confounders. Due to non-normal distributions, the data were log-transformed (lnC1,2C/C2C). Secondary outcomes included comparison of mean change scores at final visit compared to baseline for serum and urine C1,2C and C2C in the 2 treatment groups and in Flare versus No-Flare groups. RESULTS: Baseline and final visit samples were available in 130 subjects for serum analysis and 126 subjects for urinalysis. No significant difference was seen between placebo and glucosamine groups in the serum C1,2C/C2C ratio, with a mean (SD) change from baseline to final visit of 0.8 (27.8) and -0.1 (1.8), respectively (mean difference 0.9; 95% CI -6.0, 7.7, p = 0.80). Similarly, no differences between treatment groups were seen for mean change in urine C1,2C/C2C (p = 0.82), or for mean change in C2C or C1,2C. In linear regression analysis, after adjustment for sex, radiographic severity, baseline lnC1,2C/C2C ratio, WOMAC function, and flare status, treatment was not a significant predictor of final serum or urine lnC1,2C/C2C ratio. When those who experienced flare were contrasted with those without flare, there was a nonsignificant trend toward a difference in mean baseline to final visit change score for serum C1,2C/C2C ratio (p = 0.12). In addition, in the multivariable linear regression analysis, flare status showed a borderline association with final visit serum lnC1,2C/C2C ratio (p = 0.16). CONCLUSION: No statistically significant effect of glucosamine sulfate on type II collagen fragment levels in serum or urine was observed for knee OA over 6 months. Further research is necessary to elucidate which biopathologic systems, if any, are affected by glucosamine treatment. While collagen degradation products may be of value in predicting progression, at least as defined by clinical flare, a larger dataset would be needed to prove this.