目的  探讨慢性肾脏病(chronic kidney disease,CKD)G3～5患者冠状动脉钙化(coronary artery calcification,CAC)的影响因素，构建并验证列线图模型。 方法  采用单中心回顾性病例对照研究，纳入2017年1月─2022年12月于大连医科大学附属第二医院收治的1376例患者。收集临床资料，采用Agatston评分和冠状动脉钙化积分辅助评估软件评定冠状动脉钙化积分(coronary artery calcification score,CACS)，分为非重度钙化(0≤CACS≤400)和重度钙化(CACS＞400)。 结果  建模组和验证组分别纳入963例和413例患者。多因素Logistic回归分析：CKD G5(OR=2.872，95%CI:1.400～6.555,P=0.007)、年龄(OR=1.036，95%CI:1.018～1.055,P＜0.001)、对数B型利钠肽(OR=1.419，95%CI:1.236～1.638,P＜0.001)、低密度脂蛋白胆固醇(OR=1.915，95%CI:1.550～2.381,P＜0.001)、甲状旁腺激素(OR=1.000，95%CI:1.000～1.001,P=0.039)、钙磷乘积(OR=1.321，95%CI:1.117～1.563,   P=0.001)、抑酸药物使用(OR=1.756，95%CI:1.134～2.739,P=0.012)、糖尿病(OR=2.121，95%CI:1.347～3.335,P=0.001)为重度钙化独立危险因素，四碘甲状腺原氨酸(OR=0.920，95%CI:0.857～0.986，      P=0.021)、镁离子(OR =0.028，95%CI:0.007～0.109,P＜0.001)、钠离子(OR=0.923，95%CI:0.878～0.970,P=0.002)、降血脂药物使用(OR=0.290，95%CI:0.180～0.460,P＜0.001)为保护因素。建模组、验证组的曲线下面积(area under the curve,AUC)分别为0.848(95%CI:0.814～0.881)、0.836(95%CI:0.784～0.887)，预测区分度较高。Hosmer-Lemeshow检验和校准曲线表明建模组、验证组模型预测一致性较高(平均绝对误差=0.008、0.036，P=0.478、0.215)。决策曲线分析(decision curve analysis，DCA)分析显示，在建模组(阈值为0.01～0.78)和验证组(阈值为0.04～0.52)中，模型临床净获益较高。 结论  建立的列线图模型可有效预测CKD G3～5患者重度CAC风险，具有良好的鉴别度与临床实用性，可为临床干预提供量化决策依据。

Objective  To explore the influencing factors of coronary artery calcification (CAC) in patients with chronic kidney disease (CKD) stages G3～5 and to develop and validate a nomogram prediction model. Methods  This single-center retrospective case-control study included 1,376 patients admitted to the Second Affiliated Hospital of Dalian Medical University between January 2017 and December 2022. Clinical data were collected, and the Agatston score combined with dedicated CAC quantification software was used to quantify the coronary artery calcification score (CACS). Patients were categorized into the non-severe calcification (0≤CACS≤400) and severe calcification (CACS＞400).  Results  overall, 963 patients were assigned to the training cohort and the remaining 413 to the validation cohort. Multivariate logistic regression analysis identified CKD stage G5 (OR=2.872, 95% CI：1.400～6.555, P=0.007), age (OR=1.036, 95% CI：1.018～1.055, P＜0.001), log-transformed B-type natriuretic peptide (LogBNP) (OR=1.419, 95% CI：1.236～1.638, P＜0.001), low-density lipoprotein cholesterol (OR=1.915, 95% CI：1.550～2.381, P＜0.001), intact parathyroid hormone (OR=1.000, 95% CI：1.000～1.001, P=0.039), calcium–phosphorus product (OR=1.321, 95% CI：1.117～1.563, P=0.001), use of acid-suppressive drugs (OR=1.756, 95% CI：1.134～2.739, P=0.012), and diabetes mellitus (OR=2.121, 95% CI： 1.347～3.335, P=0.001) as independent risk factors for severe CAC. Thyroxine(T4) (OR=0.920, 95% CI：0.857～0.986, P=0.021), serum magnesium (OR=0.028, 95% CI：0.007～0.109, P＜0.001), serum sodium (OR=0.923, 95% CI：0.878～0.970, P=0.002), and use of lipid-lowering drugs (OR=0.290, 95% CI：0.180～0.460, P＜0.001) were protective factors. The area under the curve (AUC) for the training and validation cohorts was 0.848 (95% CI：0.814～0.881) and 0.836 (95% CI：0.784～0.887), respectively, indicating good discriminative ability. The Hosmer–Lemeshow test and calibration curves demonstrated good agreement between predicted and observed outcomes in both cohorts (mean absolute error: 0.008 and 0.036; P=0.478 and 0.215). Decision curve analysis (DCA) showed favorable clinical net benefit across threshold probabilities of 0.01～0.78 in the training cohort and 0.04～0.52 in the validation cohort. Conclusion: The established nomogram effectively predicts the risk of severe CAC in patients with CKD stages G3～5, showing good discrimination and clinical utility, and provides quantitative evidence for individualized clinical decision-making.