Abstract
Purpose (the aim of the study): People with prior intra-articular knee injury are at substantial risk of post-traumatic osteoarthritis (PTOA). Systemic biomarkers may detect early disease activity, aid diagnosis, and monitor progression. We hypothesised that individuals 1–7 years following an intra-articular knee injury would present a significantly different multivariate profile of systemic biomarker concentrations compared with highly matched controls.
Methods: Seventy-six participants took part in the study. Groups were matched for sex (20 females per group), age (26 ± 6 vs 25 ± 6 years, p = 0.33), fat mass index (6.2 ± 3.1 vs 5.9 ± 3.7 kg/m2, p = 0.27), fat free mass index (17.8 ± 2.3 vs 17.6 ± 2.5 kg/m2, p = 0.29), and maximal oxygen uptake (44.4 ± 7.5 vs 45.4 ± 9.7 mL/kg/min, p = 0.37). Median time since surgery was 24 (12-72) months. Participants attended within two hours of waking fasted, avoiding exercise, caffeine, and alcohol for 24 hours. After 30 minutes non-weightbearing, venous blood (10mL) was collected from the antecubital fossa using a butterfly needle. Samples coagulated for 20 minutes at room temperature and were centrifuged (1000g, 15 minutes, 4°C). Serum nordicARG™ (aggrecan degradation fragment) was quantified using a sandwich chemiluminescent immunoassay, while nordicC2M™, nordicPRO-C2™, nordicC3M™, and nordicPRO-C4™ were quantified using competitive chemiluminescent immunoassays (IDS-iSYS Multi-Discipline Automated Analyzer) and were analysed singly given high analytical precision (intra-assay CV: 4–9%; inter-assay CV: 6–15%). Serum COMP was measured in duplicate (intra-assay CV: 2%; inter-assay CV: 9%) using sandwich ELISA. Data were ln-transformed, and normality assessed using Shapiro–Wilk tests. Group differences were tested via permutational multivariate analysis of variance (PERMANOVA), with dispersion homogeneity confirmed and distance-based redundancy analysis used to visualise separation and biomarker loadings. One-tailed paired t-tests applied Benjamini–Hochberg correction (m = 6, q < 0.05). ROC analyses assessed discrimination (AUC) using a combined COMP + C3M + PRO-C4 model; the Youden index defined optimal sensitivity–specificity. Leave-one-pair-out cross-validation estimated model performance.
Results: Multivariate analysis revealed a significant difference between those after a knee injury and uninjured matched controls (R2 = 0.05, F = 4.10, p = 0.001). Canonical loading indicated that COMP, PRO-C4, and C3M drove the group separation (Figure 1). Consequently, a second PERMANOVA including only these three biomarkers revealed a greater multivariate magnitude of difference (R2 = 0.09, F = 7.65, p = 0.0004). Univariate analyses demonstrated moderate-to-large paired effects for COMP (Injured: 2.53 ± 0.26 U/L; Control: 2.34 ± 0.25 U/L; dz = 0.60; p = 0.001; q = 0.004; 95% CI = 0.08–0.29), C3M (Injured: 6.26 ± 0.14 ng/mL; Control: 6.16 ± 0.16 ng/mL; dz = 0.50; p = 0.004; q = 0.01; 95% CI = 0.03–0.17) and PRO-C4 (Injured: 8.73 ± 0.12 ng/mL; Control: 8.67 ± 0.12 ng/mL; dz = 0.40; p = 0.02; q = 0.04; 95% CI = 0.01–0.12). Receiver operating characteristic analysis of the combined COMP + C3M + PRO-C4 model demonstrated good discrimination between groups (AUC = 0.77, 95% CI 0.66–0.87), which remained robust under leave-one-pair-out cross-validation (AUC = 0.73, 95% CI 0.61–0.84; optimism = 0.04) (Figure 2).
Conclusions: This is the first study to robustly detect systemic biomarker differences after intra-articular knee injury. This multivariate approach provides a foundation for development of a biomarker panel to detect those at risk of PTOA, monitor progression, and act as a therapeutic target.
Methods: Seventy-six participants took part in the study. Groups were matched for sex (20 females per group), age (26 ± 6 vs 25 ± 6 years, p = 0.33), fat mass index (6.2 ± 3.1 vs 5.9 ± 3.7 kg/m2, p = 0.27), fat free mass index (17.8 ± 2.3 vs 17.6 ± 2.5 kg/m2, p = 0.29), and maximal oxygen uptake (44.4 ± 7.5 vs 45.4 ± 9.7 mL/kg/min, p = 0.37). Median time since surgery was 24 (12-72) months. Participants attended within two hours of waking fasted, avoiding exercise, caffeine, and alcohol for 24 hours. After 30 minutes non-weightbearing, venous blood (10mL) was collected from the antecubital fossa using a butterfly needle. Samples coagulated for 20 minutes at room temperature and were centrifuged (1000g, 15 minutes, 4°C). Serum nordicARG™ (aggrecan degradation fragment) was quantified using a sandwich chemiluminescent immunoassay, while nordicC2M™, nordicPRO-C2™, nordicC3M™, and nordicPRO-C4™ were quantified using competitive chemiluminescent immunoassays (IDS-iSYS Multi-Discipline Automated Analyzer) and were analysed singly given high analytical precision (intra-assay CV: 4–9%; inter-assay CV: 6–15%). Serum COMP was measured in duplicate (intra-assay CV: 2%; inter-assay CV: 9%) using sandwich ELISA. Data were ln-transformed, and normality assessed using Shapiro–Wilk tests. Group differences were tested via permutational multivariate analysis of variance (PERMANOVA), with dispersion homogeneity confirmed and distance-based redundancy analysis used to visualise separation and biomarker loadings. One-tailed paired t-tests applied Benjamini–Hochberg correction (m = 6, q < 0.05). ROC analyses assessed discrimination (AUC) using a combined COMP + C3M + PRO-C4 model; the Youden index defined optimal sensitivity–specificity. Leave-one-pair-out cross-validation estimated model performance.
Results: Multivariate analysis revealed a significant difference between those after a knee injury and uninjured matched controls (R2 = 0.05, F = 4.10, p = 0.001). Canonical loading indicated that COMP, PRO-C4, and C3M drove the group separation (Figure 1). Consequently, a second PERMANOVA including only these three biomarkers revealed a greater multivariate magnitude of difference (R2 = 0.09, F = 7.65, p = 0.0004). Univariate analyses demonstrated moderate-to-large paired effects for COMP (Injured: 2.53 ± 0.26 U/L; Control: 2.34 ± 0.25 U/L; dz = 0.60; p = 0.001; q = 0.004; 95% CI = 0.08–0.29), C3M (Injured: 6.26 ± 0.14 ng/mL; Control: 6.16 ± 0.16 ng/mL; dz = 0.50; p = 0.004; q = 0.01; 95% CI = 0.03–0.17) and PRO-C4 (Injured: 8.73 ± 0.12 ng/mL; Control: 8.67 ± 0.12 ng/mL; dz = 0.40; p = 0.02; q = 0.04; 95% CI = 0.01–0.12). Receiver operating characteristic analysis of the combined COMP + C3M + PRO-C4 model demonstrated good discrimination between groups (AUC = 0.77, 95% CI 0.66–0.87), which remained robust under leave-one-pair-out cross-validation (AUC = 0.73, 95% CI 0.61–0.84; optimism = 0.04) (Figure 2).
Conclusions: This is the first study to robustly detect systemic biomarker differences after intra-articular knee injury. This multivariate approach provides a foundation for development of a biomarker panel to detect those at risk of PTOA, monitor progression, and act as a therapeutic target.
| Original language | English |
|---|---|
| DOIs | |
| Publication status | E-pub ahead of print - 23 Apr 2026 |
| Event | Osteoarthritis Research Society International (OARSI) World Congress 2026 - Palm Beach, Palm Beach, USA United States Duration: 23 Apr 2026 → 26 Apr 2026 https://congress.oarsi.org/ |
Conference
| Conference | Osteoarthritis Research Society International (OARSI) World Congress 2026 |
|---|---|
| Country/Territory | USA United States |
| City | Palm Beach |
| Period | 23/04/26 → 26/04/26 |
| Internet address |
Fingerprint
Dive into the research topics of 'Elevated Serum COMP, C3M, and PRO-C4 in Individuals 1–7 Years After Knee Injury: Systemic Indicators of Post-Traumatic Osteoarthritis Activity'. Together they form a unique fingerprint.Cite this
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS