糖尿病性周围神经病理性疼痛患者疼痛灾难化风险预测模型的构建与验证

Construction and validation of a risk prediction model for pain catastrophization in patients with diabetic peripheral neuropathic pain

摘要:

目的:分析糖尿病性周围神经病理性疼痛(DPNP)患者疼痛灾难化(PC)的危险因素,构建PC风险预测模型并进行验证。方法:选取本单位2021年1月~2024年12月收治的DPNP患者,根据患者是否发生PC分为PC组与无PC组,采用单因素及多因素Logistic回归分析DPNP患者PC的危险因素,构建PC风险列线图预测模型,绘制工作特征受试者曲线(ROC)及校正曲线评估列线图预测模型的准确性及有效性。结果:共搜集到DPNP患者150例,其中PC患者60例、无PC患者90例。PC组女性占比、小学及以下占比、糖尿病≥10年占比、规律运动占比、多重用药占比、合并症≥4个占比、疼痛持续时间＞6个月占比、DDS评分、PSSS评分、NRS评分均高于无PC组(P＜0.05),PC组疼痛阈值低于无PC组(P＜0.05)；Logistic回归分析显示,女性、小学及以下学历、规律运动、多重用药、合并症≥4个、疼痛阈值高、DDS评分高、PSSS评分高、NRS评分高是DPNP患者发生PC的危险因素(P＜0.05)；风险列线图模型显示,女性、小学及以下学历、规律运动、多重用药、合并症≥4个、疼痛阈值高、DDS评分高、PSSS评分高、NRS评分高会增加DPNP患者发生PC的风险；ROC曲线预测风险列线图模型的AUC为0.880(95%CI:0.746~0.969),灵敏度为90.00%(54/60),特异度为91.11%(82/90),准确度为90.67%(136/150)；风险列线图模型拟合度较高(χ2=3.427,df=6,P=0.330),Bootstrap内部验证结果显示,C-index为0.824。结论:DPNP患者发生PC主要与性别、文化程度、运动、用药、合并症、疼痛程度等因素有关,基于其危险因素构建的风险列线图预测模型对DPNP患者发生PC的预测价值较高。

Abstract:

Objective: To analyze the risk factors of pain catastrophization (PC) in patients with diabetic peripheral neuropathic pain (DPNP), construct a PC risk prediction model and verify it. Method: A total of 150 patients with DPNP admitted to our unit from January 2021 to December 2024 were selected and divided into the PC group (60 cases) and the non-PC group (90 cases) according to whether the patients developed PC. Univariate and multivariate Logistic regression analyses were used to analyze the risk factors of PC in patients with DPNP, and a nomogram prediction model for PC risk was constructed. The operating characteristic receiver ROC curve (ROC) and the correction curve were drawn to evaluate the accuracy and effectiveness of the nomogram prediction model. Result: The proportions of females, primary school education and below, diabetes for ≥10 years, regular exercise, polymedication, comorbidities ≥4, pain duration > 6 months, pain threshold, DDS score, PSSS score, and NRS score in the PC group were all higher than those in the non-PC group (P < 0.05). The pain threshold in the PC group was lower than that in the non-PC group (P < 0.05). Logistic regression analysis showed that female gender, educational attainment of primary school or below, regular exercise, multiple drug use, ≥4 comorbidities, high pain threshold, high DDS score, high PSSS score, and high NRS score were risk factors for PC in patients with DPNP (P < 0.05); The risk nomogram model shows that being female, having an educational attainment of primary school or below, regular exercise, polymedication, having ≥4 comorbidities, a high pain threshold, a high DDS score, a high PSSS score, and a high NRS score will increase the risk of PC in patients with DPNP. The AUC of the risk nomogram model predicted by the ROC curve was 0.880 (95%CI: 0.746-0.969), the sensitivity was 90.00% (54/60), the specificity was 91.11% (82/90), and the accuracy was 90.67% (136/150). The risk nomogram model had a relatively high fitting degree (χ2=3.427, df=6, P=0.330), and the internal validation results of Bootstrap showed that the C-index was 0.824. Conclusion: The occurrence of PC in patients with DPNP is mainly related to factors such as gender, educational level, exercise, medication, comorbidities, and pain degree. The risk nomogram prediction model constructed based on these risk factors has a relatively high predictive value for the occurrence of PC in patients with DPNP.