Tier 1 — highest

Effect of collagen supplementation on osteoarthritis symptoms: a meta-analysis of randomized placebo-controlled trials

Juan Mario García-Coronado, Lorena Martínez-Olvera, Rodrigo E. Elizondo-Omaña, Carlos Alberto Acosta-Olivo, Félix Vilchez-Cavazos, Luis Ernesto Simental-Mendía, Mario Simental-Mendía
International Orthopaedics 2019 43(3):531–538

Bibliography

PubMed
PMID 30368550
DOI
10.1007/s00264-018-4211-5
Funding
No funding source is disclosed in the paper.
Competing interests
The authors declared no conflict of interest. Their listed affiliations were academic and public clinical institutions in Mexico; no commercial collagen-manufacturer affiliation is listed in the paper.

Study snapshot

DesignSystematic review and meta-analysis of randomised placebo-controlled trials (Cochrane risk-of-bias framework, random-effects model, generic inverse variance method, I² heterogeneity statistic)
ModelAdults with osteoarthritis (primarily knee OA), Kellgren-Lawrence grades 1–4 across included trials
Sample5 RCTs pooled (n=519 total: 277 collagen, 242 placebo)
InterventionOrally administered collagen supplements across trials: hydrolysed collagen 10 g/day (three trials); undenatured type II collagen 40 mg/day (one trial); hydrolysed collagen 2 g/day (one trial). Sources included chicken, porcine, and bovine.
DurationIntervention periods 10–48 weeks across the pooled trials
EndpointsChange in total WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index) score; WOMAC subscores (pain, stiffness, functional limitation); Visual Analog Scale (VAS) pain score

What the study showed, in plain terms

This 2019 meta-analysis pooled five randomized placebo-controlled osteoarthritis trials involving 519 participants. The included interventions were heterogeneous: hydrolysed collagen at gram doses and undenatured Type II collagen at a milligram dose were analysed together.

The pooled analysis favoured collagen interventions for total WOMAC and VAS pain, while the WOMAC pain and physical-function subscores were not individually significant. Heterogeneity was high for total WOMAC and very high for VAS pain.

The review therefore supports a pooled symptom signal across heterogeneous collagen interventions, not a clean class-effect estimate for hydrolysed collagen. Mixing hydrolysed collagen with undenatured Type II collagen, the small number of trials and inconsistent subscale results materially limit interpretation.

Key findings

  • Total WOMAC index: weighted mean difference -8.00 (95% CI -13.04 to -2.95, p=0.002) in favour of collagen versus placebo, with high between-study heterogeneity (I²=78%).
  • WOMAC stiffness subscore: WMD -0.41 (95% CI -0.74 to -0.08, p=0.01) in favour of collagen.
  • WOMAC pain subscore: WMD -0.22 (95% CI -1.58 to 1.13, p=0.75) — not significant.
  • WOMAC functional limitation subscore: WMD -0.62 (95% CI -5.77 to 4.52, p=0.81) — not significant.
  • VAS pain score: WMD -16.57 mm (95% CI -26.24 to -6.89, p<0.001) in favour of collagen, with very high heterogeneity (I²=96%).
  • All five included trials showed low risk of bias for incomplete outcome data and selective outcome reporting per Cochrane framework. Two trials had incomplete information on randomisation or blinding.
  • Pooled sample: 519 adults with osteoarthritis (primarily knee), 277 receiving collagen and 242 receiving placebo, across interventions ranging 2 g/day hydrolysed collagen to 40 mg/day undenatured type II collagen, over 10 to 48 weeks.

What this study can and cannot tell us

  • Only five RCTs met inclusion criteria — a small meta-analytic base. Statistical pooling is possible but confidence intervals remain wide and any single trial has substantial leverage on the result.
  • High between-study heterogeneity for both WOMAC (I²=78%) and VAS (I²=96%). This means the included trials are estimating meaningfully different effects, likely due to variation in collagen type, dose, source, treatment duration, and OA severity. Pooled estimates should be read with caution when heterogeneity is this high.
  • The WOMAC pain subscore, VAS pain, and total WOMAC ostensibly measure related constructs but produced divergent results (pain subscore n.s., VAS strongly significant). The authors attribute this to the WOMAC pain subscore measuring pain across five specific activities while VAS measures overall pain intensity — but this discrepancy is a real concern the reader should weigh.
  • Included trials tested different collagen preparations (hydrolysed collagen 2 g and 10 g doses, undenatured type II collagen 40 mg) as if they were interchangeable. They are not — hydrolysed collagen and undenatured type II collagen are hypothesised to work by very different mechanisms and should arguably not be pooled at all.
  • Two of the five trials ran for only 10–13 weeks, potentially insufficient time for the full osteoarthritis symptom effect to develop.
  • Publication bias is not formally assessed. A funnel plot analysis or Egger's test would have strengthened the conclusion.
  • All primary trials were themselves subject to variable risk of bias — two had unclear randomisation sequence generation or allocation concealment.
  • Meta-analytic reanalyses published subsequently (e.g. Do collagen supplements reduce symptoms of osteoarthritis? Int Orthop 2021) have questioned whether the effect sizes reported here are as strong as they appear, and note that the minimum clinically significant difference on VAS is ~13 mm — the observed 16.57 mm is only slightly above this threshold.
  • The analysis is limited to osteoarthritis symptoms. It does not address whether collagen affects underlying cartilage structure, disease progression, or need for eventual surgical intervention.

Related papers in the corpus

Tier 1
Effect of collagen supplementation on osteoarthritis symptoms: a meta-analysis of randomized placebo-controlled trials
Juan Mario García-Coronado et al. · International Orthopaedics · 2019
Tier 2
Whey protein but not collagen peptides stimulate acute and longer-term muscle protein synthesis with and without resistance exercise in healthy older women: a randomized controlled trial
Sara Y. Oikawa et al. · The American Journal of Clinical Nutrition · 2020
Tier 2
The Influence of Specific Bioactive Collagen Peptides on Body Composition and Muscle Strength in Middle-Aged, Untrained Men: A Randomized Controlled Trial
Denise Zdzieblik et al. · International Journal of Environmental Research and Public Health · 2021

Citation network

Articles citing this research paper

7 articles

Biohack Blueprint analyses that reference this study in their evidence base.

Editorial review

Reviewed by the Biohack Blueprint research team

Last verified

Dr. Khaled Mahmoud, MBBChCritical Care Medicine, Longevity Science & Editorial Direction Dr. Amr Yasser, PhDCardiometabolic, Musculoskeletal, Gut & Bioavailability Dr. Abdulla Mostafa, PhDCellular Senescence, Neurological, Mitochondrial & Brain-Ageing Biology