High serum glucose levels are associated with a higher perceived age

Abstract

Estimating perceived age by facial photographs is a good estimate of health in elderly populations. Previously, we showed that familial longevity is marked by a more beneficial glucose metabolism already at middle age. As glucose is also related to skin aging, this study aimed to investigate the association between glucose metabolism and perceived age. Perceived age was assessed using facial photographs and non-fasted glucose and insulin were measured in 602 subjects from the Leiden Longevity Study. Non-diabetic subjects (n = 569) were divided in three strata according to their glucose levels, and diabetic subjects (n = 33; as a proxy of long-term hyperglycemic exposure) were included as a fourth stratum. Considered confounding factors were gender, chronological age, current smoking, body mass index, photo-damage score, and insulin levels. Perceived age was increased from 59.6 years (SE = 0.3) in the first stratum to 61.2 years (SE = 0.6) in diabetic subjects (p for trend = 0.002). In non-diabetic subjects only, perceived age was increased from 59.6 years (SE = 0.3) in the first stratum to 60.6 years (SE = 0.3) in the third stratum (p for trend = 0.009). Continuously, perceived age increased 0.40 years (SE = 0.14, p = 0.006) per 1 mmol/L increase in glucose level in non-diabetic subjects. The present study demonstrates that, also among non-diabetic subjects, higher glucose levels are associated with a higher perceived age. Future research should be focused on elucidating possible mechanisms linking glucose levels to perceived age.

Fig. 1

Association of glucose levels with perceived age. The figure depicts the association between glucose level and perceived age (n = 602). Non-diabetic subjects (n = 569) were divided over tertiles of glucose and diabetic subjects (n = 33) included as a fourth category. Mean perceived age and obtained p values were adjusted for gender, chronological age, body mass index, current smoking, photo-damage score, and Ln (insulin) level as described in “Methods.” For the non-diabetic subjects, the mean perceived age (SE) was 59.6 years (0.3) in the low, 60.2 years (0.3) in the medium, and 60.6 years (0.3) in the high glucose tertile. For the diabetic subjects, the mean perceived age (SE) was 61.2 years (0.6). In a model using only the non-diabetic subjects, we observed a statistical significant positive trend (p _trend = 0.009) for the increase in perceived age dependent on glucose tertiles. Inclusion of the diabetic subjects in the model also revealed a highly significant positive trend (p _trend = 0.002)