Researchers Marzieh Soheili and Hamed Gilzad Kohan of Western New England University built a quantitative systems pharmacology, or QSP, model that connects three layers of biology into a single simulation: the pharmacokinetics of glucosamine absorption and delivery to synovial fluid, the turnover of glycosaminoglycans, the sponge-like matrix molecules that give cartilage its resilience, and the clinical endpoints that patients and radiologists actually measure, namely joint space width on X-rays and pain scores on the WOMAC scale. The goal was to test whether the conflicting trial results could be reconciled by differences in drug exposure at the target tissue, rather than by mysterious differences in biology.

To build the model, the team conducted a systematic review of the literature through PubMed, retrieving 177 randomized controlled trials of glucosamine in osteoarthritis. From 145 eligible studies, they selected five landmark datasets for calibration and validation: the three-year structural trials by Reginster and Pavelka, the GAIT pain trial, the Spanish GUIDE trial for independent validation, and the MOVES combination-therapy trial as an exploratory stress test. The model was implemented as a system of ordinary differential equations in MATLAB, with five free parameters estimated against eighteen aggregate observations from the published trials and all pharmacokinetic inputs fixed from literature values.

The calibration performance was strikingly good. The model reproduced the three-year joint space width treatment effect with a predicted value of 0.211 millimeters against an observed 0.240 millimeters, a 12 percent error, and it tracked the 24-week pain trajectories of both the placebo and glucosamine arms of GAIT within roughly ten percent. External validation against the GUIDE trial, which used the same crystalline glucosamine sulfate formulation and a true placebo arm, found that every observed pain score fell within the model’s 90 percent prediction intervals, with a mean absolute error of 3.8 WOMAC points. Notably, the model under-predicted the between-arm treatment difference in GUIDE, predicting 1.4 points of benefit versus the observed 5.0, which the authors attribute to their deliberate choice of a structure-dominated pain mechanism with only a minor direct analgesic component.

But the study’s most consequential finding is a cautionary one. The model could only reproduce the European structural benefit because it assumed an anti-catabolic potency anchor, an IC50 of 3.0 micrograms per milliliter, that places achievable joint concentrations on the steep part of the drug’s inhibition curve. When the team re-calibrated the model across the concentration range at which glucosamine’s anti-inflammatory effects are actually observed in laboratory studies, roughly 10 to 1,000 micrograms per milliliter, the predicted structural benefit collapsed below the minimal clinically important difference of 0.1 millimeters and the fit degraded substantially. Measured steady-state synovial concentrations after a standard 1,500 milligram dose are only about 0.78 micrograms per milliliter, one to three orders of magnitude below the levels that suppress cartilage degradation in cell experiments.

This potency mismatch offers a mechanistic explanation for the inconsistent trial record. If glucosamine’s true potency at its molecular target corresponds to the in vitro range, then even optimistic assumptions about oral absorption cannot deliver enough drug to the joint to meaningfully inhibit the IL-1beta and NF-kappaB signaling pathways that drive matrix breakdown. Under that scenario, the predicted disease-modifying effect largely disappears, and the positive European results become difficult to explain by pharmacology alone. The authors are explicit that their optimistic predictions, including a minimum effective dose of about 750 milligrams per day, an 83 percent responder rate at the standard dose, and consistent benefit across all severity stages, are all conditional on the same potency assumption and on calibration data from industry-sponsored trials that independent studies have not replicated.

The model also weighs in on a formulation debate that has shaped the glucosamine market for years. The base model assumed that crystalline glucosamine sulfate is about twice as bioavailable as the hydrochloride salt, which would partly explain why European trials using the sulfate form succeeded where GAIT, using the hydrochloride, did not. But recent pharmacokinetic work found no significant difference in systemic exposure between crystalline and regular glucosamine sulfate, and crystallographic analyses have shown that commercial glucosamine sulfate products are actually physical mixtures of glucosamine chloride and potassium or sodium salts rather than true sulfate salts. Because all of these forms dissociate in gastric fluid to release the same active glucosamine cation, the salt-form explanation for the trial discrepancy now looks shaky, and the model’s own sensitivity analysis shows that equalizing bioavailability shifts the minimum effective dose upward to 1,500 milligrams per day, precisely the dose at which independent trials have failed.

Beyond diagnosing the controversy, the framework generates forward-looking insights that illustrate why regulators are increasingly embracing model-informed drug development. Simulations predicted a near-linear relationship between oral bioavailability and treatment effect, suggesting that raising bioavailability from 22 to 40 percent could boost the structural benefit by roughly 73 percent without any dose increase. Because glucosamine is absorbed mainly through glucose transporters rather than passive diffusion, conventional permeability enhancers may not help, but preclinical strategies such as chitosan-based formulations, liposomes, and a peptide prodrug targeting the intestinal PepT1 transporter have shown promise in laboratory studies. The duration simulations also found that structural benefit continues accumulating through five years without plateau, supporting long-term treatment for patients who respond.

The work arrives at a moment when the regulatory landscape is shifting in its favor. The FDA Modernization Act 2.0, signed in December 2022, removed the blanket mandate for animal testing and explicitly authorized computational models as alternatives, and QSP submissions to the agency have grown steadily, exceeding 80 in 2023 alone. Osteoarthritis, which affects an estimated 595 million people worldwide and is projected to rise by 75 to 95 percent by 2050, still has no approved disease-modifying drug, and the authors argue that mechanistic models calibrated to human data could reduce the late-stage failures that have plagued the field. The complete MATLAB pipeline, including all calibration, bootstrap, and simulation scripts, has been released publicly on GitHub for independent replication.

Ultimately, the study’s value lies less in endorsing glucosamine than in stating precisely what would have to be true for it to work. The model shows that a disease-modifying effect requires both sufficient anti-catabolic potency and adequate joint exposure, and it demonstrates that current evidence does not establish that either condition is met at clinically achievable concentrations. For a supplement whose benefits have been argued about for a quarter century, that kind of quantitative clarity, mapping exactly where the pharmacology falls short and what a better formulation would need to achieve, may prove more useful than another round of contradictory trials. The glucosamine controversy, it turns out, may have been a pharmacokinetic question all along, and now it finally has numbers attached.

Subject of Research: A quantitative systems pharmacology model of glucosamine sulfate pharmacokinetics, cartilage matrix dynamics, and clinical outcomes in knee osteoarthritis

Article Title: A mechanistic quantitative systems pharmacology framework for glucosamine sulfate in knee osteoarthritis: linking exposure, cartilage biology, and clinical outcomes

Article References: A mechanistic quantitative systems pharmacology framework for glucosamine sulfate in knee osteoarthritis: linking exposure, cartilage biology, and clinical outcomes. (n.d.). https://doi.org/10.3389/jpps.2026.16471

Image Credits: AI Generated

DOI: 10.3389/jpps.2026.16471

Keywords: glucosamine, osteoarthritis, quantitative systems pharmacology, cartilage, bioavailability, joint space width, WOMAC, disease-modifying osteoarthritis drug, model-informed drug development, pharmacokinetics, NF-kappaB, glycosaminoglycan