Large-scale retrospective analyses of the effect of iron deficiency anemia on hemoglobin A1c concentrations

Unlocking Optimal Glycemic Interpretation: Redefining HbA1c Analysis in Female Patients With Diabetes and Iron-Deficiency Anemia Using Machine Learning Algorithms

OBJECTIVE

In response to the nuanced glycemic challenges faced by women with iron deficiency anemia (IDA) associated with diabetes, this study uses advanced machine learning algorithms to redefine hemoglobin (Hb)A1c measurement values. We aimed to improve the accuracy of glycemic interpretation by recognizing the critical interaction between erythrocytes, iron, and glycemic levels in this specific demographic group.

METHODS

This retrospective observational study included 17,526 adult women with HbA1c levels recorded from 2017 to 2022. Samples were classified as diabetic, prediabetic, or non-diabetic based on HbA1c and fasting blood glucose (FBG) levels for distribution analysis without impacting model training. Support Vector Machines, Linear Regression, Random Forest, and K-Nearest Neighbor algorithms as machine learning (ML) methods were used to predict HbA1c levels. Following the training of the model, HbA1c values were predicted for the IDA samples using the trained model.

RESULTS

According to our results, there has been a 0.1 unit change in HbA1c values, which has resulted in a clinical decision change in some patients.

DISCUSSION

Using ML to analyze HbA1c results in women with IDA may unveil distinctions among patients whose HbA1c values hover near critical medical decision thresholds. This intersection of technology and laboratory science holds promise for enhancing precision in medical decision-making processes.

Michaelis-Menten kinetic modeling of hemoglobin A1c status facilitates personalized glycemic control

INTRODUCTION

Discrepancy between measured HbA1c and HbA1c calculated from plasma glucose is associated with higher risk for diabetic complications. However, quantification of this difference is inaccurate due to the imperfect linear conversion models. We propose to introduce a mathematical formula that correlates with the observational data and supports individualized glycemic control.

METHODS

We analysed 175,437 simultaneous plasma glucose and HbA1c records stored in our laboratory database. Employing the Michaelis-Menten (MM) equation, we compared the calculated HbA1c levels to the measured HbA1c levels. Data from patients with multiple records were used to establish the patients' glycemic status and to assess the predictive power of our MM model.

RESULTS

HbA1c levels calculated with the MM equation closely matched the population's average HbA1c levels. The Michaelis constant (Km) had a negative correlation with HbA1c (r2 = 0.403). Using personalized Km values in the MM equation, 85.1% of HbA1c predictions were within 20% error (ADAG calculation: 78.4%). MM prediction also performed better in predicting pathologic HbA1c levels (0.904 AUC vs. 0.849 AUC for ADAG).

CONCLUSION

MM equation is an improvement over linear models and could be readily employed in routine diabetes management. Km is a reliable and quantifiable marker to characterize variations in glucose tolerance.

Association between liver and chronic kidney disease on hemoglobin A1c concentrations

INTRODUCTION

HbA1c is the gold standard for measuring long-range glycemic control in patients with type-2 diabetes mellitus. Conditions such as CKD or LD can lead to spurious HbA1c test results. There is conflicting literature about the relationship between HbA1c, LD, and CKD.

METHODS

Results for HbA1c concentrations were retrieved from 2015- to 2019. We evaluated over 2,500 test results with LD and 20,000 results with CKD compared to over 21,000 test results without LD, iron deficiency anemia, or CKD. Patients were classified as having LD if they had high ALT and AST concentrations. If they have abnormal serum creatinine and BUN or low eGFR based on age-based reference ranges for CKD. Kruskal-Wallis statistical analyses method was used to test whether the two samples followed the same distribution and significance. Results The median HbA1c concentration was 5.8% (40 mmol/l) among LD classified patients in both males and females vs. 5.4% (36 mmol/l) (P<0.001) for females and 5.6% (38 mmol/l) (P<0.001) for males without LD. A significant difference in median HbA1c concentrations were also observed between CKD samples (female: 5.7% (39 mmol/l), male: 6.0% (42 mmol/l)) and non-CKD samples (female: 5.4% (36 mmol/l), male: 5.6% (38 mmol/l)) (P<0.001). Depending on the population's CKD stage, median concentrations of % HbA1c are increased from stage-1 through stage-4 and fell in Stage-5. Conclusion Patients with high AST and ALT concentrations or CKD can have increased HbA1c concentrations compared to normal patients. When using HbA1c concentrations to monitor diabetes, healthcare professionals should consider LD or CKD status before making any therapeutic decisions.