br Discussion Previous observational epidemiologic studies h
Discussion Previous observational epidemiologic studies have obtained inconsistent results examining the relationship between LTL and cancer risk. Increased LTL due to germline genetic variation was found to be associated with increased risk for nine of 22 primary cancers in a Mendelian randomization study  while shorter measured LTL was associated with increased risks for some, but not all, cancers in two meta-analyses [10,34]. To our knowledge, only three previous RCTs examined exercise-only effects (i.e., no diet intervention) on LTL [, , ]. None showed a statistically significant difference in LTL change between exercise and control groups. Two of the trials [36,37] were six months duration and were small: one in 16 obese, middle-aged women  and the other in 49 sedentary, overweight, 68-year old adults (71% women) . The Nutrition and Exercise in Women (NEW) Trial was the most similar to the ALPHA Trial, comparing LTL change in healthy, postmenopausal, overweight/obese women assigned to 12-months of aerobic exercise (N = 106), diet (N = 105), diet-plus-exercise (N = 108), or no exercise (N = 80) . Another intervention study that PKF118-310 was non-randomized and six months long (N = 50 overweight adults age 67–68 years), showed that LTL increased and decreased almost equally between exercise and control groups , as in our study (Table 4). This latter study  and others [35,39] further showed significant inverse associations between baseline LTL and LTL change, as in our study (r = −0.62). There are several possible reasons we did not observe a statistically significant intervention effect on LTL, the first relating to time. Oscillations in LTL are known to occur  and longer interventions, beyond 12 months, may be necessary to detect exercise effects. The qPCR assay are prone to variability, despite using analytical triplicates. The variability from the assays likely also contributed random error , therefore decreasing the precision of our effect estimates and decreasing statistical power. Another reason relates to our study population. Some propose that exercise may protect LTL by ‘buffering’ attrition caused by psychological stress [41,42] and we did not account for stress in our study. In addition, there is some evidence that body fat precludes a protective effect of exercise on telomeres  and ALPHA Trial participants were generally overweight. Biologically we expected exercise to lower telomere attrition by decreasing chronic inflammation , oxidative stress [44,45], and DNA damage [46,47] and by increasing antioxidant defense mechanisms . However, we observed no significant effect of the ALPHA Trial intervention on indicators of oxidative damage (8-hydroxy-2′-deoxyguanosine, 8-isoprostaglandin F2α) or anti-oxidant enzymes (superoxide dismutase, catalase) relative to controls  supporting our finding of no effect on LTL, although concentrations of the inflammatory biomarker C-reactive protein decreased . Our exploratory analyses consistently showed greater telomere attrition in the exercise group relative to controls, though not statistically significantly. Sjogren et al. also noted a non-statistically significant, inverse trend between LTL change and exercise duration . In ALPHA Trial participants, this tendency was observed specifically for women who were older, obese, or had lower fitness levels or shorter telomeres at baseline (Tables S1–S4). The reasons for these findings are unclear. An inverted U-shaped dose-response curve was reported previously in other populations, indicating shorter telomeres only in men and women with lower and higher (not moderate) activity levels [50,51]. Some propose an adaptive benefit from moderate exercise that induces anti-oxidant response mechanisms over time, but is overwhelmed by high levels of exercise . We did not observe a U-shaped curve in our study (Fig. 2).