The Sensitivity and Specificity of the Glucose Challenge Test in a Universal Two-Step Screening Strategy for Gestational Diabete

Diabetes Care e1

The Sensitivity and Specificity of the Katrien Benhalima,1 Paul Van Crombrugge,2 GlucoseChallengeTestinaUniversalTwo- Carolien Moyson,1 Johan Verhaeghe,3 Step Screening Strategy for Gestational Sofie Vandeginste,4 Hilde Verlaenen,4 Chris Vercammen,5 Toon Maes,5 Diabetes Mellitus Using the 2013 World Els Dufraimont,6 Christophe De Block,7 Yves Jacquemyn,8 Farah Mekahli,9 Health Organization Criteria Katrien De Clippel,10 Annick Van Den https://doi.org/10.2337/dc18-0556 Bruel,11 Anne Loccufier,12 Annouschka Laenen,13 Roland Devlieger,3 and Chantal Mathieu1

The International Association of Diabetes couldlimitthenumberofOGTTs.TheGCT of the GCT was not used to treat patients, and Pregnancy Study Groups (IADPSG) has been used in combination with the GCTthresholdswerenotprespecified. The recommendsauniversalone-stepscreen- 100 g OGTT or the 75 g OGTT with various diagnosis of GDM was based on the ing strategy with the 75 g oral glucose diagnosticcriteria,butdataarelackingon 2013WHOcriteria(1,2).Ofallparticipants, tolerance test (OGTT) for gestational thesensitivityandspecificityoftheGCTin 1,811 (90.3%) received both a GCT and diabetes mellitus (GDM) (1). Since the conjunction with the 2013 WHO criteria OGTT between 24 and 28 weeks of preg- adoption of the IADPSG recommenda- for GDM. nancy.Thereceiveroperatingcharacteristic tion by the World Health Organization We performed a multicentric prospec- curve showed an area under the curve of (WHO), theIADPSG criteriaare commonly tive cohort study, the Belgian Diabetes in 0.77 (95% CI 0.74–0.81) for the GCT. Based referred to as the 2013 WHO criteria (2). Pregnancy Study (BEDIP-N), between 2014 on the 75 g OGTT, GDM prevalence was The IADPSG recommendation remains and2017,enrolling2,006womenbetween 12.5% (n 5 231). By using a universal two- controversialduetothesignificantincrease 6 and 14 weeks of pregnancy (4). Partic- step screening strategy with the commonly in GDM prevalence, increased workload, ipants without prediabetes or diabetes in used GCT thresholds 140 mg/dL (7.8 the need for a fasting test, and the risk early pregnancy (defined by the American mmol/L) and 130 mg/dL (7.2 mmol/L), for increased medicalization of care (3). Diabetes Association criteria) received both GDM prevalence varied from 7.5 to 9.1% Several professional associations there- a GCT and 75 g OGTT between 24 and 28 (Table 1). The GCT threshold of 140 mg/dL fore still recommend a universal two-step weeks of pregnancy. Participants and health (7.8 mmol/L) only had a sensitivity of screening strategy, using a nonfasting 50 g care providers were blinded for the result 59.6%. To achieve sensitivity rates $70%, glucose challenge test (GCT) to determine of the GCT (4). The GCTs were analyzed the threshold of the GCT would need to be whether an OGTT should be performed (3). centrally at the laboratory of the univer- reduced to at least 130 mg/dL (7.2 mmol/L), The GCT is easier to perform and is generally sity hospital of Leuven. Because the GCT and applying lower thresholds to the GCT e-LETTERS bettertoleratedthananOGTT.Inaddition,a has not yet been validated in conjunction would increase sensitivity rates to $77% but two-step screening strategy with a GCT with the 2013 WHO criteria and the result would lead to low specificity rates varying – OBSERVATIONS

1Department of Endocrinology, University Hospital Gasthuisberg, KU Leuven, Leuven, Belgium 2Department of Endocrinology, Onze-Lieve-Vrouwziekenhuis Ziekenhuis Aalst-Asse-Ninove, Aalst, Belgium 3Department of Gynaecology and Obstetrics, University Hospital Gasthuisberg, KU Leuven, Leuven, Belgium 4Department of Obstetrics and Gynecology, Onze-Lieve-Vrouwziekenhuis Ziekenhuis Aalst-Asse-Ninove, Aalst, Belgium 5Department of Endocrinology, Imelda Ziekenhuis, Bonheiden, Belgium 6Department of Obstetrics and Gynecology, Imelda Ziekenhuis, Bonheiden, Belgium 7Department of Endocrinology-Diabetology and Metabolic Diseases, Antwerp University Hospital, Edegem, Belgium 8Department of Gynaecology, Obstetrics and Fertility, Antwerp University Hospital, Edegem, Belgium 9Department of Endocrinology, Kliniek Sint-Jan, Brussels, Belgium 10Department of Obstetrics and Gynecology, Kliniek St-Jan, Brussels, Belgium 11Department of Endocrinology, AZ Sint-Jan, Brugge, Belgium 12Department of Obstetrics and Gynecology, AZ Sint-Jan, Brugge, Belgium 13Leuven Biostatistics and Statistical Bioinformatics Centre, KU Leuven, Leuven, Belgium Corresponding author: Katrien Benhalima, [email protected]. Received 13 March 2018 and accepted 27 March 2018. © 2018 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for proﬁt, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license. See accompanying articles, pp. XXXX, XXXX, XXXX, XXXX, XXXX, XXXX, XXXX, XXXX, and XXXX. Diabetes Care Publish Ahead of Print, published online May 10, 2018 e2 2013 World Health Organization GDM Criteria Diabetes Care

from 64.2 to 56.0% for a GCT threshold of site and Sylva Van Imschoot from the AZ Sint-Jan 125 mg/dL (6.9 mmol/L) and 120 mg/dL site for their help with the recruitment and study (6.7 mmol/L), respectively (Table 1). For a assessments. The authors thank the research assistants, paramedics, and physicians of all the GCTthresholdof130mg/dL(7.2mmol/L), participating centers for their support and all the 8.2) 7.4) 6.6) 6.0) 5.5) the positive posttest probability was women who participated in the study. – – – – – 26.4%, the negative posttest probability Funding. This investigator-initiated study was was 5.5%, and 65.1% of all OGTTs could funded by the Belgian National Lottery, the Fund % (95% CI) of Academic Studies of UZ Leuven, and the Fund be avoided compared with a universal

: negative likelihood ratio. Yvonne and Jacques François-de Meurs of the

2 one-stepscreeningstrategywiththe75g King Baudouin Foundation. K.B. is the recipient

Negative posttest probability, OGTT (Table 1). ofaClinicalDoctoralScholarshipoftheFundof TheBEDIP-Nstudyis,toourknowledge, Academic Studies of UZ Leuven. R.D. is the the first study that has prospectively recipient of a Fundamenteel Klinisch Navorser- fi schapFondsWetenschappelijkOnderzoekVlaan- evaluated the sensitivity and speci city deren. 37.5)34.3)31.2)28.7)25.6) 6.9 (5.7 6.2 (5.1 5.5 (5.6 4.9 (4.0 4.5 (3.7 – – – – – of the GCT in a universal two-step screen- The sponsors of the study had no role in the ingstrategyforGDMusingthe2013WHO design of the study; in the collection, handling, analysis, or interpretation of the data; or in the % (95% CI) criteria. The GCT has been used in combination with the 100 g OGTT or the 75 g decision to write and submit the manuscript for : positive likelihood ratio; LR publication. 1 OGTTwithvariousdiagnosticcriteriasuch Duality of Interest. No potential conflicts of Positive posttest probability, as the Carpenter and Coustan criteria, the interest relevant to this article were reported. National Diabetes Data Group criteria, the Author Contributions. K.B., P.V.C., and C.M. 0.58)0.53)0.49) 31.7 (26.1 0.45) 29.1 (23.9 0.43) 26.4 (21.7 24.3 (20.0 21.6 (17.8 – – – – – 1999 WHO criteria, or the Canadian Di- conceived the project. K.B. performed the liter- aturereview.K.B.andC.M.wrotethefirstdraftof (95% CI) abetesAssociationcriteriaandhasshown the manuscript. C.M. prepared the data. A.L. 2 variable sensitivity rates between 70 performed the statistical analysis. All authors and 88% and specificity rates between contributed to the study design, including data 69and89%whenusingaGCTthresholdof collection, data interpretation, and manuscript number without GDM. LR 3.6) 0.50 (0.42 3.1) 0.44 (0.37 2.7) 0.39 (0.32 2.4) 0.35 (0.27 2.0) 0.32 (0.24

– – – – – revision. K.B. is the guarantor of this work and, as 5 140 mg/dL (7.8 mmol/L) and sensitivity (95% CI) LR such,hadfullaccesstoallthedatainthestudyand N fi 1 rates between 88 and 99% and speci city takes responsibility for the integrity of the data LR rates between 66 and 77% when using a and the accuracy of the data analysis. cut-off; N GCT thresholdof130 mg/dL (7.2 mmol/L) / , n (5). However, many studies had a high or unclear bias because the result of the References screening test was used to determine 1. Metzger BE, Gabbe SG, Persson B, et al.; International Association of Diabetes and Preg- whether further testing was needed for nancy Study Groups Consensus Panel. Interna- 82.9), 1,282/1,583 3.1 (2.7 78.1), 1,204/1,583 2.8 (2.4 72.4), 1,111/1,583 2.4 (2.2 66.5), 1,016/1,583 2.2 (2.0 58.4), 886/1,583 1.9 (1.7 – – – – – GDM and not all patients received a con- tional Association of Diabetes and Pregnancy city, % (95% CI), number with GCT fi firmatory OGTT if the GCT was below a StudyGroupsrecommendationson thediagnosis 5 certain threshold (5). Our study avoided and classification of hyperglycemia in pregnancy. n Speci – these limitations as both health care pro- Diabetes Care 2010;33:676 682 2. WorldHealthOrganization.Diagnosticcriteria city: N / fi viders and participants were blinded for the n and classification of hyperglycaemia first de- GCT, thus avoiding any bias in screening. tectedinpregnancy:aWorldHealthOrganization In conclusion, we show now that the guideline. Diabetes Res Clin Pract 2014;103:341– GCT has a moderate diagnostic accuracy 363 66.1), 136/228 81.0 (79.0 72.3), 151/228 76.1 (73.9 78.1), 165/228 70.2 (67.9 82.9), 177/228 64.2 (61.8 86.8), 187/228 56.0 (53.5 – – – – – in a universal two-step screening strat- 3. Benhalima K, Mathieu C, Damm P, et al. A proposalfortheuseofuniformdiagnosticcriteria egy for GDM using the 2013 WHO cri-

0.05551. for gestational diabetes in Europe: an opinion paper teria. A GCT threshold of 140 mg/dL 3 by the European Board & College of Obstetrics and (7.8mmol/L)hadonlyasensitivityof59.6% Gynaecology (EBCOG). Diabetologia 2015;58:1422– number with GDM. Speci 1429

5 and can therefore not be recommended ) Sensitivity, % (95% CI),

n 4. Benhalima K, Van Crombrugge P, Verhaeghe J, N in a two-step approach for GDM using the et al. The Belgian Diabetes in Pregnancy Study 2013 WHO criteria. To achieve sensitivity (BEDIP-N), a multi-centric prospective cohort study

cut-off; rates $70%, the threshold of the GCT on screening for diabetes in pregnancy and gesta- city of the GCT across different thresholds GDM, % ( $ Prevalence of would need to be reduced to at least tional diabetes: methodology and design. BMC 130 mg/dL (7.2 mmol/L). Pregnancy Childbirth 2014;14:226 5. Donovan L, Hartling L, Muise M, Guthrie A, Vandermeer B, Dryden DM. Screening tests for )

n gestational diabetes: a systematic review for the

%( Acknowledgments. The authors thank Inge U.S. Preventive Services Task Force. Ann Intern – number with GCT BecksteddefromtheAntwerpUniversityHospital Med 2013;159:115 122 Abnormal GCTs, 5 n — Sensitivity andﬁ speci ivity: 140 mg/dL135 mg/dL 23.9 (441)130 mg/dL 29.1 (537)125 mg/dL 34.9 (645)120 mg/dL 7.5 40.8 (136) (754) 8.3 48.4 (151) (895) 59.6 9.1 (53.0 (165) 66.2 9.8 (59.7 (177) 10.3 72.4 (187) (66.1 77.6 (71.7 82.0 (76.4 Table 1 Threshold GCT $ $ $ $ $ Sensit Conversion factor for SI units for glucose in mmol/L: