BACKGROUND: Measurement of glucose in multiple Field Laboratories requires rigorous standardization when patients and caregivers are masked, unless predefined thresholds are met. Local misclassification of participants at the thresholds can introduce recruitment bias and adversely affect the integrity of study findings. PURPOSE: To describe the challenges and the approach to meeting them in measuring glucose, HbA1c, and C-peptide in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study. HAPO is an observational epidemiologic study of 25 000 pregnant women from 15 centres in 10 countries, designed to clarify unanswered questions on associations of maternal glycemia, less severe than overt diabetes mellitus, with risks of adverse pregnancy outcome. METHODS: Glucose tolerance (75 g two-hour OGTT) is assessed locally at 24-32 weeks' gestation, with results masked if fasting and two-hour plasma glucose are </=5.8 mmol/L and </=11.1 mmol/L, respectively. For analysis of outcomes HAPO utilizes measurements of glucose, C-peptide, and HbA1c on frozen samples in a Central Laboratory; measurement of HbA1c on stored, frozen blood also creates special challenges. A common external quality assessment scheme was used to standardize glucose measurements between Field Centre and Central Laboratories before and during data collection. RESULTS: Agreement between Field Centres and the Central Laboratory in rates above the masking thresholds has been excellent (Kappa = 0.79, P < 0.001), with rates of 1.8 and 1.7% respectively. Percent technical error for Central Laboratory OGTT measurements of glucose, C-peptide, and HbA1c, were 2.0, 4.2, and 2.0, respectively. LIMITATIONS: It is not possible to assess if comparable results could have been obtained with a less rigorous approach to standardization. CONCLUSIONS: HAPO has been successful in ensuring comparability of glucose measurement between its Central Laboratory and Field Centre laboratories and in measuring its key metabolites with accuracy and precision. Recruitment bias based on glucose measurement has been avoided.
BACKGROUND: Measurement of glucose in multiple Field Laboratories requires rigorous standardization when patients and caregivers are masked, unless predefined thresholds are met. Local misclassification of participants at the thresholds can introduce recruitment bias and adversely affect the integrity of study findings. PURPOSE: To describe the challenges and the approach to meeting them in measuring glucose, HbA1c, and C-peptide in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study. HAPO is an observational epidemiologic study of 25 000 pregnant women from 15 centres in 10 countries, designed to clarify unanswered questions on associations of maternal glycemia, less severe than overt diabetes mellitus, with risks of adverse pregnancy outcome. METHODS:Glucose tolerance (75 g two-hour OGTT) is assessed locally at 24-32 weeks' gestation, with results masked if fasting and two-hour plasma glucose are </=5.8 mmol/L and </=11.1 mmol/L, respectively. For analysis of outcomes HAPO utilizes measurements of glucose, C-peptide, and HbA1c on frozen samples in a Central Laboratory; measurement of HbA1c on stored, frozen blood also creates special challenges. A common external quality assessment scheme was used to standardize glucose measurements between Field Centre and Central Laboratories before and during data collection. RESULTS: Agreement between Field Centres and the Central Laboratory in rates above the masking thresholds has been excellent (Kappa = 0.79, P < 0.001), with rates of 1.8 and 1.7% respectively. Percent technical error for Central Laboratory OGTT measurements of glucose, C-peptide, and HbA1c, were 2.0, 4.2, and 2.0, respectively. LIMITATIONS: It is not possible to assess if comparable results could have been obtained with a less rigorous approach to standardization. CONCLUSIONS: HAPO has been successful in ensuring comparability of glucose measurement between its Central Laboratory and Field Centre laboratories and in measuring its key metabolites with accuracy and precision. Recruitment bias based on glucose measurement has been avoided.
Authors: Boyd E Metzger; Steven G Gabbe; Bengt Persson; Thomas A Buchanan; Patrick A Catalano; Peter Damm; Alan R Dyer; Alberto de Leiva; Moshe Hod; John L Kitzmiler; Lynn P Lowe; H David McIntyre; Jeremy J N Oats; Yasue Omori; Maria Ines Schmidt Journal: Diabetes Care Date: 2010-03 Impact factor: 17.152
Authors: C M Ackerman; L P Lowe; H Lee; F Chen; E Hughes; P Cholod; A R Dyer; M G Hayes; B E Metzger; W L Lowe; M Urbanek Journal: J Clin Endocrinol Metab Date: 2010-05-05 Impact factor: 5.958
Authors: E Sidahmed; M L Cornellier; J Ren; L M Askew; Y Li; N Talaat; M S Rapai; M T Ruffin; D K Turgeon; D Brenner; A Sen; Z Djuric Journal: J Hum Nutr Diet Date: 2013-09-20 Impact factor: 3.089
Authors: Lynn P Lowe; Boyd E Metzger; William L Lowe; Alan R Dyer; Thomas W McDade; H David McIntyre Journal: J Clin Endocrinol Metab Date: 2010-09-15 Impact factor: 5.958
Authors: Rachel Kadakia; Octavious Talbot; Alan Kuang; James R Bain; Michael J Muehlbauer; Robert D Stevens; Olga R Ilkayeva; Lynn P Lowe; Boyd E Metzger; Christopher B Newgard; Denise M Scholtens; William L Lowe Journal: J Clin Endocrinol Metab Date: 2019-10-01 Impact factor: 5.958
Authors: Tatjana Radaelli; Kristen Anne Farrell; Larraine Huston-Presley; Saeid Baradaran Amini; John Patrick Kirwan; Harold David McIntyre; Patrick Michael Catalano Journal: Diabetes Care Date: 2009-12-23 Impact factor: 19.112
Authors: Amanda M Perak; Nicola Lancki; Alan Kuang; Darwin R Labarthe; Norrina B Allen; Svati H Shah; Lynn P Lowe; William A Grobman; Denise M Scholtens; Donald M Lloyd-Jones; William L Lowe Journal: Am J Obstet Gynecol Date: 2020-08-05 Impact factor: 8.661