Comparison of glucose measurement on dried blood spots versus plasma samples in pregnant women with and without anemia

ABSTRACT

ABSTRACT Objective Compare the concordance degree between plasma glucose and glucose measurements on Dried Blood Spots (DBS) during pregnancy. Subjects and methods Glucose measurement was performed in pregnant women after a fast of 8-12 hours. Venous blood was collected with sodium fluoride, the plasma was separated, and glucose measured by the enzymatic oxidase glucose method. Capillary blood samples were collected and analyzed by DBS. For statistics, the paired Student's t test, interclass correlation coefficient (ICC), graphic approach of Altman and Bland, and survival - concordance plot were used. Results 307 pregnant women were evaluated, 88.6% without diabetes and 11.4% with previous diabetes. The glucose ranged from 66 to 190 mg/dL [3.66 to 10.55 mmol/L] in plasma and from 53 to 166 mg/dL [2.94 to 9.21 mmol/L] in DBS. The glucose average values were 88.1 ± 12 mg/dL [4.98 ± 0.67 mmol/L] in plasma and 89.2 ±11,5 mg/dL, [4.95 ± 0.64 mmol/L] in DBS - p-value = 0.084. The ICC value was moderate (0.510), and Pearson's correlation coefficient was r = 0.507 p < 0.001. Altman and Bland's graph showed that difference between the values obtained by both methods is -24.62 to 22.3 mg/dL [-1.37 to 1.24 mmol/L]. Significant fixed bias (-1,16 average difference) and proportional bias (r = 0.056; p = 0.33) were not observed. Anemia was associated with differences between plasma glucose and DBS measurements (p = 0.031). Conclusion Capillary glucose in DBS correlates with plasma glucose; however, the methods do not present good concordance. The presence of anemia worsens this result.