Telomere length has been identified as a potential biomarker of aging 
. Our previous study examined the effect of associations between diet, body composition, and lifestyle factors on telomere length in the NHS 
. In the present study, we examine the associations of both rotating night shifts and sleep duration with telomere length. Our results suggest that sleep duration is positively associated with telomere length, except for those with more than 10-hour sleep. Individuals with long sleep duration (10+ hours) had shortened telomere length due to two possible reasons. Long sleep duration is likely to be an indicator of poor physical condition. Previous research showed that long sleep (≥9 hours) was associated with decreased physical performance in older women compared with mid-range sleepers (7–8 hours) 
. On the other hand, people who sleep more than 10 hours can be very athletic. Our study showed that people with high level of moderate or vigorous physical activity had shorter telomere length than those with modest level (unpublished data).
Short sleep duration has been associated with a decreased risk of Parkinson's and increased risks of various cancers 
. Three cohort studies have suggested that breast cancer risk decreased with increasing sleep duration. 
, but in our own cohort we did not observe an association between sleep duration and breast cancer risk 
. Another study also reported an association between short sleep duration and an increased risk of colorectal adenomas 
. Individuals sleeping fewer than 6 hours per night had an almost 50% increased risk of colorectal adenomas (OR
1.47, P for trend
0.02) compared with individuals sleeping at least 7 hours per night 
. Similarly, with the exception of Parkinson's disease and melanoma, short telomere length in PBL has been suggested to be associated with an increased risk of breast, bladder, lung, and stomach cancers 
. Therefore, we hypothesized that there may be a positive association between sleep duration and telomere length. The current study supports the hypothesis that sleep duration is positively associated with telomere length, but the mechanisms underlying the association are still unknown. Melatonin, a hormone closely related to sleep, has been demonstrated to defend against oxidative stress, promote DNA repair and inhibit tumor development 
. Some studies have shown that short telomere length is associated with high oxidative stress 
. Recently, a link between short sleep duration and low levels of melatonin has been suggested 
. In a cohort study of sleep duration and breast cancer, the levels of melatonin in urine exhibited a dose-dependent positive association with the number of hours of sleep 
. Another study also showed that serum melatonin levels were lower in individuals with<6 hours of sleep than those with >9 hours 
. Therefore, sleep duration probably influences telomere length via a melatonin-mediated pathway. The finding that patients with obstructive sleep apnea syndrome had both an abnormal melatonin secretion pattern 
and short telomere length 
was compatible with this hypothesis. More studies are needed to elucidate the possible mechanism.
We observed that long sleep duration was significantly associated with long telomere length among women younger than 50 years old but not among those over 50 years old. Because telomere is shortened over age, it is possible that telomere attrition is more pronounced and more sensitive to environmental stress in early age and the attrition becomes slower over time. In addition, the effect of sleep duration on telomere length may be mediated through a melatonin-related pathway. High melatonin levels are a recognized protective factor against oxidative stress 
and long sleep duration has been linked to high melatonin levels 
. A significant trend of decreasing melatonin levels with increasing age was observed previously 
. The variation of amplitude (night-day ratio of melatonin) was bigger in younger individuals than in older ones 
. Our previous study also observed that younger women appeared to have higher urinary melatonin levels than older women, particularly in premenopausal women (≤39 yr versus ≥49 yr) 
. Therefore, it would be predicted that the association between sleep duration and telomere length is more apparent among younger women.
Longer years of rotating night shifts work have also been associated with increased risk of various cancers, such as breast, colon, prostate, and endometrium 
. However, in our previous study, we also suggested that both long duration of rotating night shifts work and short telomere length are associated with a decreased risk of melanoma 
and Parkinson's disease 
. We found that women with a longer history of rotating night shifts work tended to have shorter telomere length, but the association was not statistically significant. Consistent with our findings, a cross-sectional analysis of the Sister Study showed no association between work at night (P
0.83), rotating nightshift work (P
0.44) and telomere length 
In conclusion, in this large prospective cohort study, we found that sleep duration was positively associated with telomere length, especially in women under 50 years old. Women with longer duration of rotating night shifts tended to have short telomere length, but it was not statistically significant. To our knowledge, this is the first study to report associations between rotating night shifts, sleep duration, and telomere length in a cohort study. Further research is needed to confirm these possible associations.