Bonomo M, Gandini M L, Mastropasqua A, Begher C, Valentini U, Faden D, Morabito A
Diabetes and Metabolic Unit, I Division of Obstetrics and Gynecology, Niguarda Hospital, Milan, Italy.
Am J Obstet Gynecol. 1998 Jul;179(1):179-85. doi: 10.1016/s0002-9378(98)70270-4.
Our purpose was to examine the validity of 140 mg/dL cutoff value in oral glucose challenge test screening for gestational diabetes mellitus when including in the group to be identified women fulfilling more inclusive Carpenter and Coustan criteria for 100-g oral glucose tolerance testing interpretation and gravid women with borderline glucose intolerance.
We reanalyzed data of a multicenter study performed on 704 pregnant women screened at the twenty-fourth to twenty-eighth week with a 50-g oral glucose challenge test followed by a universal 100-g oral glucose tolerance test. We used receiver-operator characteristic curve analysis, assembling positive and negative groups according to the different criteria adopted in oral glucose tolerance test interpretation (National Diabetes Data Group or Carpenter-Coustan) and in assignment of women with borderline glucose intolerance. Besides the statistical cutoff value, defined by the Youden index (Sensitivity + Specificity - 1), we also selected a "high-sensitivity" cutoff value, identified by the maximal sensitivity associated with >70% specificity.
With use of National Diabetes Data Group criteria, the statistical and high-sensitivity cutoff values were set at 142 mg/dL when the positive group included only women with positive oral glucose tolerance test results and at 140 mg/dL when it also included subjects with borderline glucose intolerance. With use of Carpenter-Coustan criteria, the statistical cutoff value was set at 141 mg/dL when the positive group included only women with positive oral glucose tolerance test results and at 140 mg/dL when it also included subjects with borderline glucose intolerance; the high-sensitivity cutoff value was set at 140 mg/dL when the positive group included only women with positive oral glucose tolerance test results and at 136 mg/dL when it also included subjects with borderline glucose intolerance.
We suggest maintaining the 140 mg/dL oral glucose challenge test threshold if the diagnostic target is to recognize only women with positive results of the oral glucose tolerance test. To prevent perinatal risks in pregnancies complicated by borderline glucose intolerance, with Carpenter-Coustan criteria a lower cutoff value (136 mg/dL) could be hypothesized to improve test sensitivity, allowing more extensive diagnosis of "borderline" subjects; however, the higher economic costs resulting from the increased false-positive rate and the limited improvement obtainable in sensitivity currently do not justify its generalized use.
我们的目的是在口服葡萄糖耐量试验筛查妊娠期糖尿病时，研究140mg/dL的临界值在纳入满足更具包容性的卡彭特和库斯坦100g口服葡萄糖耐量试验解读标准的待识别女性以及葡萄糖耐量临界异常的孕妇群体时的有效性。
我们重新分析了一项多中心研究的数据，该研究对704名在孕24至28周接受50g口服葡萄糖耐量试验筛查，随后进行常规100g口服葡萄糖耐量试验的孕妇进行了研究。我们使用受试者操作特征曲线分析，根据口服葡萄糖耐量试验解读（美国国家糖尿病数据组或卡彭特 - 库斯坦标准）以及葡萄糖耐量临界异常女性的分类所采用的不同标准，将阳性和阴性组进行汇总。除了由约登指数（敏感度 + 特异度 - 1）定义的统计临界值外，我们还选择了一个“高敏感度”临界值，该临界值由与>70%特异度相关的最大敏感度确定。
采用美国国家糖尿病数据组标准时，当阳性组仅包括口服葡萄糖耐量试验结果为阳性的女性时，统计临界值和高敏感度临界值设定为142mg/dL；当阳性组还包括葡萄糖耐量临界异常的受试者时，临界值设定为140mg/dL。采用卡彭特 - 库斯坦标准时，当阳性组仅包括口服葡萄糖耐量试验结果为阳性的女性时，统计临界值设定为141mg/dL；当阳性组还包括葡萄糖耐量临界异常的受试者时，临界值设定为140mg/dL；当阳性组仅包括口服葡萄糖耐量试验结果为阳性的女性时，高敏感度临界值设定为140mg/dL；当阳性组还包括葡萄糖耐量临界异常的受试者时，高敏感度临界值设定为136mg/dL。
我们建议，如果诊断目标仅为识别口服葡萄糖耐量试验结果为阳性的女性，维持口服葡萄糖耐量试验140mg/dL的阈值。为预防伴有葡萄糖耐量临界异常的妊娠中的围产期风险，采用卡彭特 - 库斯坦标准时，可以假设较低的临界值（136mg/dL）可提高试验敏感度，从而对“临界”受试者进行更广泛的诊断；然而，假阳性率增加导致的更高经济成本以及目前在敏感度方面可获得的有限改善并不足以证明其广泛应用的合理性。
