The Impact of Age and Sex on Fasting Plasma Glucose and Glycated Haemoglobin (HbA1c) in the Non-diabetes Population.

Abstract

Introduction: We previously reported sex differences in the distribution of glycated haemoglobin (HbA1c) for men/women aged < 50 years vs older individuals, with implications for delayed diabetes diagnosis. Here, we explored whether this pattern was also seen in matched fasting plasma glucose (FPG) levels.

Methods: We extracted data on same-day, paired HbA1c and FPG levels from clinical biochemistry laboratory databases from Mersey and West Lancashire Teaching Hospitals NHS Trust (n = 10,153) and Cambridge University Hospitals NHS Foundation Trust (n = 10,022) between Jan 2019 and Dec 2023. Only cases with a single, general-practice HbA1c test were utilised to minimise the risk of including non-diagnostic tests and tests from specialist care (e.g. endocrinology, antenatal services; final dataset: n = 17,271). We examined the links of HbA1c and FPG levels to age and sex.

Results: Median HbA1c levels were 1 mmol/mol lower in women aged < 45 years compared to men aged < 45 years but not in those aged ≥ 45 years. This pattern was not seen with FPG, where median levels in women were 0.1-0.2 mmol/L lower across all ages. The HbA1c:FPG ratio was significantly higher in women than men in the 45-54 and ≥ 55 years age groups (p = 0.004, Z-score = 2.9 and p =  < 0.001, Z-score = 8.9, respectively) but not in the < 45 years age group (p = 0.649, Z-score = 0.5). We confirmed our previous finding that median HbA1c levels in women aged ≥ 55 years and 45-55 years were the same as those in men (39 and 37 mmol/mol, respectively) and that for women aged < 45 years, the median HbA1c (34 mmol/mol) was 1 mol/mol lower than for men (35 mmol/mol). This is reflected in the Z-scores, which showed the largest deviation from zero in the < 45 years age group (- 9.1) and the smallest in the older age group (- 2.9).

Conclusion: We showed differences in HbA1c and FPG patterns with age between men and women, with implications for the diabetes diagnostic threshold for HbA1c in pre-menopausal women, the underdiagnosis of type 2 diabetes in younger women, and missed opportunities for intervention. We propose that a suggested change to HbA1c reference ranges in this group warrants serious consideration and detailed evaluation.