In this longitudinal study of 608 patients with stable coronary artery disease, we observed three distinct telomere trajectories: 45% exhibited telomere shortening, 32% maintained the same telomere length, and 23% lengthened their telomeres during 5 years of follow-up. We also identified four independent predictors of telomere shortening in patients with coronary artery disease: longer baseline telomere length, older age, male sex, and higher waist to hip ratio. These observations suggest bidirectional regulation of leukocyte telomere length, and raise the possibility that telomeric aging may be reversed.
Prior studies have observed telomere lengthening in approximately 12–24% of healthy individuals 
, but the trajectory of telomere length has not previously been evaluated in patients with coronary artery disease. Given the physiological burden of chronic illness, one might expect telomere shortening to occur in a greater proportion of patients with coronary artery disease. However, less than half of our sample experienced telomere shortening, and almost a quarter actually lengthened their telomeres during the 5-year follow-up. Moreover, individuals with the longest telomeres experienced the greatest amount of shortening, while those with shorter telomeres either maintained or increased their length. These results demonstrate for the first time that the inverse association previously seen in healthy individuals 
extends to patients with coronary artery disease. Consistent with mathematical models of telomere shortening 
, these observations further suggest that there may be negative feedback regulation of leukocyte telomere length in humans. An intriguing candidate for mediating such regulation is the enzyme telomerase, which is active in hematopoietic stem and progenitor cells, and at low levels in peripheral blood leukocytes 
Surprisingly, once baseline telomere length was taken into consideration, only a few other candidate variables retained an independent effect on telomere trajectory. Specifically, and in contrast with prior cross-sectional studies 
, we found no significant associations between telomere trajectory and ethnicity, socioeconomic status, blood pressure, insulin resistance, smoking, body mass index, physical activity, or consumption of multivitamins. These findings suggest that the epidemiology of longitudinal telomere dynamics is fundamentally distinct from the epidemiology of cross-sectional telomere length. Although many variables may be associated with cross-sectional telomere length at a single time point, reflecting the cumulative lifetime burdens of genetic and environmental exposures, most of these effects appear to be dwarfed by the apparent negative feedback from baseline telomere length to telomere trajectory. Moreover, confounding by measured or unmeasured factors, may have contributed to associations of lifestyle factors with cross-sectional telomere length in previous studies.
In this study of individuals with established coronary artery disease, increasing age and male sex were associated with accelerated telomere attrition, even after accounting for differences in baseline telomere length. These findings conflict with prior studies of healthy individuals, which have reported no significant independent effect of sex
, and a bimodal effect of increasing age
. Further studies are needed to examine the effects of age and gender on telomere maintenance and telomerase activity in persons with and without cardiovascular disease.
The association of obesity with short telomere length has been observed 
, but the effects of body mass index and waist-to-hip ratio on telomere dynamics have not been previously reported. We found no independent association of body mass index with telomere trajectory, suggesting that the biological effect is driven primarily by abdominal obesity. Furthermore, we observed no attenuation of this association after adjustment for adipokines and systemic inflammatory mediators that are typically associated with abdominal obesity. Indeed, abdominal obesity appeared to have an even stronger effect on telomere shortening after adjusting for these variables.
Oxidative stress may provide a potential link between abdominal obesity and telomere shortening. Epidemiologic studies have demonstrated a robust association between abdominal obesity and oxidative stress
. Excess production of free radicals causes premature cellular senescence and accelerates age-associated tissue damage in animal models of visceral obesity
. Moreover, oxidative stress directly exerts a negative effect on telomere length maintenance, both through inhibition of telomerase activity
and direct erosion of GGG triplets in telomeric DNA
. Further studies are warranted to elucidate the mechanisms whereby abdominal obesity accelerates telomeric aging in vivo.
No cardioprotective medications were found to have an independent effect on telomere trajectory. In particular, we found no association of statin use with change in telomere length. These findings support and extend the prior observation that, while statins attenuate the excess risk of coronary events conferred by short telomeres, their use is not associated with longer telomeres
Among the strengths of the present study is the measurement of a wide range of candidate variables that have been associated with telomeric aging. The longitudinal study design allowed us to identify the determinants of leukocyte telomere trajectory in a large cohort of patients with stable coronary artery disease. However, several limitations should be considered in the interpretation of our results. First, our measurements were restricted to telomere length in circulating leukocytes and do not necessarily reflect telomere trajectory in other cell compartments such as myocardium, endothelium, or the atherosclerotic plaque. Second, the quantitative PCR technique employed in this study measures the mean telomere length across all chromosomes present in the participant's blood sample. However, evidence from rodents suggests that the shortest telomere, rather than the mean telomere length, may be the more important determinant of cell viability and chromosomal stability
. As such, our use of mean telomere length could have resulted in a loss of precision with regard to ascertainment of the shortest telomere length in each cell. Third, we were not able to perform assays of telomerase activity, which might further clarify the mechanisms of telomere lengthening in a subset of patients. Fourth, our evaluation of abdominal obesity was based on waist-to-hip ratio, but more sophisticated imaging techniques such as computed tomography would likely provide more accurate classification in this regard. Fifth, the study sample consisted only of individuals with established coronary artery disease. As such, our findings may not be applicable to healthy individuals, and we are not able to directly estimate the effect of coronary artery disease on the rate of telomere length change.
Finally, as with any repeated measurement of a continuous variable, the possibility of regression to the mean should be considered. This is however extremely unlikely, because baseline telomere length was not used as an inclusion criterion. Indeed, each and every participant with baseline and follow-up telomere measurements was included in the study. Moreover, we measured baseline and follow-up telomere length in duplicate or triplicate for every sample, thereby limiting the effects of random fluctuation due to test conditions. The coefficients of variability achieved approach those of the Southern blot technique, which remains the gold standard for telomere length assessment.
In summary, we report that leukocyte telomere trajectory in a cohort of persons with coronary artery disease is powerfully influenced by baseline telomere length in a pattern suggestive of negative feedback regulation. Age, male sex, and abdominal obesity also independently predicted telomere shortening. Future studies will further elucidate the mechanisms, significance, and reversibility of telomeric aging in cardiovascular disease.