Analysis of self-monitoring of blood glucose metrics in gestational diabetes mellitus and their association with infants born large for gestational age: A historical observational cohort study of 879 pregnancies

Introduction Self-monitoring of blood glucose (SMBG) is the standard of care for women with gestational diabetes mellitus (GDM). This study aimed to review SMBG profiles in women with GDM and to examine how glucose metrics derived from SMBG relate to fetal overgrowth and infants born large for gestational age (LGA, >90th percentile). Material and Methods This was a single-center, historical, observational cohort study of 879 GDM pregnancies in Sweden. The diagnosis of GDM was based on a universal 75 g oral glucose tolerance test at gestational week 28 or 12 in high-risk women. The glucose metrics derived from the SMBG profiles were calculated. Treatment targets for glucose were <5.3 mmol/L fasting, and <= 7.8 mmol/L 1-h postprandial. The median (interquartile range) number of glucose measurements in the analysis for each woman was 318 (216-471), including 53 (38-79) fasting glucose measurements. Associations between glucose metrics and LGA were analyzed using binary logistic regression analysis adjusted for maternal age, body mass index, smoking, nulliparity, and European/non-European origin. Receiver operating characteristic (ROC) curves were used to evaluate glucose levels for LGA prediction. Differences in means were tested using analysis of variance. Results The proportion of LGA infants was 14.6%. Higher mean glucose levels and smaller proportion of readings in target (glucose 3.5-7.8 mmol/L) were significantly associated with LGA (odds ratio [95% confidence interval]: 3.06 [2.05-4.57] and 0.94 [0.92-0.96], respectively). The strongest association was found with mean fasting glucose (3.84 [2.55-5.77]). The ability of mean fasting glucose and overall mean glucose to predict LGA infants in the ROC curves was fair, with areas under the curve of 0.738 and 0.697, respectively (p < 0.001). The corresponding discriminating thresholds were 5.3 and 6.1 mmol/L, respectively. Mean glucose levels increased and readings in target decreased with increasing body mass index category and at each step of adding pharmacological treatment, from diet alone to metformin and insulin (p < 0.001). Conclusions Higher mean glucose levels and a smaller proportion of readings within the target range were associated with an increased risk of LGA. Suboptimal glucose control is associated with obesity and the need for pharmacological treatment.