BACKGROUND: Clinical determination of bone fracture healing remains qualitative, typically determined through the combination of plain film radiographs, clinical assessment, and patient-reported pain. Importantly, standard radiographs can only detect bone formation in the fracture site after sufficient tissue mineralization has occurred, restricting utility to the later stages of bone repair. A more rigorous method for determining fracture healing progression could significantly improve patient care. Quantitative biomarkers have gained diagnostic value in many clinical indications. Multiple bone turnover biomarkers have been successfully utilized for diagnosis and clinical management of osteoporosis. However, there remains limited evidence for the correlation and diagnostic efficacy of these biomarkers when applied to fracture repair.
METHODS AND FINDINGS: Here we present data from a large cohort of participants without (n = 111) or with (n = 153) fracture with the primary outcome of interest our blood-based biomarker which detects the degradation product of the trimeric collagen X protein (CXM). Collagen X is a transient extracellular matrix protein synthesized by hypertrophic chondrocytes during the soft callus phase of endochondral fracture repair. Our data show that healthy patients with an age range of 21-85 years without fracture (enrolled 2018-2020) have low circulating levels of CXM (median = 563.4 pg/mL; interquartile range (IQR) [483.2, 771.1]) that do not vary independently with age (r = 0.04, p = 0.26) or sex (p = 0.42). Fracture data includes patients with an acute closed or low grade open (Gustilo type I or II) tibial or femoral shaft fracture from a secondary analysis of the VitaShock study (NCT02786498, n = 102: enrolled through the University of Maryland 2016-2019), complemented by a new prospectively enrolled observational study that recruited from the Oregon Health & Science University level 1 academic trauma center (n = 51: enrolled 2019-2023) that were followed until the patient was deemed clinically healed or until they failed to return for follow-up. We show that in serum CXM positively correlates to the bone biomarkers N-terminal propeptide of type I procollagen (P1NP, r = 0.50, p < 0.0001), C-terminal telopeptide of type I collagen (CTX, r = 0.40, p < 0.0001), and osteocalcin (r = 0.26, p = 0.0411); with a negative correlation to the obesity biomarker leptin (r = -0.31, p = 0.0138). Further, patients with early healing exhibited a peak in CXM at 6 weeks of 1,092 pg/mL (95% confidence interval (CI) [804.8, 1,379]), which was significantly higher than patients with normal healing of 630.8 pg/mL (95% CI [399.9, 861.8]; p = 0.016). We then show that we can reliably transfer this assay from serum collection through venipuncture to a dried blood spot (DBS) collected by finger prick (r = 0.75, Serum = 0.4217*DBS + 181.9, p < 0.0001). Using DBS, the prospective clinical observational study finally suggests that median time-to-peak CXM is at 25.5 days following the elimination of outliers (n = 6) using Robust Regression and Outlier Removal (ROUT, Q = 1%). This study did not find significant differences in CXM expression according to age, sex, or bone that are likely due to the main limitation of an observational study.
CONCLUSION: This rigorous data set supports the future prospective use of the CXM biomarker collected by finger prick in interventional fracture studies and/or in observational studies requiring quantitative assessments of long-bone healing.