Association of CHI3L1 with bone turnover markers in Iraqi postmenopausal women with osteopenia and osteoporosis Authors: Eman Mal Allah Jasim, Salwa Hameed Naser Al-Rubae’i Published: 2026 ABSTRACT IntroductionChitinase-3-like protein 1 (CHI3L1) has been implicated in bone remodeling, but its relationship with bone turnover markers across different degrees of postmenopausal bone loss remains unclear. This study evaluated CHI3L1 together with CTX-1, NTX-1, and osteonectin (OSN) in Iraqi postmenopausal women with osteopenia (ON) and osteoporosis (OP).Materials and MethodsNinety postmenopausal women aged 45–65 years were classified into control, ON, and OP groups (n = 30 each) according to lumbar spine BMD and T-scores. Serum biomarkers were measured by ELISA. Group differences were assessed by one-way ANOVA with Tukey's HSD test. Pearson correlations were adjusted using the Benjamini–Hochberg false discovery rate, and multivariable regression was used to evaluate associations with BMD after adjustment for age, BMI, WHpR, and WHtR. ROC analysis assessed discriminatory performance.ResultsCHI3L1 and NTX-1 increased progressively from controls to ON and OP, whereas OSN decreased. CTX-1 was elevated in ON and OP versus controls but did not differ between ON and OP. CHI3L1, CTX-1, and NTX-1 were inversely associated with BMD, whereas OSN was positively associated (all p < 0.001). CTX-1 showed the highest AUC for ON versus controls (0.974), while NTX-1 showed the highest AUC for OP versus controls (0.999) and OP versus ON (0.968).ConclusionThese biomarkers showed distinct associations with skeletal status and BMD. CHI3L1 may provide complementary information alongside established bone turnover markers, while NTX-1 exhibited the highest observed AUC for OP in this cohort. Larger independent studies are required for validation. Introduction Osteoporosis (OP) is a progressive skeletal disorder characterized by reduced bone mass and deterioration of bone microarchitecture, which increases skeletal fragility and susceptibility to fractures. Osteopenia (ON), defined by a bone mineral density (BMD) T-score between −1.0 and − 2.5, represents a clinically relevant state of reduced bone mass rather than merely a transitional category before osteoporosis.1, 2, 3 Although fracture risk is greatest in individuals with markedly reduced BMD, fractures also occur in a substantial proportion of postmenopausal women whose T-scores remain above the diagnostic threshold for osteoporosis.4 This is particularly important after menopause, as bone loss accelerates during the menopausal transition and continues into the postmenopausal period.5 Dual-energy X-ray absorptiometry (DEXA) remains the principal clinical method for BMD assessment and densitometric classification of osteoporosis.1, 2 More broadly, biomedical research has increasingly incorporated physiologically relevant experimental models and mechanism-directed molecular approaches to investigate complex disease processes.6, 7 This emphasis on mechanistic investigation is also evident in osteoporosis research, where cellular and zebrafish models have been used to examine bone-loss pathways and osteoprotective responses under different pathological conditions.8, 9 Bone loss after menopause reflects a disturbance in the normal balance between bone formation and resorption. Estrogen deficiency favors osteoclast activity and accelerates bone resorption, gradually shifting skeletal remodeling toward net bone loss.10 Because BMD represents the accumulated structural consequence of this process, biochemical markers of bone turnover can provide complementary information about ongoing skeletal metabolic activity. Elevated concentrations of bone turnover markers have been associated with increased fracture risk in postmenopausal women, and the β-C-terminal telopeptide of type I collagen (β-CTX-I) is currently recognized as a reference biochemical marker of bone resorption.11 Recent evidence has also shown that higher circulating β-CTX levels are associated with lower BMD at clinically relevant skeletal sites and with greater fracture risk in patients with osteoporosis.12 These observations support the assessment of circulating markers alongside BMD to investigate the biological changes accompanying the progression from reduced bone mass to established osteoporosis. At the biochemical level, changes in bone remodeling can be assessed through molecules released during matrix degradation or involved in matrix formation and mineralization. C-terminal telopeptide of type I collagen (CTX-1) and N-terminal telopeptide of type I collagen (NTX-1) are telopeptide fragments generated during osteoclast-mediated degradation of type I collagen and are therefore used as biochemical indicators of bone resorption.13 NTX-1 has also been associated with subsequent loss of bone density, supporting its investigation as an additional indicator of resorptive activity.14 Osteonectin (OSN), also known as secreted protein acidic and rich in cysteine (SPARC), differs from these telopeptide markers because it is an extracellular matrix protein involved in osteoblast function and bone mineralization. Experimental evidence indicates that OSN can regulate osteoblast mineralization in a concentration-dependent manner.15 The combined assessment of CTX-1, NTX-1, and OSN may therefore provide complementary information on collagen degradation and matrix mineralization across different degrees of bone loss. Chitinase-3-like protein 1 (CHI3L1), also known as YKL-40, has recently attracted attention as a potential regulator of bone remodeling. Experimental evidence suggests that its role in skeletal metabolism is complex. Increased CHI3L1 expression has been linked to enhanced osteoclastogenesis through the early growth response 1 (EGR1)/methyltransferase-like 3 (METTL3)/ N6-methyladenosine (m6A)/CHI3L1 regulatory axis in an ovariectomy-induced model of osteoporosis.16 A subsequent mechanistic study showed that CHI3L1 interacts with interleukin-13 receptor alpha 2 (IL13Rα2) and enhances receptor activator of nuclear factor-κB ligand (RANKL)-induced mitogen-activated protein kinase (MAPK) and protein kinase B (AKT) signaling, thereby promoting osteoclast differentiation and bone resorption.17 In contrast, other experimental findings indicate that CHI3L1 can promote osteoblast differentiation and suppress osteoclastogenesis, suggesting that its effects may vary with the cellular and metabolic context.18 This context-dependent behavior provides a rationale for examining circulating CHI3L1 alongside direct markers of collagen degradation, particularly NTX-1, to determine whether they reflect distinct aspects of the remodeling disturbance associated with ON and OP. Recent studies in Iraqi postmenopausal women have examined inflammatory mediators in relation to osteoporosis, indicating growing interest in the contribution of systemic inflammation to postmenopausal bone loss.19 However, the relationship between circulating CHI3L1 and biochemical markers representing different components of bone remodeling remains insufficiently characterized in this population, particularly across both ON and OP. Assessing CHI3L1 alongside CTX-1, NTX-1, and OSN may help determine whether changes in inflammatory and matrix-remodeling activity occur alongside increased bone resorption and declining bone mass. By evaluating CHI3L1 alongside CTX-1, NTX-1, and OSN across normal BMD, ON, and OP within the same Iraqi postmenopausal cohort, the present study addresses this population-specific gap and provides an integrated assessment of inflammatory, collagen-resorptive, and matrix-related changes across varying degrees of bone loss. We hypothesized that circulating concentrations of these markers would differ by skeletal status and that CHI3L1 would be associated with bone resorption markers and BMD. Therefore, the present study aimed to compare serum CHI3L1, CTX-1, NTX-1, and OSN concentrations among Iraqi postmenopausal women with normal BMD, ON, and OP; to examine their relationships with anthropometric and densitometric variables; and to explore their ability to discriminate among the studied skeletal status groups. Materials and methods Study population A total of 90 Iraqi postmenopausal women aged 45–65 years were enrolled in the study. Postmenopausal status was defined as the absence of menstruation for at least 12 consecutive months. The sample size was determined using an a priori power analysis for a one-way analysis of variance (ANOVA) with three independent groups. Assuming a significance level of 0.05, a statistical power of 90% (1 − β = 0.90), and an expected Cohen's effect size of f = 0.40, the required total sample size was <90 participants. Therefore, 30 participants were allocated to each group, resulting in a total sample size of 90 participants, allocated equally into three groups: control (n = 30), ON (n = 30), and OP (n = 30). Group classification was based on BMD and T-score measurements obtained by DEXA. Eligible participants were postmenopausal women without metabolic skeletal disorders other than ON or OP, as applicable. Exclusion criteria included renal or hepatic dysfunction, gastrointestinal disorders associated with malabsorption, malignancy, endocrine disorders known to affect bone metabolism, and current or recent use of medications known to influence bone turnover or vitamin D metabolism. Anthropometric measurements were obtained under standardized conditions using calibrated equipment. Body weight was measured in kilograms, standing height in meters, and waist and hip circumferences in centimeters. BMI was calculated as weight divided by height squared (kg/m2). WHpR was calculated as waist circumference divided by hip circumference, whereas WHtR was calculated as waist circumference divided by height. All participants provided written informed consent before enrollment. The study was approved by the Ethics Committee of the College of Science, Mustansiriyah University (Ref. BCSMU/0125/0052C; October 2025). Participants were recruited at Al-Wasiti Teaching Hospital, Baghdad, Iraq, between 15 October 2025 and 10 January 2026. Specimen collection For biochemical analysis, approximately 5 mL of venous blood was drawn from each participant into a gel-containing collection tube. Samples were left at 25 °C for 30 min to allow clot formation, then centrifuged at 1500 ×g for 10 min. The serum fraction was then carefully separated, aliquoted, and maintained at −80 °C until the laboratory measurements were performed. Biochemical assays Serum concentrations of CHI3L1, CTX-1, NTX-1, and OSN were quantified using commercially available double-antibody sandwich ELISA kits supplied by Shanghai Coon Koon Biotech Co., Ltd., China. The catalog numbers, analytical ranges, and sensitivities were as follows: CHI3L1, CK-bio-10,841, 1–32 ng/mL and 0.1 ng/mL; CTX-1, CK-bio-11,060, 1–48 ng/mL and 1.0 ng/mL; NTX-1, CK-bio-11,061, 0.5–16 nmol/L and 0.1 nmol/L; and OSN, CK-bio-12,735, 0.5–10 ng/mL and 0.1 ng/mL, respectively. According to the manufacturer, the intra-assay and inter-assay coefficients of variation for the CHI3L1, CTX-1, and OSN kits were < 7% and < 10%, respectively. All assays were performed according to the manufacturer's instructions. Standards and serum samples were added to antibody-precoated wells, followed by horseradish peroxidase (HRP)-conjugate reagent and incubation at 37 °C for 60 min. The plates were washed five times, after which chromogenic solutions A and B were added, and the plates were incubated at 37 °C for 15 min, protected from light. The reaction was terminated with stop solution, and absorbance was measured at 450 nm within 15 min using a microplate reader (HUMAN, Germany). Serum concentrations were calculated using assay-specific standard curves generated from the corresponding standards. All standards and serum samples were assayed in duplicate, and the mean absorbance of the duplicate wells was used to calculate the final concentration. Bone assessment BMD and T-scores were measured using a DEXA bone analyzer (DMS STRATOS, France). Measurements were obtained at the lumbar spine (L1–L4). Participants were classified according to their T-scores as follows: control, T-score ≥ −1.0; ON, T-score between −1.0 and − 2.5; and OP, T-score ≤ −2.5. These criteria were applied consistently in assigning participants to the three study groups. Statistical analyses The normality of continuous variables was assessed using the Shapiro–Wilk test, and no significant deviations from normality were detected (p > 0.05). Continuous variables are presented as mean ± standard deviation (SD). Differences among the control, ON, and OP groups were evaluated using one-way analysis of variance (ANOVA), followed by Tukey's honestly significant difference (HSD) post hoc test for pairwise comparisons. Eta squared (η2) was calculated to quantify the magnitude of the overall group effect. Relationships among biochemical, anthropometric, and densitometric variables were examined using Pearson's correlation coefficient. To account for multiple testing in the correlation analyses, p-values were adjusted using the Benjamini–Hochberg false discovery rate (FDR) procedure. FDR correction was applied separately to 12 correlations for each biomarker, excluding self-correlations, and an FDR-adjusted p-value <0.05 was considered statistically significant. Multivariable linear regression analyses were performed to evaluate the independent associations of CHI3L1, CTX-1, NTX-1, and OSN with BMD after adjustment for age, BMI, WHpR, and WHtR. Separate regression models were constructed for each biomarker, with BMD entered as the dependent variable. Unstandardized regression coefficients (B), standardized coefficients (β), 95% confidence intervals (CI), and coefficients of determination (R2 and adjusted R2) were reported. ROC curve analysis was performed to evaluate the discriminatory performance of the investigated biomarkers for distinguishing ON from controls, OP from controls, and OP from ON. AUC, optimal cut-off values, sensitivity, and specificity were determined for each comparison. Unless otherwise specified for FDR-adjusted correlation analyses, statistical significance was defined as p < 0.05. Statistical analyses were performed using IBM SPSS Statistics version 26.0, whereas GraphPad Prism version 10.6.1 was used for graphical presentation and visualization of the results. Results Comparison of anthropometric and densitometric characteristics among the control, ON, and OP groups No statistically significant differences were observed among the three groups in age, WHpR, or WHtR (p = 0.480, p = 0.214, and p = 0.710, respectively). The corresponding η2 values were 0.025, 0.035, and 0.008, indicating small or negligible group effects for these variables. BMI differed significantly among the groups (p = 0.036; η2 = 0.074). Post hoc analysis showed that this difference was confined to the comparison between the ON and OP groups (p = 0.030), whereas the control group did not differ significantly from either the ON or OP group (p = 0.646 and p = 0.212, respectively). In contrast, pronounced differences were observed in BMD and T-score across the three groups (both p < 0.0001). The effect sizes were substantial for both BMD (η2 = 0.821) and T-score (η2 = 0.840), indicating that a large proportion of the variability in these measurements was attributable to the skeletal-status group. Pairwise comparisons showed significant differences among the control, ON, and OP groups for both parameters (all p < 0.0001). BMD and T-score decreased progressively from the control group to ON, reaching their lowest values in the OP group. The detailed findings are presented in Table 1, and representative DEXA images from each study group are shown in Fig. 1.Table 1Anthropometric and densitometric characteristics of the control, ON, and OP groups.VariablesControlMean ± SDOsteopeniaMean ± SDOsteoporosisMean ± SDp-valueη2Number303030––Age (years)58.13 ± 4.6056.70 ± 6.1356.53 ± 5.940.4800.025BMI (Kg/m2)31.71 ± 6.3833.08 ± 5.9229.11 ± 5.440.0360.074WHpR0.90 ± 0.070.94 ± 0.160.94 ± 0.080.2140.035WHtR0.64 ± 0.090.66 ± 0.080.66 ± 0.100.7100.008BMD (g/cm2)1.08 ± 0.100.87 ± 0.040.72 ± 0.06< 0.00010.821T-score0.14 ± 0.81−1.73 ± 0.32−3.06 ± 0.51< 0.00010.840Fig. 1DEXA reports of selected postmenopausal women from the study cases: A: control; B: osteopenia patient; and C: osteoporosis patient.Fig. 1 Table 1Anthropometric and densitometric characteristics of the control, ON, and OP groups.VariablesControlMean ± SDOsteopeniaMean ± SDOsteoporosisMean ± SDp-valueη2Number303030––Age (years)58.13 ± 4.6056.70 ± 6.1356.53 ± 5.940.4800.025BMI (Kg/m2)31.71 ± 6.3833.08 ± 5.9229.11 ± 5.440.0360.074WHpR0.90 ± 0.070.94 ± 0.160.94 ± 0.080.2140.035WHtR0.64 ± 0.090.66 ± 0.080.66 ± 0.100.7100.008BMD (g/cm2)1.08 ± 0.100.87 ± 0.040.72 ± 0.06< 0.00010.821T-score0.14 ± 0.81−1.73 ± 0.32−3.06 ± 0.51< 0.00010.840 Anthropometric and densitometric characteristics of the control, ON, and OP groups. DEXA reports of selected postmenopausal women from the study cases: A: control; B: osteopenia patient; and C: osteoporosis patient. Data are presented as mean ± SD. Comparisons among the three groups were performed using one-way ANOVA followed by the HSD post hoc test for pairwise comparisons. η2 was calculated as the effect size for the overall group comparison. A p-value <0.05 was considered statistically significant. Comparison of CHI3L1, CTX-1, NTX-1, and OSN among ON, OP, and control groups The serum concentrations of all investigated biomarkers differed significantly among the three study groups (all overall p < 0.0001). CHI3L1 concentrations increased progressively from the control group (8.46 ± 1.24 ng/mL) to the ON group (9.94 ± 1.09 ng/mL), reaching the highest level in the OP group (13.71 ± 3.16 ng/mL). Pairwise comparisons showed significant differences between control and ON (p = 0.006), control and OP (p < 0.0001), and ON and OP (p < 0.0001). The overall group effect for CHI3L1 was large (η2 = 0.544). CTX-1 concentrations also differed significantly among the groups (overall p < 0.0001). Higher concentrations were observed in both the ON (16.50 ± 1.67 ng/mL) and OP (15.76 ± 2.02 ng/mL) groups compared with the control group (12.98 ± 1.02 ng/mL) (both p < 0.0001). However, the difference between the ON and OP groups was not statistically significant (p = 0.186). NTX-1 showed a progressive increase across the groups, from 5.30 ± 0.54 nmol/L in the control group to 5.95 ± 0.82 nmol/L in the ON group and 7.96 ± 0.59 nmol/L in the OP group. All pairwise comparisons were statistically significant (p < 0.0001). A large group effect was observed for NTX-1 (η2 = 0.752). In contrast, OSN concentrations decreased progressively with worsening skeletal status, from 4.81 ± 0.52 ng/mL in the control group to 3.55 ± 0.76 ng/mL in ON and 2.75 ± 0.39 ng/mL in OP. Significant differences were observed for all pairwise comparisons (p < 0.0001), with a large overall effect size (η2 = 0.693). The comparisons of serum biomarker concentrations are summarized in Table 2 and illustrated in Fig. 2.Table 2Comparison of serum CHI3L1, CTX-1, NTX-1, and OSN concentrations among the control, ON, and OP groups.VariablesControlMean ± SDOsteopeniaMean ± SDOsteoporosisMean ± SDp-valueη2CHI3L1 (ng/mL)8.46 ± 1.249.94 ± 1.0913.71 ± 3.16< 0.00010.544CTX-1 (ng/mL)12.98 ± 1.0216.50 ± 1.6715.76 ± 2.02< 0.00010.474NTX-1 (nmol/L)5.30 ± 0.545.95 ± 0.827.96 ± 0.59< 0.00010.752OSN (ng/mL)4.81 ± 0.523.55 ± 0.762.75 ± 0.39< 0.00010.693Fig. 2Serum concentrations of CHI3L1, CTX-1, NTX-1, and OSN in the control, ON, and OP groups. Data are presented as mean ± SD. The indicated p-values represent pairwise post hoc comparisons between groups.Fig. 2 Table 2Comparison of serum CHI3L1, CTX-1, NTX-1, and OSN concentrations among the control, ON, and OP groups.VariablesControlMean ± SDOsteopeniaMean ± SDOsteoporosisMean ± SDp-valueη2CHI3L1 (ng/mL)8.46 ± 1.249.94 ± 1.0913.71 ± 3.16< 0.00010.544CTX-1 (ng/mL)12.98 ± 1.0216.50 ± 1.6715.76 ± 2.02< 0.00010.474NTX-1 (nmol/L)5.30 ± 0.545.95 ± 0.827.96 ± 0.59< 0.00010.752OSN (ng/mL)4.81 ± 0.523.55 ± 0.762.75 ± 0.39< 0.00010.693 Comparison of serum CHI3L1, CTX-1, NTX-1, and OSN concentrations among the control, ON, and OP groups. Serum concentrations of CHI3L1, CTX-1, NTX-1, and OSN in the control, ON, and OP groups. Data are presented as mean ± SD. The indicated p-values represent pairwise post hoc comparisons between groups. Data are presented as mean ± SD. Comparisons among the three groups were performed using one-way ANOVA followed by HSD post hoc pairwise comparisons. Eta squared (η2) was calculated to quantify the magnitude of the overall group effect. A p-value <0.05 was considered statistically significant. Multivariable regression analysis Multivariable linear regression analyses were performed to evaluate the associations of the investigated serum biomarkers with BMD after adjustment for age, BMI, WHpR, and WHtR. CHI3L1 was significantly and inversely associated with BMD (B = −0.039, β = −0.702, 95% CI: −0.047 to −0.030, p < 0.001), with the model explaining 51.8% of the variance in BMD (R2 = 0.518; adjusted R2 = 0.489). CTX-1 also showed a significant inverse association with BMD (B = −0.043, β = −0.566, 95% CI: −0.056 to −0.029, p < 0.001), accounting for 36.4% of the variance (R2 = 0.364; adjusted R2 = 0.326). A stronger inverse association was observed for NTX-1 (B = −0.095, β = −0.754, 95% CI: −0.114 to −0.076, p < 0.001), with the corresponding model explaining 56.9% of the variance in BMD (R2 = 0.569; adjusted R2 = 0.543). In contrast, OSN was positively associated with BMD (B = 0.124, β = 0.765, 95% CI: 0.102 to 0.146, p < 0.001), and the model including OSN showed the highest explanatory power among the four biomarkers (R2 = 0.619; adjusted R2 = 0.597). The complete regression results are presented in Table 3.Table 3Multivariable linear regression analysis of the associations between serum biomarkers and BMD after adjustment for age, BMI, WHpR, and WHtR.BiomarkerBSEStandardized β95% CIp-valueR2Adjusted R2CHI3L1−0.0390.004−0.702−0.047 to −0.030<0.0010.5180.489CTX-1−0.0430.007−0.566−0.056 to −0.029<0.0010.3640.326NTX-1−0.0950.010−0.754−0.114 to −0.076<0.0010.5690.543OSN0.1240.0110.7650.102–0.146<0.0010.6190.597 Table 3Multivariable linear regression analysis of the associations between serum biomarkers and BMD after adjustment for age, BMI, WHpR, and WHtR.BiomarkerBSEStandardized β95% CIp-valueR2Adjusted R2CHI3L1−0.0390.004−0.702−0.047 to −0.030<0.0010.5180.489CTX-1−0.0430.007−0.566−0.056 to −0.029<0.0010.3640.326NTX-1−0.0950.010−0.754−0.114 to −0.076<0.0010.5690.543OSN0.1240.0110.7650.102–0.146<0.0010.6190.597 Multivariable linear regression analysis of the associations between serum biomarkers and BMD after adjustment for age, BMI, WHpR, and WHtR. Separate regression models were constructed for each biomarker, with BMD entered as the dependent variable. B, unstandardized regression coefficient; SE, standard error; β, standardized regression coefficient; CI, confidence interval; R2, coefficient of determination. A p-value <0.05 was considered statistically significant. Correlation In the ON group, Pearson correlation analysis followed by Benjamini–Hochberg FDR correction showed a significant positive correlation between CHI3L1 and CTX-1 (r = 0.550, FDR-adjusted p = 0.020). Although CHI3L1 was positively correlated with NTX-1 (r = 0.450), this association did not remain statistically significant after correction for multiple comparisons. No significant correlations were observed between CHI3L1 and OSN, BMD, or the evaluated anthropometric variables after FDR correction. Similarly, CTX-1, NTX-1, and OSN showed no statistically significant associations with age, anthropometric measurements, or BMD in the ON group after correction for multiple testing. The complete correlation results are presented in Table 4.Table 4Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the ON group.ParametersCHI3L1CTX-1NTX-1OSNrp-valuerp-valuerp-valuerp-valueAge (years)−0.0090.962−0.1590.7600.1290.927−0.0760.916Height (m)−0.0510.8610.1380.7600.0720.9270.0700.916Weight (kg)0.2230.3540.1450.760−0.0990.927−0.3060.600BMI (kg/m2)0.2500.3540.1080.760−0.0550.927−0.3110.600Waist (cm)0.1550.5510.1090.7600.3260.4720.1090.916Hip (cm)−0.0890.7680.0900.763−0.0040.9830.0410.916WHpR0.2500.354−0.0090.9620.1720.9270.0370.916WHtR−0.2360.354−0.1840.7600.1060.927−0.1460.916BMD (g/cm2)−0.2500.3540.0390.914−0.0940.9270.1290.916CHI3L1 (ng/mL)––––––––CTX-1 (ng/mL)0.5500.020––––––NTX-1 (nmol/L)0.4500.0760.1100.760––––OSN (ng/mL)−0.2500.354−0.1310.760−0.0200.983–– Table 4Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the ON group.ParametersCHI3L1CTX-1NTX-1OSNrp-valuerp-valuerp-valuerp-valueAge (years)−0.0090.962−0.1590.7600.1290.927−0.0760.916Height (m)−0.0510.8610.1380.7600.0720.9270.0700.916Weight (kg)0.2230.3540.1450.760−0.0990.927−0.3060.600BMI (kg/m2)0.2500.3540.1080.760−0.0550.927−0.3110.600Waist (cm)0.1550.5510.1090.7600.3260.4720.1090.916Hip (cm)−0.0890.7680.0900.763−0.0040.9830.0410.916WHpR0.2500.354−0.0090.9620.1720.9270.0370.916WHtR−0.2360.354−0.1840.7600.1060.927−0.1460.916BMD (g/cm2)−0.2500.3540.0390.914−0.0940.9270.1290.916CHI3L1 (ng/mL)––––––––CTX-1 (ng/mL)0.5500.020––––––NTX-1 (nmol/L)0.4500.0760.1100.760––––OSN (ng/mL)−0.2500.354−0.1310.760−0.0200.983–– Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the ON group. Values are presented as Pearson's correlation coefficients (r) and Benjamini–Hochberg false discovery rate (FDR)-adjusted p-values. FDR correction was applied separately to 12 correlations for each biomarker, excluding self-correlations. An FDR-adjusted p-value <0.05 was considered statistically significant. In the OP group, Pearson correlation analysis after FDR correction showed that CHI3L1 remained significantly associated with several biochemical, anthropometric, and densitometric variables. CHI3L1 was positively correlated with CTX-1 (r = 0.700, p < 0.0001), NTX-1 (r = 0.600, p < 0.0001), and WHpR (r = 0.450, p = 0.026). In contrast, significant inverse correlations were observed between CHI3L1 and BMD (r = −0.450, p = 0.026), BMI (r = −0.407, p = 0.026), height (r = −0.407, p = 0.043), and OSN (r = −0.450, p = 0.026). NTX-1 was also significantly and inversely correlated with BMI (r = −0.487, p = 0.036). No statistically significant correlations were observed between CTX-1, NTX-1, or OSN and the remaining anthropometric or densitometric variables after correction for multiple comparisons. The complete FDR-adjusted correlation results for the OP group are presented in Table 5.Table 5Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the OP group.ParametersCHI3L1CTX-1NTX-1OSNrp-valuerp-valuerp-valuerp-valueAge (years)0.0150.939−0.0280.8850.2070.326−0.0060.973Height (m)−0.4070.043−0.2350.312−0.3120.2370.1280.750Weight (kg)0.2960.2000.2240.3120.3830.148−0.2110.526BMI (kg/m2)−0.4070.026−0.3200.202−0.4870.036−0.2720.348Waist (cm)0.2060.3600.3410.1610.2420.263−0.1550.710Hip (cm)−0.0850.7850.0810.8030.0870.7050.0430.973WHpR0.4500.0260.4110.1440.2680.237−0.3160.348WHtR0.0340.938−0.0490.868−0.0300.876−0.0130.973BMD (g/cm2)−0.4500.026−0.3600.161−0.2690.2370.0690.957CHI3L1 (ng/mL)––––––––CTX-1 (ng/mL)0.700<0.0001––––––NTX-1 (nmol/L)0.600<0.00010.2640.271––––OSN (ng/mL)−0.4500.026−0.3390.161−0.2890.237–– Table 5Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the OP group.ParametersCHI3L1CTX-1NTX-1OSNrp-valuerp-valuerp-valuerp-valueAge (years)0.0150.939−0.0280.8850.2070.326−0.0060.973Height (m)−0.4070.043−0.2350.312−0.3120.2370.1280.750Weight (kg)0.2960.2000.2240.3120.3830.148−0.2110.526BMI (kg/m2)−0.4070.026−0.3200.202−0.4870.036−0.2720.348Waist (cm)0.2060.3600.3410.1610.2420.263−0.1550.710Hip (cm)−0.0850.7850.0810.8030.0870.7050.0430.973WHpR0.4500.0260.4110.1440.2680.237−0.3160.348WHtR0.0340.938−0.0490.868−0.0300.876−0.0130.973BMD (g/cm2)−0.4500.026−0.3600.161−0.2690.2370.0690.957CHI3L1 (ng/mL)––––––––CTX-1 (ng/mL)0.700<0.0001––––––NTX-1 (nmol/L)0.600<0.00010.2640.271––––OSN (ng/mL)−0.4500.026−0.3390.161−0.2890.237–– Pearson correlation coefficients between serum biomarkers and anthropometric and densitometric variables in the OP group. Values are presented as Pearson's correlation coefficients (r) and Benjamini–Hochberg false discovery rate (FDR)-adjusted p-values. FDR correction was applied separately to 12 correlations for each biomarker, excluding self-correlations. An FDR-adjusted p-value <0.05 was considered statistically significant. ROC-based discriminatory performance of serum biomarkers ROC curve analysis was performed to evaluate the ability of the investigated serum biomarkers to discriminate women with ON from controls. CTX-1 exhibited the highest discriminatory performance, with an AUC of 0.974 (p < 0.0001). At a cut-off value of 14.62 ng/mL, CTX-1 showed 90.0% sensitivity and 96.7% specificity. OSN also demonstrated high discriminatory performance, with an AUC of 0.918 (p < 0.0001); a cut-off value of 4.33 ng/mL yielded 90.0% sensitivity and 93.3% specificity. CHI3L1 showed moderate-to-good discrimination between the ON and control groups, with an AUC of 0.802 (p < 0.0001). At a cutoff of 8.80 ng/mL, sensitivity was 86.7%, and specificity was 56.7%. NTX-1 exhibited an AUC of 0.760 (p = 0.001), with 80.0% sensitivity and 66.7% specificity at a cut-off value of 5.42 nmol/L. Thus, among the investigated biomarkers, CTX-1 exhibited the highest AUC for distinguishing ON from controls in this cohort. The ROC findings are summarized in Table 6.Table 6ROC curve analysis of serum biomarkers for discrimination between women with ON and controls.ParameterAUCBest cut-off pointSensitivitySpecificityp-ValueCHI3L1 (ng/mL)0.8028.8086.7%56.7%< 0.0001CTX-1 (ng/mL)0.97414.6290%96.7%< 0.0001NTX-1 (nmol/L)0.7605.4280%66.7%0.001OSN (ng/mL)0.9184.3390%93.3%< 0.0001 Table 6ROC curve analysis of serum biomarkers for discrimination between women with ON and controls.ParameterAUCBest cut-off pointSensitivitySpecificityp-ValueCHI3L1 (ng/mL)0.8028.8086.7%56.7%< 0.0001CTX-1 (ng/mL)0.97414.6290%96.7%< 0.0001NTX-1 (nmol/L)0.7605.4280%66.7%0.001OSN (ng/mL)0.9184.3390%93.3%< 0.0001 ROC curve analysis of serum biomarkers for discrimination between women with ON and controls. AUC, area under the ROC curve. Sensitivity and specificity correspond to the reported optimal cut-off values. A p-value <0.05 was considered statistically significant. For discrimination between the OP and control groups, all investigated biomarkers showed high AUC values. NTX-1 exhibited the highest AUC in this cohort (AUC = 0.999, p < 0.0001). At a cut-off value of 6.06 nmol/L, NTX-1 achieved 100% sensitivity and 96.67% specificity. CHI3L1 also showed a high discriminatory performance, with an AUC of 0.986 (p < 0.0001); a cut-off value of 10.41 ng/mL yielded 93.3% sensitivity and 100% specificity. OSN demonstrated a similarly high AUC of 0.984 (p < 0.0001), with 100% sensitivity and 96.67% specificity at a cut-off value of 3.44 ng/mL. CTX-1 showed an AUC of 0.897 (p < 0.0001), with 80.0% sensitivity and 93.33% specificity at a cut-off value of 14.08 ng/mL. Thus, NTX-1 exhibited the highest AUC for distinguishing women with OP from controls in this cohort. The corresponding ROC findings are presented in Table 7.Table 7ROC curve analysis of serum biomarkers for discrimination between women with OP and controls.ParameterAUCBest cut-off pointSensitivitySpecificityp-ValueCHI3L1 (ng/mL)0.98610.4193.3%100%< 0.0001CTX-1 (ng/mL)0.89714.0880.00%93.33%< 0.0001NTX-1 (nmol/L)0.9996.06100%96.67%< 0.0001OSN (ng/mL)0.9843.44100%96.67%< 0.0001 Table 7ROC curve analysis of serum biomarkers for discrimination between women with OP and controls.ParameterAUCBest cut-off pointSensitivitySpecificityp-ValueCHI3L1 (ng/mL)0.98610.4193.3%100%< 0.0001CTX-1 (ng/mL)0.89714.0880.00%93.33%< 0.0001NTX-1 (nmol/L)0.9996.06100%96.67%< 0.0001OSN (ng/mL)0.9843.44100%96.67%< 0.0001 ROC curve analysis of serum biomarkers for discrimination between women with OP and controls. AUC, area under the ROC curve. Sensitivity and specificity correspond to the reported optimal cut-off values. A p-value <0.05 was considered statistically significant. For direct discrimination between the ON and OP groups, NTX-1 exhibited the highest AUC among the investigated biomarkers (AUC = 0.968, p < 0.001). At a cut-off value of 6.99 nmol/L, NTX-1 achieved 96.7% sensitivity and 90.0% specificity. CHI3L1 also showed good discriminatory performance, with an AUC of 0.896 (p < 0.001); a cut-off value of 11.14 ng/mL provided 80.0% sensitivity and 80.0% specificity. OSN showed an AUC of 0.839 (p < 0.001), with 80.0% sensitivity and 70.0% specificity at a cut-off value of 3.004 ng/mL. In contrast, CTX-1 did not significantly discriminate between the ON and OP groups (AUC = 0.599, p = 0.188). Thus, NTX-1 exhibited the highest AUC for distinguishing OP from ON in this cohort. The corresponding ROC findings are presented in Table 8.Table 8ROC curve analysis of serum biomarkers for discrimination between women with OP and ON.ParameterAUCBest cut-off pointSensitivitySpecificityP-ValueCHI3L1 (ng/mL)0.89611.1480%80%<0.001CTX-1 (ng/mL)0.599–––0.188NTX-1 (nmol/L)0.9686.9996.7%90%<0.001OSN (ng/mL)0.8393.00480%70%<0.001 Table 8ROC curve analysis of serum biomarkers for discrimination between women with OP and ON.ParameterAUCBest cut-off pointSensitivitySpecificityP-ValueCHI3L1 (ng/mL)0.89611.1480%80%<0.001CTX-1 (ng/mL)0.599–––0.188NTX-1 (nmol/L)0.9686.9996.7%90%<0.001OSN (ng/mL)0.8393.00480%70%<0.001 ROC curve analysis of serum biomarkers for discrimination between women with OP and ON. AUC, area under the ROC curve. Sensitivity and specificity correspond to the reported optimal cut-off values. A p-value <0.05 was considered statistically significant. Discussion The present study showed distinct biomarker patterns across skeletal status groups among Iraqi postmenopausal women. CHI3L1 and NTX-1 increased across the control, ON, and OP groups, whereas OSN decreased, and CTX-1 was elevated in both ON and OP, with no significant difference between them. All four biomarkers remained significantly associated with BMD after adjustment for age, BMI, WHpR, and WHtR. CHI3L1 was also associated with bone resorption markers, particularly in OP. ROC analysis showed that CTX-1 had the highest AUC for ON versus controls, whereas NTX-1 had the highest AUC for OP versus controls and for OP versus ON. These findings suggest that the biomarkers may reflect complementary aspects of postmenopausal bone remodeling. The progressive increase in circulating CHI3L1 from the control group to ON and OP, together with its inverse association with BMD after adjustment for age and anthropometric indices, supports a relationship between CHI3L1 and the severity of skeletal deterioration in this cohort. This pattern is consistent with population-based evidence showing that higher circulating YKL-40 concentrations in women were associated with poorer bone-related ultrasound parameters and higher CTX concentrations, suggesting that circulating CHI3L1 may accompany increased bone resorptive activity rather than simply reflect reduced bone mass.20 At the mechanistic level, Wang et al. demonstrated that CHI3L1 was upregulated in ovariectomized osteoporotic mice and that activation of the EGR1/METTL3/m6A/CHI3L1 axis promoted osteoclast differentiation, providing a biological basis for the higher CHI3L1 concentrations observed in women with greater bone loss in the present study.16 This interpretation is further supported by evidence that CHI3L1 can interact with IL13Rα2 on osteoclast precursors and enhance RANKL-induced MAPK and AKT signaling, thereby increasing osteoclast formation and expression of osteoclast-related genes.17 Nevertheless, the role of CHI3L1 in bone metabolism appears to be context-dependent. Park et al. reported that CHI3L1 enhanced bone morphogenetic protein 2 (BMP2)- mediated osteoblast differentiation, increased osteoprotegerin (OPG) expression, suppressed osteoclastogenesis, and was lower in patients with osteoporosis, which contrasts with the higher circulating concentrations observed in our OP group.18 The discrepancy may reflect differences in biological compartment, experimental model, population characteristics, or the stage and metabolic context of bone loss. Accordingly, the present findings should not be interpreted as evidence that CHI3L1 exerts an exclusively pro-resorptive effect; rather, they suggest that increased circulating CHI3L1 in these postmenopausal women is associated with a remodeling environment characterized by lower BMD and greater resorptive activity. The behavior of the collagen degradation markers provided further evidence of increased resorptive activity across the skeletal states studied, although CTX-1 and NTX-1 did not exhibit identical patterns. CTX-1 concentrations were significantly higher in both ON and OP than in controls, but no significant difference was detected between ON and OP, whereas NTX-1 increased progressively across the three groups and retained a strong inverse association with BMD in the adjusted regression model. This distinction is biologically plausible because CTX and NTX originate from different terminal regions of type I collagen and, although both reflect collagen degradation, their circulating concentrations are influenced by different analytical and biological sources of variation. Earlier comparative work showed that serum CTX and NTX are correlated measures of bone resorption but do not respond identically to changes in skeletal turnover, supporting their interpretation as related rather than interchangeable markers.21 The progressive increase in NTX-1 observed in the present study is consistent with longitudinal evidence from the SWAN cohort, in which higher NTX concentrations during early postmenopause were independently associated with faster subsequent loss of lumbar spine and femoral neck BMD.22 In parallel, population-based data on β-CTX have shown that increased bone turnover is more frequent in untreated women with osteoporosis than in those with osteopenia or normal BMD, confirming the close relationship between collagen degradation and skeletal deterioration after menopause.23 The absence of a further rise in CTX-1 from ON to OP in our cohort therefore should not be interpreted as evidence that resorption ceases to increase with disease severity. CTX is particularly sensitive to pre-analytical conditions, including time of sampling and fasting status; substantial circadian variation has been demonstrated, and fasting markedly reduces this variability.24 Such biological variation, together with the modest sample size, may partly contribute to the overlap in CTX-1 concentrations between the ON and OP groups. In contrast, the more consistent increase in NTX-1 and its stronger adjusted association with BMD in the present cohort suggest that NTX-1 may have captured differences in collagen degradation across these skeletal-status categories more clearly than CTX-1, although this interpretation requires confirmation in larger cohorts using standardized pre-analytical conditions. The progressive reduction in OSN concentrations from the control group to ON and OP, in contrast to the increasing pattern of collagen resorption markers, was further supported by the positive association between OSN and BMD in the adjusted regression model. This finding is consistent with the established involvement of OSN in bone matrix formation and mineralization, rather than in collagen degradation. In a randomized study of postmenopausal women with osteoporosis, Zhang et al. reported a positive correlation between circulating SPARC and lumbar spine BMD at baseline; moreover, treatment with recombinant parathyroid hormone (PTH) increased both plasma SPARC and lumbar BMD, and the changes in these two parameters were positively correlated.25 More recent experimental evidence also supports a contribution of SPARC to skeletal integrity. Kaneda et al. demonstrated that loss of SPARC adversely affected bone structural and biomechanical properties in mice, with reductions in femoral BMD being particularly evident in females.26 These observations are consistent with the present finding that lower circulating OSN levels accompany more advanced bone loss, and that higher OSN levels are independently associated with greater BMD. Nevertheless, because circulating OSN levels, rather than bone-specific expression or osteoblast activity, were measured, the present data cannot establish that reduced OSN directly causes impaired bone formation. Instead, the decline in OSN may reflect alterations in extracellular matrix turnover and mineralization occurring alongside the progressive loss of bone mass. The associations between CHI3L1 and anthropometric variables observed in the OP group suggest that its circulating concentration may also reflect a broader metabolic and inflammatory environment accompanying skeletal deterioration. In particular, CHI3L1 was positively associated with WHpR but inversely associated with BMI after FDR correction. These apparently divergent relationships should be interpreted with caution, because BMI reflects overall body mass relative to height, whereas WHpR more specifically captures body fat distribution. Recent population-based evidence has shown that circulating YKL-40 is associated with insulin resistance and metabolic syndrome, even after accounting for obesity and waist circumference, indicating that CHI3L1 may reflect metabolic dysfunction not adequately captured by general adiposity alone.27 This interpretation is also relevant to the emerging relationship between CHI3L1 and musculoskeletal deterioration. In patients with type 2 diabetes, Zhang et al. found that higher serum CHI3L1 concentrations were independently associated with sarcopenia, and that the prevalence of sarcopenia increased across CHI3L1 categories.28 Although that study addressed sarcopenia rather than osteoporosis, it supports the possibility that elevated circulating CHI3L1 levels accompany a broader metabolic–musculoskeletal phenotype rather than represent a bone-specific signal. Experimental findings further indicate that CHI3L1 can interact with pathways involved in insulin signaling and lipid metabolism. In a high-fat-diet model, hepatic CHI3L1 expression increased in parallel with impaired insulin signaling and lipid accumulation, whereas genetic deletion or functional inhibition of CHI3L1 improved hepatic insulin sensitivity and reduced lipid deposition.29 These observations provide a possible metabolic context for the association between circulating CHI3L1 and bone loss, since metabolic dysfunction, inflammation, and skeletal remodeling may occur concurrently after menopause. Nevertheless, metabolic variables such as fasting glucose, insulin resistance, lipid profile, hepatic fat accumulation, and body composition were not assessed in the present study. Therefore, the observed relationships with WHpR and BMI do not establish a metabolic mechanism and should instead serve as a basis for future studies that incorporate direct measures of metabolic status and body composition. The ROC analyses revealed distinct discriminatory profiles among the investigated biomarkers. CTX-1 exhibited the highest AUC for distinguishing ON from controls, whereas NTX-1 showed the highest AUC for both OP versus controls and the direct comparison between OP and ON. In particular, the AUC of 0.968 for NTX-1 in the OP-versus-ON comparison suggests that the progressive increase in this marker was not limited to distinguishing normal from reduced bone mass but also tracked the difference between the two degrees of skeletal impairment observed in this cohort. Previous studies support the ability of collagen degradation products to distinguish women by skeletal status, although the magnitude of discrimination has varied substantially across populations and analytical methods. De la Piedra et al. reported high ROC areas for urinary CTX and NTX when comparing postmenopausal women with osteoporosis with healthy postmenopausal controls, with AUC values of 0.957 and 0.934, respectively.30 More recently, Feng et al. found that β-CTX discriminated perimenopausal women with bone loss with relatively high sensitivity and specificity, while combining β-CTX with estradiol yielded an AUC of 0.950.31 These findings support the biological plausibility of substantial separation in bone-resorption markers between skeletal-status groups, but they also demonstrate that performance estimates depend on the population studied, the biological specimen, assay characteristics, and the reference definition of bone loss. The population-specific nature of the present findings should also be considered when interpreting the biomarker concentrations and their discriminatory performance. Bone turnover markers are influenced not only by menopausal status and skeletal condition but also by ethnicity, body composition, environmental exposures, dietary patterns, and analytical conditions. In the Study of Women's Health Across the Nation (SWAN) cohort, longitudinal changes in NTX across the menopausal transition differed significantly among ethnic groups, with the magnitude of the increase greatest in Japanese women and smallest in African American women, even after adjustment for relevant covariates.32 Similar heterogeneity has been demonstrated among Asian postmenopausal populations. In the Pan-Asia Menopause study, baseline concentrations of several bone turnover markers differed significantly across nine ethnic groups, and these differences persisted independently of age and BMI.33 A systematic review of studies conducted across Asian countries likewise documented substantial inter-country and even intra-country variation in serum bone turnover marker concentrations, including CTX, with differences in participant selection and analytical methodology identified as potential contributors.34 These observations are particularly relevant to the present Iraqi cohort because population-specific reference data for CHI3L1, NTX-1, CTX-1, and OSN in postmenopausal women remain limited. Consequently, the concentrations and ROC-derived cut-off values identified here should not be assumed to apply directly to women from other ethnic or geographic populations. Rather, the present study provides initial population-specific evidence that requires replication in geographically diverse cohorts using standardized sampling and assay procedures. Several limitations should be acknowledged. The relatively small sample size and single-center design may limit the precision of the estimated associations and reduce the generalizability of the findings. Although the regression models were adjusted for age, BMI, WHpR, and WHtR, residual confounding by unmeasured factors, such as years since menopause, vitamin D status, calcium intake, physical activity, and other lifestyle variables, cannot be excluded. In addition, the lack of standardized fasting and sampling time may have influenced bone turnover marker measurements, particularly CTX-1. Finally, the ROC-derived cut-off values were obtained from the same cohort and require validation in larger, independent, multicenter populations before clinical application. Conclusions The present study demonstrated distinct changes in circulating CHI3L1, CTX-1, NTX-1, and OSN across different degrees of bone loss in Iraqi postmenopausal women. Higher CHI3L1, CTX-1, and NTX-1 concentrations and lower OSN concentrations were associated with reduced BMD, and these associations remained significant after adjustment for age, BMI, WHpR, and WHtR. The relationships of CHI3L1 with bone resorption markers further suggest that it may reflect inflammatory and remodeling processes accompanying skeletal deterioration. ROC analysis showed that CTX-1 had the highest AUC for distinguishing ON from controls, whereas NTX-1 exhibited the highest AUC for both OP versus controls and the direct comparison between OP and ON. These findings support the potential value of combining CHI3L1 with established bone turnover markers to characterize postmenopausal bone loss. However, the identified associations and cut-off values should be interpreted as exploratory and require confirmation in larger, multicenter, and independently validated cohorts before clinical application. CRediT authorship contribution statement Eman Mal Allah Jasim: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Salwa Hameed Naser Al-Rubae’i: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ethical approval Ethical approval was granted by the Ethics Committee of the College of Science, Mustansiriyah University (Ref. BCSMU/0125/0052C; October 2025). Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Prof. Dr. Salwa Hameed Naser Al-Rubae'i reports administrative support and equipment, drugs, or supplies were provided by Mustansiriyah University. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.