Comparison of the ability to diagnose gestational diabetes mellitus between glycated albumin or fructosamine and hemoglobin A1c-a meta-analysis of diagnostic studies.

BACKGROUND

While the oral glucose tolerance test (OGTT) is the gold standard for diagnosing gestational diabetes mellitus (GDM), hemoglobin A1c (A1C), glycated albumin (GA), and fructosamine (Fruc) have the potential to conveniently classify pregnant women as having or not having GDM because two or more measurements are not necessary, unlike with OGTT. Compared with A1C, GA or Fruc, which reflects more recent glycemic indices, may more accurately detect GDM. This meta-analysis compared the diagnostic ability of GDM between GA or Fruc and A1C.

METHOD

Comprehensive literature searches were performed for studies published up to Sep.11, 2023. Selected studies were limited to those that attempted to identify GDM using both GA and/or Fruc and A1C to compare these indicators. Two authors extracted the data. Disagreements were resolved by a third author. Study quality was assessed by a revised tool for quality assessment of diagnostic accuracy in studies (QUADUS-2). The dataset consisting of true-positives, false-positives, false-negatives, and true-negatives was pooled with a bivariate between-study model and hierarchical summary receiver operating characteristic model. Pooled relative sensitivity and specificity based on a random-effects model were calculated to compare the diagnostic ability of A1C and GA and/or Fruc for GDM.

RESULT

There were 20 eligible studies. Pooled positive and negative likelihood ratios with 95% confidence intervals (CI) were 5.11 (2.30-11.37) and 0.53 (0.40-0.71), respectively, if GA or Fruc was used and 3.75 (2.61-11.39) and 0.53 (0.42-0.66), respectively, if A1C was used. Relative sensitivity and specificity (95% CI) for GA or Fruc compared with A1C was 1.00 (0.94-1.07) and 0.99 (0.97-1.01), respectively.

CONCLUSION

The insufficiently low value of the negative likelihood ratio did not support the single use of GA/Fruc or A1C in screening for GDM suggesting the necessity of combining other risk factors with glycemic indicators. Although there was no statistically significant difference in diagnostic ability between GA or Fruc and A1C, the value of the positive likelihood ratio indicated that GA or Fruc was somewhat more useful than A1C as a rule-in test for the diagnosis of GDM.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42023461975.