Recent experimental work has identified CXCL9 as a promoter for the differentiation of osteoclast progenitors into osteoclasts, with resultant bone resorption. However, no human study has validated an association between this chemokine and osteoporosis or fracture risk. We conducted a matched case–control study nested in the prospective, population-based Singapore Chinese Health Study. Fifty-five men and 119 women with incident hip fractures, occurring median 6.2 years after blood collection, were matched individually to controls by age at recruitment, sex, and duration of blood storage. Serum chemokines, CXCL9 and CXCL10, were measured using immunoassays. Multivariable conditional logistic regression models that included age at blood collection, body mass index, smoking, and diabetes as covariates were used to estimate odds ratios (OR) and 95% confidence intervals (CI) for association with hip fracture risk. Predictive utility of chemokine for hip fracture risk was examined by comparing area under receiver operating characteristic curves (AUC) between prognostic models with and without the chemokine. Increasing CXCL9 levels were associated with increasing hip fracture risk in men but not in women (p_interaction = 0.002);​ comparing extreme quartiles, the OR (95% CI) in the highest quartile was 10.35 (1.90–56.39) in men (p_trend = 0.002) but 1.46 (0.59–3.60) in women (p_trend = 0.32). Adding CXCL9 to a prognostic model that already incorporated age and other risk factors improved the AUC (95% CI) from 0.65 (0.55–0.76) to 0.74 (0.65–0.83) for the predictive utility of hip fractures in men but not in women. Conversely, the association between CXCL10 and hip fracture risk was not statistically significant in either sex.