超重及肥胖人群中不同胰岛素抵抗计算指数准确性评价

尉耘翠，祝茸，夏丽莉，刘芬，何燕

(首都医科大学公共卫生学院，北京100069)



超重及肥胖人群中不同胰岛素抵抗计算指数准确性评价

尉耘翠，祝茸，夏丽莉，刘芬，何燕

(首都医科大学公共卫生学院，北京100069)

摘要：目的观察不同胰岛素抵抗(IR)计算指数评价超重及肥胖人群中IR的准确性，寻找超重及肥胖人群更为理想的判定临界值。方法研究对象为406例超重及肥胖的非糖尿病患者，计算纳入者的稳态模型胰岛素抵抗指数(HOMA-IR)、定量胰岛素敏感性检测指数(QUICKI)、McAuley指数(McA)，以空腹胰岛素≥12 mIU/L为IR金标准，评价各指数判定IR的准确性。结果 406例超重及肥胖者中，IR 114例、非IR 292例。以HOMA-IR≥2.6判定IR时，IR 138例、非IR 268例，其灵敏度为94.74%、特异度为89.73%，约登指数为0.84；以HOMA-IR≥2.68作为判定切点时，其灵敏度为93.3%、特异度为92.9%，约登指数为0.86。以QUICKI≤0.339判定IR时，IR 210例、非IR196例，其灵敏度为100%、特异度为67.12%，约登指数为0.67；以QUICKI≤0.335作为判定切点时，其灵敏度为79.8%、特异度为99.0%，约登指数为0.79。以McA≤5.8判定IR时，IR 403例、非IR3例，其灵敏度为98.25%、特异度为0.34%，约登指数为-1.41；以McA≤4.29作为判定切点时，其灵敏度为78.0%、特异度为82.9%，约登指数为0.61。 结论在超重及肥胖人群中HOMA-IR能更准确地评估IR；在这部分人群中判定IR时应考虑肥胖及脂代谢对胰岛素水平的影响作用，适当调整判定临界值，以便能提高评价IR的准确性。

关键词：超重；肥胖；胰岛素抵抗；胰岛素抵抗指数；稳态模型胰岛素抵抗指数；定量胰岛素敏感性检测指数；McAuley指数

北京市自然科学基金资助项目(713202)。

近年肥胖的发生率逐年上升[1]。研究发现，其与胰岛素抵抗(IR)有关[2]。目前，IR的判定金标准为高胰岛素正葡萄糖钳夹技术[3]，但此法操作复杂，临床普及困难。目前，大量研究采用空腹胰岛素(FINS)≥12 mIU/L作为判定非糖尿病人群IR最为理想的指标[4.5]。另外，学者还提出多项IR简易指数，如稳态模型胰岛素抵抗指数(HOMA-IR)[6]、定量胰岛素敏感性检测指数(QUICKI)[7]及McAuley指数(McA)[8]等。但是，以上指标评价多基于全人群基础，超重及肥胖人群仍缺乏准确的IR效能评价标准。2014年5月～2015年8月，我们分析了HOMA-IR、QUICKI、McA评估超重及肥胖人群IR的灵敏度及特异度，旨在寻找判定IR的理想临界值，为临床早期预防IR提供依据。现报告如下。

1资料与方法

1.1临床资料纳入标准：年龄≥20岁，BMI≥24 kg/m2，自愿参加，未口服影响糖、脂代谢药物，近3个月内未进行任何节食或减肥者。排除标准：患有内分泌疾病、肝肾疾病者。研究对象为406例超重及肥胖的非糖尿病患者，其中男252例、女154例，年龄20～67(43.97±10.28)岁；超重395例，肥胖11例； BMI 24.00～30.86 kg/m2，腰围/臀围比0.70～1.01。

1.2HOMA-IR、QUICKI、McA判定IR的效能分析10 h过夜禁食后于次日清晨抽取静脉血2 mL,采用葡萄糖氧化酶方法测定空腹血糖(FBG)，采用标准酶分光光度技术测定HDL-C、LDL-C、TC、TG，采用ELISA法测定FINS。根据上述结果计算HOMA-IR、 QUICKI、McA。计算公式：HOMA-IR= FINS×(FBG/22.5)；QUICKI= 1/(logFINS+logFBG)；McA=exp(2.63-0.28lnFINS-0.31lnTG)。当HOMA-IR≥2.6、QUICKI≤0.339、McA≤5.8被认定为IR。以FINS≥12 μU/L值为金标准[4,5]评价HOMA-IR、 QUICKI、McA指数判定IR的灵敏度及特异度。

1.3统计学方法采用SPSS17.0统计软件。采用ROC曲线分析法判定诊断界限值。P<0.05为差异有统计学意义。

2结果

以FINS≥12 μU/L值为金标准，406例超重及肥胖者中IR 114例、非IR 292例。以HOMA-IR≥2.6判定IR时，IR 138例、非IR 268例；其灵敏度为94.74%、特异度为89.73%，约登指数为0.84，ROC曲线下面积(AUC)为0.98(95%CI：0.97～0.99,P≤0.001)。当以HOMA-IR≥2.68作为判定切点时，其灵敏度为93.3%、特异度为92.9%，约登指数为0.86。以QUICKI≤0.339判定IR时，IR 210例、非IR196例；其灵敏度为100%、特异度为67.12%，约登指数为0.67,AUC为0.96(95%CI：0.95～0.98,P<0.01)。当以QUICKI≤0.335作为判定切点时，其灵敏度为79.8%、特异度为99.0%，约登指数为0.79。以McA≤5.8判定IR时，IR 403例、非IR3例；其灵敏度为98.25%、特异度为0.34%，约登指数为-1.41；AUC为0.96(95%CI：0.95～0.98，P<0.01)。当以McA≤4.29作为判定切点时，其灵敏度为78.0%、特异度为82.9%，约登指数为0.61。

3讨论

大量研究表明，肥胖、高血压、糖尿病等与IR相关的多种疾病均与FINS水平密切相关[9]。在非糖尿病人群中，FINS与高胰岛素正糖钳夹技术测定的葡萄糖代谢率相关性较好，可用以准确预测机体IR的程度[4]。权威研究将FINS作为最简单有效的IR指标，但糖尿病患者FINS与金标准相关性较差。其主要原因是FINS水平由胰岛素分泌和胰岛素敏感性二者共同决定，而糖尿病患者即使已呈显著IR状态，但会由于胰岛素分泌缺陷而使胰岛素水平不高。因此，糖尿病人群不宜采用FINS作为IR判定标准。以往研究发现，FINS水平与BMI及TG水平均存在显著正相关。现有研究表明，这种相关性不仅与BMI相关，还在很大程度上与脂肪分布情况相关。因此，向心性肥胖的患者更容易患代谢综合征[10]。成年人降低体质量可降低胰岛素水平，增加胰岛素敏感性，提示减轻体质量对防止IR的发生具有一定作用[11]。目前,几种IR计算指标的标准设定均基于全人群基础，未考虑脂代谢的作用。因此,超重及肥胖人群IR临界值的选择应更为保守。

HOMA-IR是另一种被广泛使用的IR指标。既往研究[12，13]在大样本中验证了HOMA-IR与胰岛素钳夹技术测定结果具有较好相关性，可较好地用于临床和流行病学调查[14]。目前，国内多项IR相关研究以所研究人群HOMA-IR 的P75值作为IR判定界值，发现其临界变化很大，也有研究认为其参考值应因性别、年龄而异[15]。国外研究常以HOMA-IR≥2.6或近似值作为IR的参考值，变化范围较小。本研究发现，相对于QUICKI及McA指数，HOMA-IR判定IR的准确性更好。多项研究显示，HOMA-IR在检验IR程度具有较高的灵敏度和特异度，且比QUICKI更为可靠[16]。Conwell等[17]也报道了相似结果，不同时点的HOMA-IR与胰岛素敏感性呈现显著负相关。儿童及青少年人群中FINS、HOMA-IR、QUICKI与静脉葡萄糖耐量实验结果也具有强相关性。本研究发现，超重及肥胖人群中HOMA-IR判定IR准确定最好，而当HOMA-IR≥2.68时其判定IR的准确性最佳。

前期研究表明，QUICKI在不同时间点与敏感性呈显著正相关[17]。本研究中以常用标准QUICKI≤0.339来判定IR时特异度较低。进一步分析我们发现，在超重及肥胖人群中以QUICKI≤0.335作为IR判定切点为时，其准确性最好。因此，超重及肥胖人群判定IR时应考虑降低临界值。而McA指数也是检验IR较为准确的间接指标，尤其是在葡萄糖代谢最小估计模型下其灵敏度和特异度都很高[8,18]。以目前常用标准McA≤5.8判定超重及肥胖人群IR时，特异度仅为0.34%，提示在中国超重及肥胖人群中此IR判定界值偏高，特异度较低，应合理下调判定切点。我们发现在超重及肥胖人群中，当以McA≤4.29作为判定切点时其判定IR准确性最好，能更可靠地判定IR人群，有助于开展代谢性疾病的早期预防。

综上所述，超重及肥胖人群中HOMA-IR具有更好的IR判定准确性，在这部分人群中判定IR时应考虑肥胖及脂代谢对胰岛素水平的影响作用，适当调整判定临界值，以便为IR的早发现、早预防提供准确依据。

参考文献：

[1] Nguyen DM，El-Serag HB. The epidemiology of obesity[J].Gas North Am,2010, 39(1):1-7.

[2] Sobngwi E, Mbanya JC, Unwin NC, et al. Physical activity and its relationship with obesity, hypertension and diabetes in urban and rural Cameroon[J]. Int J Obes Relat Metab Disord, 2002,26(7):1009-1016.

[3] DeFronzo RA, Lilly L. The triumvirate: beta-cell, muscle, liver. A collusion responsible for NIDDM[J]. Diatebets,1988,37(6):667-687.

[4] Laakso M. How good a marker is insulin level for insulin resistance[J]. Am J Epidemiol, 1993,137(9):959-965.

[5] Scott RA, Lagou V, Welch RP, et al. Large-scale association analyses identify new loci influencing glycemic traits and provide insight into the underlying biological pathways[J]. Nat Genet, 2012,44(9):991-1005.

[6] Matthews DR, Hosker JP, Rudenski AS,et al. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man[J]. Diabetologia, 1985,28(7):412-419.

[7] Katz A, Nambi SS, Mather K, et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans[J]. J Clin Endo Metab, 2000,85(7):2402-2410.

[8] Mc Auley KA, Williams SM, Mann JI, et al. Diagnosing insulin resistance in the general population[J]. Diabetes Care, 2001,24(3):460-464.

[9] Haffner SM. Epidemiology of type 2 diabetes: risk factors[J]. Diabetes Care, 1998,21(3):3-6.

[10] Korsic M, FisterK ,Ivankovic D, et al. Visceral obesity[J]. Lijec Vjesn, 2011,133(7-8):284-287.

[11] Srinivasa NRG, Prema G, Priya G, et al. Comparison between serum insulin levels and its resistance with biochemical, clinical and anthropometric parameters in South Indian children and adolescents[J]. Indian J Clin Biochem, 2011,26(1):22-27.

[12] Hffner SM, Kennedy E, Gonzalez C, et al. A prospective analysis of the HOMA model.The Mexico City Diabetes Study[J].Diabetes care, 1996,19(10):1138-1141.

[13] Emoto M, Nishizawa Y, Maekawa K, et al. Homeostasis model assessment as a clinical index of insulin resistance in type 2 diabetic patients treated with sulfonylurea[J]. Diabetes care, 1999,22(5):818-822.

[14] 贾伟平,项坤三,陈蕾,等.上海地区40岁以上自然人群中IR现况及特征分析[J].上海医学,2001,24(4):199-202.

[15] Gayoso DP, Otero GA, Rodriguez AMX, et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a general adult population: effect of gender and age; EPIRCE cross-sectional study[J]. BMC Endo Dis, 2013,13(1):47.

[16] Simsek E, Karabay E, Aras S, et al. Investigating for insulin resistance and type 2 diabetes mellitus in obese children[J]. Turk J Endo Metab, 2005,1(2):17-22.

[17] Conwell LS, Trost SG, Brown WJ, et al. Indexes of insulin resistance and secretion in obese children and adolescents: A validation study[J]. Diabetes Care, 2004,27(2):314-319.

[18] Richard M, Blanche S, Yolette SJ, et al. Fasting insulin levels as a measure of insulin resistance in American blacks[J]. J Med, 2003,34(1-6):31-38.

Evaluation of different indices of insulin resistance in overweight and

obese population

YUYun-cui,ZHURong,XIALi-li,LIUFen,HEYan

(CapitalMedicalUniversity,Beijing100069,China)

Abstract:ObjectiveTo observe the accuracy of several indices in assessing insulin resistance (IR) in the overweight/obesity population and to find more desirable cut-off point of IR. Methods This study involved 406 overweight/obesity participants without overt diabetes. Homeostasis model assessment of insulin resistance (HOMA-IR) index, Quantitative insulin sensitivity check index (QUICKI) and McAuley index (McA index) were calculated and compared with fasting insulin levels (FINS) (≥12 mIU/L, which was the reference criteria for determined IR). ResultsOf the 406 participants, 114 cases of IR and 292 cases of non-IR were found. When using HOMA-IR≥2.6 as the cut-off point of IR, there were 138 cases of IR and 268 cases of non-IR. The sensitivity was 94.74% and the specificity was 89.73%. Youden index was 0.84. When using HOMA-IR≥2.68 as the cut-off point, the sensitivity and specificity was 93.3% and 92.9%, respectively. Youden index was elevated to 0.86. When using QUICKI≤0.339 as the cut-off point of IR, there were 210 cases of IR and 196 cases of non-IR. The sensitivity was 100% and the specificity was 67.12%. Youden index was 0.67. When using QUICKI≤0.335 as the cut-off point, the sensitivity and specificity was 79.8% and 99.0%, respectively. Youden index was increased to 0.79. When using McA≤5.8 as the cut-off point of IR, there were 403 cases of IR and 3 cases of non-IR. The sensitivity was 98.25% and the specificity was only 0.34%. Youden index was -1.41. When using McA≤4.29 as the cut-off point, the sensitivity and specificity was 78.0% and 82.9%, respectively. Youden index was increased to 0.61. ConclusionsHOMA-IR is more accurate in detecting IR in overweight/obese population. The thresholds of these three indices in overweight/obese population should be adjusted appropriately considering the influence of obesity and lipid metabolism on insulin levels to increase the accuracy in evaluating IR.

Key words:overweight; obesity; insulin resistance; insulin resistance index; homeostasis model assessment of insulin resistance index; quantitative insulin sensitivity check index; McAuley index

收稿日期：(2015-09-10)

通信作者简介：何燕(1962-)，女，博士，教授，博士生导师，主要研究方向为慢性病流行病学。E-mail： yanhe1220@126.com

作者简介：第一尉耘翠(1990-)，女，硕士研究生在读，主要研究方向为慢性病流行病学。E-mail： yuyuncui7@126.com

基金项目：国家自然科学基金资助项目(30971178, 30670778)；“十二五”国家科技支撑计划重点资助项目(2015BAI09B01)；

中图分类号：R181.3

文献标志码：A

文章编号：1002-266X(2015)46-0014-03

doi：10.3969/j.issn.1002-266X.2015.46.005