In this prospective longitudinal study of 152 adolescents and young adults with T1DM, both PFT and rate of increase in PFT (mm/year) were associated with the development of MV complications during 10 years of follow-up.
Despite a significant improvement in glycemic control from puberty to early adulthood, retinopathy prevalence increased more than twofold (20% to 51%, p
< .001) throughout the study period. These findings support metabolic memory for retinopathy, as observed in the Diabetes Control and Complications Trial (DCCT),18
and underscore the importance and difficulties in maintaining tight glycemic control during early puberty and the period of growth and development.19
There was a nonsignificant increase in prevalence of renal dysfunction between baseline and final visits in keeping with the apparent disparity observed between prevalence of retinopathy and nephropathy. However, renal anatomical changes can occur in the absence of renal functional abnormalities in patients with T1DM.20
Longitudinal GEE analysis demonstrated that retinopathy was associated with greater PFT, higher mean HbA1c, female gender, longer diabetes duration, and greater BMI SDS. Early renal dysfunction (AER >7.5 µg/min) was associated with greater PFT, higher mean HbA1c, longer diabetes duration, lower BMI SDS, and DBP SDS.
In longitudinal analysis, mean HbA1c, a surrogate marker of chronic glycemic exposure, was positively associated with retinopathy, whereas HbA1c at each individual visit had an inverse association. Although the inverse association appears a paradox, we speculate this may be explained by the significant improvement in glycemic control from baseline to final visits. Further, mean HbA1c, but not “visit HbA1c,” was a significant predictor of incident renal dysfunction in both univariate and multivariate GEE models.
We also analyzed HbA1c SD as a surrogate marker of glycemic variability (results not shown), and unlike other studies,6
we found no association between HbA1c SD and complications. Although this is likely due to the small numbers in the present study, it suggests that PFT may better reflect the additive effect of early glycemic control and glycoxidative stress from glycemic variability.
Changes in plantar fascia collagen are likely to reflect widespread systemic glycation and AGE formation, which may parallel changes in the vasculature. The compounding effects of chronic hyperglycemia,21
and oxidative stress associated with T1DM are not comprehensively quantified by single assay measures such as HbA1c. Abnormal glycation of proteins, lipids, and other moieties, resulting in delayed clearance and exaggerated accumulation of AGEs, may be better assessed through tissue biopsy measures, as used in the DCCT study.4
The impractical nature of these, however, emphasizes the role of PFT as a useful, noninvasive measure of chronic glycation and metabolic burden, especially in those with early vascular complications.
Our longitudinal PFT findings are in agreement with skin biopsy AGE levels in the DCCT and their predictive value for MV complications.4
In adults with type 2 diabetes, skin autofluorescence predicted nephropathy but did not predict retinopathy.22
However, skin auto-fluorescence only detects fluorescent products, whereas PFT is a physical measure to which both fluorescent and nonfluorescent AGEs may contribute. A distinct advantage of PFT is that it uses readily available, easily reproducible, and noninvasive technology in the clinical setting.
The significant reduction in HbA1c with a parallel increase rather than decrease in PFT suggests that early glycation and formation of AGEs may irreversibly change collagen structure, reflecting the differences in half-lives of hemoglobin and skin collagen. Furthermore, these findings are consistent with “the shape of the metabolic memory of HbA1c” proposed by Lind and colleagues,18
who demonstrated that HbA1c values from 5–10 years earlier contribute the greatest risk to progression of presentretinopathy rather than concurrent visit HbA1c. The enduring effects of glycation on collagen, both in tendons and the microvasculature, have been demonstrated in animal models.2,23
The role of the AGE receptor (RAGE), through the activation of nuclear factor-kB in oxidative stress and its association with diabetes complications, has been explored.24
Daily glycemic variability may play a significant role in oxidative stress, both acutely and longer-term, which may further increase AGE production.25
The production of AGEs, activation of RAGE, and irreversible cellular changes appear to perpetuate the cycle of oxidative stress and further production of AGEs independent of HbA1c.26,27
We speculate that glycemic variability, along with metabolic memory (which may be AGE- or epigenetic-mediated, or both), may be an explanation for the role of PFT as an independent predictor of complications.
Our data support the cutoff value of AER >7.5 mg/min as a measure of early renal dysfunction and an early risk marker for subsequent development of MA.13–15
Young people with early renal dysfunction (AER >7.5 mg/min) at baseline were more likely to progress to an episode of MA (40% vs 5%, p
< .001) throughout the study period. The cumulative prevalence of MA throughout the study period was 10% and 17% at 10 and 15 years diabetes duration, respectively. This contrasts with the findings from the population-based Oxford Regional Prospective Study (ORPS), which reported a cumulative prevalence of 26% and 51% of MA after 10 and 19 years duration, respectively.28
However, participants from the ORPS had higher HbA1c and longer diabetes duration than our cohort. Our cohort was clinic-based, and participants had lower HbA1c than nonparticipants, thus selection bias may also contribute to the differences between the two studies. On the other hand, this selection bias towards patients with better glycemic control supports the robustness of PFT as a potential predictor of MV complications, since an association was found even in those with presumably lower levels of tissue glycation. Ideally, direct tissue evidence of glycation and AGE levels from plantar fascia biopsies would be available for correlation with PFT, however, this was not feasible.
The strengths of this study include the prospective longitudinal study design, repeated PFT measures, and the use of all available data points using GEE modeling. Other studies of PFT were cross-sectional in design or used a single baseline PFT measure as a predictor of MV complications; here we present repeated measures in the same individual over time. Another strength of this study is the use of the same laboratory throughout the study to quantify HbA1c, total cholesterol, and AER. Study weaknesses include the loss to follow-up, which usually related to discharge from pediatric clinical care. Study results may be influenced by the changes in ultrasound equipment and operator for the final assessment; however, these were validated and had a very reliable interobserver correlation coefficient.
In conclusion, PFT is longitudinally associated with the development of early MV complications in young people with T1DM. These findings support the concept of metabolic memory. Studies correlating PFT with other noninvasive measures of tissue glycation, such as skin autofluorescence and vascular function, will provide further insight into the pathophysiology and clinical relevance of tissue glycation in diabetes complications.