Purpose: This study aimed to assess for disparities between capillary blood glucose and plasma glucose for the O’Sullivan 1-hour glucose tests. Methodology: Subjects included were English-speaking, non-incarcerated, adult (>18 years old) pregnant patients who could give informed consent. Patients were approached at their outpatient obstetric office visits. After receiving written and verbal consent, patients were given a 50-g glucose load at the conclusion of their clinic visit. Strictly timed venous and capillary blood draws were performed at the outpatient labs. The venous and capillary blood samples were drawn within less than 5 minutes of each other and analyzed using select chemistry analyzers within less than one hour. Quality control measures, including recalibration, testing, and various quality control mechanisms, were reassessed every 24 hours. Data was tested for normal distribution with the Shapiro-Wilk test and Kolmogorov-Smirnov tests. Paired T-tests were used to compare the average difference between measurements in subjects on GraphPad. Results: One hundred patients had enrolled in the study, with ten dropping out soon after being consented. Patients were lost due to nausea and vomiting with the 50-gram glucose load (4 patients) and scheduling conflicts (6 patients). Demographics identified 84% as African American (76/90), average body mass index (BMI) being 35.4 kg/m2, and a mean age of 26. All data of the differences (capillary minus venous glucose) were found to be normally distributed. The capillary sample was significantly different from the venous sample, with the former being, on average 6.156 mg/dL higher than the latter (P 0.0008, CI 2.65 to 9.66). Subgroup analysis suggested significant differences in values among the multiparous patients (P 0.0014, CI -13.18 to -3.349), with capillary glucose being on average 8.26mg/dL higher. Similarly, patients with BMI < 30 mg/kg had significantly different sample values (P 0.0031, CI -18.75 to -4.25), with capillary glucose being 11.50 mg/dL higher on average. Patients with a gestational age of 24 weeks and beyond also had significantly different samples based on collection site (P 0.0023, CI 2.7 to 11.75), with capillary being on average 7.226 mg/dL higher. Patients with BMI 30.0 to 39.9 had no significant difference between capillary and venous glucose (P 0.2393, CI -8.72 to 2.25), with capillary being 3.24 mg/dL higher on average. Patients with BMI 40 or higher also showed no significant difference between venous and capillary glucose (P 0.1053, CI -10.11 to 1.030), with capillary glucose being on average 4.52 mg/dL higher. Nineteen (21%) patients failed their glucose test based on one or both samples. In 12 patients (13%), the two samples were in agreement for test failure. However, seven patients (7.8%) had discordant results where the glucose was less than 140 mg/dL for one sample and higher for the other. Furthermore, 4 of these patients passed by the venous sample and failed with the capillary, and the reverse in sample source occurred in 3 patients. None of the patients with discordant results failed their 3-hour confirmatory test. Discussion: All pregnant patients undergo gestational diabetes mellitus (GDM) screening, and those with diabetes risk factors such as obesity and a history of GDM undergo early diabetes screening. The two-step approach commonly used in the United States is a one-hour 50-gram oral glucose tolerance test (OGTT) using a venous plasma glucose sample. If a patient fails (glucose > 140 mg/dL), she undergoes a 3-hour OGTT. Other accepted cutoffs for the one-hour OGTT are 135 mg/dL and 130 mg/dL. Venous sampling has some factors which lead to inaccuracy, including ongoing glycolysis in samples and decreased concentration compared to capillary/arterial samples in postprandial patients. Capillary sampling can be inaccurate in patients with microvascular disease such as shock, COPD, and Raynaud’s disease. Our institution uses point-of-care testing (POCT) on capillary (finger-stick) using a glucometer with a 140 mg/dL cutoff. Studies are mixed in their endorsement of POCT for the diagnosis/screening of DM. This study suggests that capillary testing is more sensitive to hyperglycemia in postprandial patients and is more sensitive for non-obese, multiparous, and patients undergoing screening for gestational diabetes at 24 weeks GA and beyond. This study also suggests that if POCT/capillary testing is considered, the higher (140) cutoff may be the best. The findings do not show that there is a significant risk of missing diagnoses of GDM, but a higher incidence of needing 3-hour glucose tolerance testing due to its higher sensitivity.
Read full abstract