Interrelationship of hemoglobin A1c level lipid profile, uric acid, C-reactive protein levels and age in a large hospital database

Introduction

Hemoglobin A1c (HbA1c) is used to monitor glucose homeostasis and to identify risk for diabetes. As diabetic patients are frequently present with dyslipidaemia, low-grade inflammation and hyperuricemia, we tested whether HbA1c levels can be estimated having the information about lipid profile, uric acid (UA) and C-reactive protein (CRP) levels. We developed formulas to describe the association of these parameters with HbA1c levels.

Methods

Data of 9599 male and 10,817 female patients, measured between 2008 and 2018, were analysed. Patients represented a general hospital patient population with overrepresentation of those with elevated HbA1c over 5.6%. The impact of gender, age, CRP, lipid profile and UA levels on HbA1c % on HbA1c levels was tested with multiple linear regression model. The magnitude of effects of individual factors was used to develop formulas to describe the association between HbA1c and other cardiometabolic parameters. With these formulas we estimated median HbA1c values in each age in both gender and compared them to measured HbA1c levels.

Results

The developed formulas are as follow: HbA1c (estimated) in women = 0.752 + 0.237*log10(HDL/cholesterol) + 0.156*log10 (cholesterol) + 0.077*log10 (triglyceride) + 0.025*log10(CRP) +0.001*log10 (age) −0.026*log10(HDL/LDL) −0.063*log10 (uric acid)-0.075*log10 (LDL)-0.199*log10(HDL); HbA1c (estimated) in men = 1.146 + 0.08*log10 (triglyceride) + 0.046*log10(CRP) + 0.01*log10 (cholesterol) + 0.001*log10 (age) −0.014*log10(HDL)-0.018*log10(HDL/LDL)-0.025*log10(HDL/cholesterol) −0.068*log10 (LDL)-0.159*log10 (uric acid)

Between 20 and 70 years of age, estimated HbA1c matched perfectly to measured HbA1c in.

Conclusion

At population level, HbA1c levels can be estimated almost exactly based on lipid profile, CRP and uric acid levels in female patients between 20 and 70 years.