2型糖尿病患者便秘风险列线图预测模型的建立与验证

目的 探讨2型糖尿病患者发生便秘的影响因素及风险列线图预测模型的建立与验证。方法 选取2018年12月—2019年12月因2型糖尿病于我院内分泌门诊就诊的320例患者作为研究对象,采用自制调查表回顾性收集患者临床资料,采用二分类Logistic回归分析其发生便秘的影响因素。根据筛选出的影响因素建立列线图预测模型,并对模型的预测准确性及区分度进行验证。结果 年龄≥60岁、每周运动时间<2 h、血糖控制水平不佳、补充钙剂、自服通便药及焦虑抑郁情绪是2型糖尿病患者发生便秘的危险因素（P<0.05）。基于以上6项独立危险因素,建立预测2型糖尿病患者发生便秘的列线图模型,并经验证型该模型具有较好的准确度（H-L检验：χ²=5.183,P=0.486）和区分度（AUC=0.795,95%CI：0.770~0.820）。结论 本研究基于2型糖尿病患者发生便秘的6项独立危险因素分别是年龄、每周运动时间、血糖控制水平、补充钙剂、通便药及焦虑抑郁,建立的风险列线图预测模型具有可靠的预测价值,有利于医务人员筛查高风险患者和制定针对性的护理干预措施。     

妊娠期糖尿病并发先兆子痫的风险预测列线图模型构建与预防

目的 分析妊娠期糖尿病(GDM)患者并发先兆子痫的风险因素,并依此构建风险预测列线图模型,同时提出对应预防建议。方法 回顾性分析本院2019年3月～2022年3月收治的223例GDM患者临床资料,均随访至分娩结束,根据随访期间GDM患者并发先兆子痫情况将其分为并发组31例和非并发组192例,比较两组一般资料。采用多因素Logistic回归分析法分析GDM患者并发先兆子痫的危险因素,并基于此采用R3.4.5软件包绘制列线图模型,且绘制受试者工作特征(ROC)曲线对列线图模型的预测效能进行评估,绘制校准度曲线图,并采用Bootstrap法检验预测列线图预测模型的一致性。结果 多因素Logistic回归分析结果显示,年龄、BMI、FBG、TC、TG、Hcy为影响GDM患者并发先兆子痫的独立危险因素(P<0.05)。将上述危险因素作为预测指标,构建GDM患者并发先兆子痫的风险预测列线图模型,ROC曲线分析结果显示,曲线下面积(AUC)为0.830(95%CI为0.736～0.901),灵敏度为73.33%,特异度为82.22%,最大约登指数为0.555,临床应用预测准确率为87.10%;...     

列线图预测代谢综合征患者心功能减退的风险

目的建立列线图预测模型识别代谢综合征(MS)患者中可能出现心功能减退的高风险患者。方法将本院511例MS患者随机分为训练队列369例和验证队列142例。收集患者的基本信息、血液学检测指标及超声心动图指标。根据超声心动图结果,将训练队列患者划分为左室舒张功能减低(LVDD)组132例与no-LVDD组237例。通过多因素Logistic回归筛选与LVDD相关的独立影响因素。分别基于训练队列和验证队列对列线图模型的性能进行内部与外部验证。结果腰围、合并症数量、糖化血红蛋白(HbA1c)、整体纵向应变(GLS)是MS患者心功能减退的独立影响因素(P<0.05)。所构建的列线图预测模型在内部及外部验证中都具有良好的区分度(AUC:内部验证0.910,外部验证0.745)和一致性(Hosmer-Lemeshow检验:内部验证χ^(2)=5.261,P=0.729;外部验证χ^(2)=6.142,P=0.434)。决策曲线分析显示,在训练和验证队列中,模型均具有较高净收益。结论通过综合腰围、合并症数量、HbA1c和GLS建立的列线图模型可有效预测MS患者心功能减退的风险。     

血清SFRP-4在2型糖尿病肾病患者不同病理阶段的表达水平及对预后的预测价值

目的 探究血清分泌型卷曲相关蛋白-4(SFRP-4)在2型糖尿病肾病(T2DKD)患者中不同病理阶段的表达水平及对预后的预测价值。方法 选择2020年1月至2022年1月于该院就诊的140例T2DKD患者为T2DKD组,140例单纯2型糖尿病(T2DM)患者为T2DM组,另选择同期于该院体检的140例健康者为对照组。对比3组的一般资料和生化指标,并比较不同肾损伤程度T2DKD患者的一般资料及生化指标。依据预后情况,将140例T2DKD患者分为预后良好组(n=87)和预后不良组(n=53),通过单因素和多因素Logistic分析T2DKD患者预后不良的独立危险因素。应用Logistic回归模型结合限制性立方样条模型(RCS)分析T2DKD患者的SFRP-4表达水平与T2DKD患者预后不良的剂量-反应关系。依据独立因素构建列线图预测模型,并对模型进行验证。结果 随着肾损伤程度的增加,T2DKD患者T2DM病程、空腹血糖(FPG)、糖化血红蛋白(HbA1c)、肌酐(Scr)、尿素氮(BUN)、尿酸(UA)、尿β2-微球蛋白(β2-MG)和SFRP-4逐渐增高(P<0.05),白蛋白(A...     

基于列线图及机器学习的免疫检查点抑制剂相关性肺炎风险预测模型构建及验证

目的运用机器学习算法及列线图, 构建和验证免疫检查点抑制剂相关性肺炎(CIP)风险预测模型, 旨在为更好的辅助临床护理人员筛查CIP的高危人群, 提供准确直观的方法。方法采用回顾性病例对照研究。选取2019年1月至2022年2月南方医科大学珠江医院就诊的230例使用免疫检查点抑制剂治疗的肿瘤患者, 使用医院电子病历系统收集患者的资料。应用5种机器学习算法和列线图构建预测模型, 在独立测试集进行模型的验证, 最后依据评价指标AUC、准确率等评估预测模型的区分度及稳定性。结果 6种模型均提示, 肺部基础疾病、吸烟史、血清白蛋白值≤35 g/L、胸部放疗史是促进CIP发生的重要影响因素。K最近邻、支持向量机(SVM)、朴素贝叶斯、决策树和随机森林构建的CIP预测模型的AUC分别为0.647、0.696、0.930、0.870、0.934。列线图构建的模型AUC为0.813, 预测性能较好, 但低于机器学习算法中表现最佳的随机森林模型(AUC=0.934)。结论与列线图相比, 基于机器学习算法建立的CIP的风险预测模型具有更高的诊断价值, 但列线图构建的模型可更直观评估患者风险, 建议在列线图...     

结合NGAL、Cys-C的列线图模型对冠心病患者PCI术后病情无进展生存率的预测价值

目的 结合中性粒细胞明胶酶相关脂质运载蛋白(NGAL)、胱抑素C(Cys-C)建立用于预测冠心病患者经皮冠状动脉介入治疗(PCI)术后病情无进展生存率的列线图模型,并验证此模型的预测价值。方法 纳入2018年1月至2021年1月在该院行PCI术的冠心病患者161例,按PCI术后1年内是否发生心血管不良事件(MACE)分为MACE组和非MACE组,比较两组一般临床资料及PCI术后NGAL、Cys-C水平,将单因素分析中有意义的指标代入Cox风险回归模型进行多因素分析,获得影响冠心病患者PCI术后1年发生MACE的独立预测因素,重点分析NGAL、Cys-C与冠心病患者PCI术后1年发生MACE的相关性,建立预测冠心病患者PCI术后病情无进展生存率的列线图模型,并应用R软件绘制列线图,以自助抽样法和校准曲线进行列线图模型的内部验证。结果 多因素分析结果显示,糖尿病、多支病变、NGAL、Cys-C为影响冠心病患者PCI术后1年发生MACE的独立预测指标(P<0.05),基于4个独立预测因素及其影响权重建立预测冠心病患者PCI术后病情无进展生存率的列线图模型并绘制列线图,列线图模型的区分度...     

中性粒细胞与淋巴细胞比值对诊断早期2型糖尿病肾病的价值分析

目的探讨中性粒细胞与淋巴细胞比值(NLR)在2型糖尿病并发早期糖尿病肾病(DKD)患者中的变化及应用价值。方法依据预估肾小球滤过率(eGFR)及尿清蛋白与肌酐比值(ACR),将651例2型糖尿病(T2DM)患者分为早期糖尿病肾病(DKD)组376例,无肾病单纯糖尿病(DM)组275例;另选无糖尿病、年龄相近的健康体检者150例作为对照组。比较DKD、DM及对照组的NLR水平以及DKD、DM组各生化指标的水平;研究各检测指标与早期DKD相关性,用Logistic回归分析影响因素,绘制受试者工作特征曲线(ROC曲线),评估NLR对早期DKD的预测能力。结果 DKD组、DM组及对照组NLR分别为2.98±0.41、2.37±0.75以及2.01±0.60,C反应蛋白(CRP)水平分别为(8.43±2.79)mg/L、(7.54±2.33)mg/L及(3.71±2.59)mg/L。DKD组、DM组NLR及CRP值显著高于对照组,差异有统计学意义(P0.05);DM组NLR值、ACR值显著低于DKD组,差异均有统计学意义(P0.05);DM组CRP值低于DKD组,差异无统计学意义(P0.05);Logistic回归分析表明,NLR升高是DKD的危险因素(P0.01,OR=3.016);ROC曲线显示曲线下面积(AUC)=0.803(95%CI:0.770～0.835,P0.001),NLR对早期DKD有重要诊断价值,当NLR界值取2.675时,其诊断效率最高,灵敏度为51.1%,特异度为86.7%。结论 NLR升高是早期DKD的高危因素,对早期DKD有重要诊断价值。     

超声特征联合免疫组化构建列线图预测乳腺癌新辅助化疗疗效

目的 探讨原发灶超声特征联合免疫组化对预测乳腺癌新辅助化疗(NAC)疗效的价值。方法 分析接受术前NAC的80例乳腺癌患者的超声影像及临床信息,根据术后病理分为有效组和无效组。筛选NAC疗效的独立预测因素,分别构建超声特征预测模型、免疫组化预测模型、超声联合免疫组化的列线图预测模型。比较模型的预测价值及临床收益,并应用Bootstrap法对模型进行内部验证。结果 病灶大小、侧方声影、ER/PR、HER-2为NAC疗效的独立预测变量,基于4个变量绘制列线图预测模型,受试者工作特征曲线下面积(AUC)为0.868,高于单独应用超声特征(AUC=0.778)和免疫组化(AUC=0.776)的预测模型。决策曲线显示在广泛阈值范围内,列线图模型的临床净收益最高。经Bootstrap法迭代1 000次进行内部验证,平均AUC为0.870,模型稳定性良好。结论 基于NAC前原发灶的超声特征联合免疫组化构建列线图可以在一定程度上早期预测乳腺癌NAC疗效。     

基于超声内镜及临床数据预测食管胃底静脉曲张治疗后再出血风险的列线图模型建立

目的 探讨基于超声内镜(EUS)及临床数据预测食管胃底静脉曲张(EGV)治疗后再出血风险的列线图模型建立。方法 回顾性分析2020年1月~2022年12月收治的186例EGV患者,根据治疗后出血情况分为出血组(102例)、未出血组(84例),收集两组患者EUS参数及临床数据资料,采用单因素分析、多因素Logistic回归分析EGV治疗后再出血的影响因素,并构建风险预测模型,采用受试者工作特征曲线(ROC曲线)分析该模型的预测价值。结果 多因素Logistic回归分析结果显示,红色征、腹水、CTP评分、para-EVC直径、奇静脉直径为EGV治疗后再出血的独立危险因素(P<0.05)。基于以上5个指标构建的风险预测模型的AUC为0.974,敏感度、特异度为93.1%、98.8%,均高于模型中各单项指标,且该风险预测模型拟合程度较好(Hosmer-Lemeshow x_²=7.303,P=0.504)。结论 红色征、腹水、CTP评分、para-EVC直径、奇静脉直径为EGV治疗后再出血的独立预测因素,基于以上5个指标构建的风险预测模型对EGV治疗后再出血具有较高预测价值,能为EGV治疗后再出血防治提供理论依据。     

BNP、CRP、CysC等指标对心肌梗死患者预后不良的预测价值及列线图预测模型构建

目的 探讨胱抑素C(CysC)、C反应蛋白(CRP)、脑钠肽(BNP)等指标对心肌梗死患者预后的预测价值并构建列线图预测模型。方法 将2020年5月至2022年4月北京大学第三医院秦皇岛医院收治的236例心肌梗死患者纳入研究并进行为期1年的随访。根据患者是否发生不良心脑血管事件将其分为预后不良组(48例)与预后良好组(188例)。记录两组患者的一般资料,经独立样本t检验或χ²检验进行单因素分析,筛选出可能用于预测心肌梗死患者预后的指标。对单因素分析中差异有统计学意义的检测指标进行受试者工作特征(ROC)曲线分析,分析这些指标对心肌梗死患者预后的预测价值。采用多因素Logistic回归分析心肌梗死患者预后的影响因素。采用R语言软件4.0“rms”程序包构建心肌梗死患者预后的列线图预测模型,使用校正及决策曲线对列线图预测模型进行内部验证及临床预测效能评估。结果 两组患者病变范围、BNP、CRP、CysC、同型半胱氨酸(Hcy)比较,差异有统计学意义(P<0.05)。ROC曲线分析显示,BNP、CRP、CysC、Hcy的曲线下面积(AUC)分别为0.885、0.8...     

论糖尿病下肢血管病变治疗策略的选择

糖尿病下肢血管病变（lower extremity arterial disease, LEAD）是糖尿病常见的慢性并发症之一,是糖尿病患者致残、致死的主要原因,目前关于LEAD的治疗各地、各医院方案各异,本文将对目前治疗LEAD的策略进行总结比较,为临床治疗提供依据。     

Cardiovascular risk factors and disease management in type 2 diabetic patients with diabetic nephropathy

OBJECTIVE: The purpose of this study was to assess the prevalence of cardiorenal risk factors, their management in a routine clinical setting, and the actual achievement of international guideline targets in a large cohort of type 2 diabetic patients with diabetic nephropathy. RESEARCH DESIGN AND METHODS: A multicentric cross-sectional study was performed in the Campania region in Italy to evaluate cardiorenal risk factors and their management in light of international guidelines. Overall, 28,550 diabetic patients were screened in the 21 participating centers; 847 (348 male and 449 female) patients with type 2 diabetes and a clinical diagnosis of diabetic nephropathy were recruited. RESULTS: Of these subjects, 749 had microalbuminuria and 98 had macroalbuminuria. Targets for blood pressure, HbA(1c), LDL cholesterol, HDL cholesterol, and triglycerides were reached in, respectively, 17.5, 32.3, 30.7, 47, and 55.2% of the patients. Chronic renal failure (glomerular filtration rate <60 ml/min) was revealed in 41% and anemia in 23.8% of the patients. CONCLUSIONS: This is the first study to investigate a large cohort of type 2 diabetic patients with early and moderate diabetic nephropathy strictu sensu. Notably, impaired renal function can be often diagnosed in these patients even in the presence of microalbuminuria. Thus, clinical diagnosis of diabetic nephopathy allows us to identify a group of patients at very high cardiorenal risk, for whom care is really difficult. We suggest that a correct diagnosis of diabetic nephropathy should always be made and that sodium intake and anemia should be routinely evaluated in these patients.     

Poncet's disease: An unusual presentation of tuberculosis in a diabetic lady

Authors describe a 53-year-old woman who presentedto their diabetes clinic with a three week history of multiple painful and swollen joints. She had been diagnosed with type 2 diabetes 5 years back. On examination, both knee joints and left ankle were swollen. A soft tissue swelling appeared over the medial end of the left clavicle few days later. Rheumatoid arthritis, collagen vascular diseases and other common causes of polyarthritis were ruled out by appropriate investigations. Non steroidal anti-inflammatory drugs failed to give satisfactory pain relief and the arthritis persisted. Conventional cultures of synovial fluid samples including cultures for tuberculosis were negative. Computed tomography showed a space occupying lesion involving the left sternoclavicular joint. Fine needle aspiration from the lesion was performed and acidfast bacilli were demonstrated in the smear using ZiehlNeelsen stain. The explanation of her arthritis was therefore tuberculous arthritis in left sternoclavicular joint and reactive arthritis in the rest of the joints. A diagnosis of Poncet's disease was considered in her case. We treated her with standard anti-tuberculosis drugs and the arthritis resolved within a few days. She remained symptom-free at her 2 years' follow-up.     

Hyperhomocysteinemia, paraoxonase activity and risk of coronary artery disease

OBJECTIVES: Paraoxonase-1 (PON1) detoxifies homocysteine thiolactone (HcyT) in human blood and could thus delay the development of atherosclerosis. We investigated (a) PON1 activity and polymorphisms, and (b) the relationship between PON1 activity, homocysteine (Hcy) and the severity of CAD patients in Tunisian population. DESIGN AND METHODS: We used PCR-RFLP analysis to detect the Q192R and L55M variants of the PON1 gene in 100 patients with CAD and in 120 healthy controls. Paraoxonase activity was measured spectrophotometrically using phenylacetate as a substrate. Total plasma homocysteine concentrations were determined by direct chemiluminescence assay. RESULTS: We found an increased Hcy level in CAD patients compared to the control group (15.86+/-8.63 vs. 11.9+/-3.25 micromol/L respectively, P<0.001), and a decrease in PON1 activity in CAD patients as compared to the control group (117+/-56 vs. 181+/-73 U/mL respectively, P<0.001). PON1 Q192R and L55M polymorphisms were not associated with the presence of CAD (P=0.592, P=0.294, respectively). However, we found that PON1 activity is lower with the PON1 192RR than with PON1 192QQ genotypes in the study population. Furthermore, there were no association between PON1 L55M polymorphism and PON1 activity. We showed a significant decrease in PON1 activity in CAD patients presenting 0- to 3-vessel stenosis (155+/-39; 135+/-36; 103+/-22; 77+/-24 U/mL, respectively; P<0.001). CONCLUSION: In this study, we showed that low PON1 activity is associated with the PON1 192RR genotypes and associated with the severity of CAD in the Tunisian population. We hypothesize that high level of Hcy together with low PON1 activity results in an increased plasma HcyT plasma concentration leading to protein N-homocysteinylation and the development and progression of atherosclerosis.     

2型糖尿病患者高甘油三酯血症腰围表型与糖尿病肾脏病的相关性

目的 探讨2型糖尿病(T2DM)患者中高甘油三酯血症腰围表型(HTWC表型)与糖尿病肾脏病(DKD)的关系。方法 将753例T2DM患者以血甘油三酯≥1.7 mmoL/L、腰围男性≥85 cm或女性≥80 cm为切点,分为甘油三酯和腰围正常组、单纯高甘油三酯组、单纯腹型肥胖组、高甘油三酯血症腰围表型(HTWC)组,比较4组间DKD的发生率及各生化指标,并进行Logistic回归分析。结果 HTWC组患者体质量指数、腰围、臀围、收缩压、空腹血糖、尿酸、甘油三酯、尿白蛋白与肌酐比值较非HTWC组升高(均P<0.01);HTWC组DKD的发病率较其他3组高(P<0.05)。多因素Logistic回归分析表明,HTWC表型是T2DM患者合并DKD的独立危险因素(OR=2.014,95%CI=1.183～3.625)。结论 HTWC表型是T2DM患者合并DKD的独立危险因素。     

冠心病合并糖尿病患者晚期糖基化终产物测定分析

目的 探讨血清晚期糖基化终产物(AGEs)水平在冠心病(CAD)合并糖尿病(DM)患者中的意义.方法 诊断明确的CAD合并DM患者82例,根据冠状动脉造影结果分为单支病变组、双支病变组及3支病变组,分别测定血清AGEs,空腹血糖(FPG)、血脂,计算体质量指数(BMI),对3组间AGEs、FPG、血脂、BMI进行比较.结果 AGEs水平随冠状动脉病变支数增加而增高,单支病变组(35.86±11.28)ng/L,双支病变组(43.52±17.29)ng/L,3支病变组(62.09±14.23)ng/L,差异有统计学意义(P<0.05);Spearman相关分析显示,AGEs与总胆固醇(TC)、低密度脂蛋白胆固醇(LDL-C)、FPG均呈正相关(P<0.05);多元逐步线性回归分析显示,LDL-C、FPG与AGEs独立相关.结论 AGEs促进DM患者冠状动脉粥样硬化的发展,控制FPG及LDL-C可能延缓此作用.     

2型糖尿病合并早期糖尿病肾病的相关危险因素分析

目的 探讨2型糖尿病患者早期糖尿病肾病（EDN）的相关危险因素,为其防治提供临床依据。方法 将171例2型糖尿病患者按尿白蛋白排泄率（UAER）水平分为单纯糖尿病组（SDM,UAER〈20 μg/min）和早期糖尿病肾病组（EDN,UAER 20～200 μg/min）,观察两组年龄、病程、血压、体重指数（BMI）、空腹血糖（FPG）、糖化血红蛋白（HbA1c）、空腹C肽（FCP）、甘油三酯（TG）、总胆固醇（TC）、高密度脂蛋白胆固醇（HDL-C）、低密度脂蛋白胆固醇（LDL-C）等指标的差异,并进行比较分析。结果 EDN组BMI、LDL-C均较SDM组升高,而HDL-C降低,差异有统计学意义（P〈0.05）,而年龄、病程、FPG、HbA1c、FCP、TG、TC等差异无统计学意义（P〉0.05）;在50例血脂正常的患者中,SDM组的年龄、病程、TC、LDL-C等指标均低于EDN组（P〈0.05）;应用χ2检验发现,EDN在血脂正常和异常组中的比例差异有统计学意义（P〈0.05）;Logistic回归分析显示糖尿病病程、高BMI和高LDL-C与EDN的发病密切相关。结论 在2型糖尿病肾病的防治过程中,要注意控制病程、肥胖、血脂等危险因素。     

A prospective study of obesity and risk of coronary heart disease among diabetic women

OBJECTIVE: To examine the relationship of obesity, measured as BMI, and weight change to incidence of coronary heart disease (CHD) among women with diabetes. RESEARCH DESIGN AND METHODS: We followed 5,897 women with type 2 diabetes in the Nurses' Health Study for < or = 20 years. Women were aged 40-74 years and had no history of cardiovascular disease or cancer at the beginning of the follow-up period. BMI values from three time points (age 18 years, year 1976, and current) were derived from the reported height (1976) and corresponding reported weight. Weight changes between age 18 years and 1976 and after diagnosis of diabetes were calculated. Women reported diagnoses of diabetes and CHD every 2 years. Incident CHD cases were confirmed by medical record review. RESULTS: During follow-up, we documented 418 incident cases of CHD (236 of nonfatal myocardial infarction and 182 of fatal CHD). After adjustment for age, smoking, and other coronary risk factors, current BMI was strongly associated with increased risk of CHD among diabetic women. The multivariate relative risks across increasing categories of BMI (<23.0, 23.0-24.9, 25.0-26.9, 27.0-29.9, 30.0-34.9, and > or = 35.0 kg/m(2)) were 1.0, 1.58, 1.85, 1.95, 2.80, and 3.21, respectively (P for trend <0.001). Increasing BMI values from age 18 years to 1976, before diagnosis of diabetes, were also positively associated with risk of CHD. Weight gain before the diagnosis of diabetes was related to increased risk of CHD. In contrast, weight change after diagnosis of diabetes was not associated with risk of CHD. CONCLUSIONS: These findings provide strong evidence that obesity and weight gain before diagnosis of diabetes are associated with future risk of CHD among women with type 2 diabetes.     

Glycosylated hemoglobin A1c predicts coronary artery disease in non‐diabetic patients

BackgroundIn primary care, there is a need for simple and cost‐effective tool that will allow the determination of the risk of coronary artery disease (CAD). We aimed to research the value of glycosylated hemoglobin (HbA1c) in the prediction of coronary artery disease.MethodsPatients admitted to the outpatient clinic of the Cardiology for angiography were retrospectively screened. Patients with diabetes or with HbA1c of 6.5 or above were excluded. Comparative HbA1c data were obtained according to the stenosis groups. Logistic regression analysis was used to investigate the risk factors affecting stenosis positivity.ResultsOf the study group, 120 patients were without any stenosis in any coronary artery, 56 patients were with >50% stenosis in one coronary artery, and 71 patients were with >50% stenosis in more than one coronary artery. There was a statistically significant difference between HbA1c measurements according to the degree of stenosis (P = .001 and P < .01, respectively). The odd ratio for HbA1c was 6.260 (95% CI: 3,160‐12,401). According to the stenosis positivity, the cutoff point for HbA1c was found to be 5.6 and above. In the regression analysis, HbA1c was an independent risk factor for CAD. One unit increase in HbA1c level increases the risk of stenosis up to 12.4‐fold (95% CI: 5,990‐25,767).ConclusionThe study showed HbA1c can be used as an independent marker in determining the probability and severity of coronary artery disease in non‐diabetic individuals and as a useful marker in primary care predicting CAD.