Rotating Night Shift Work and Healthy Aging After 24 Years of Follow-up in the Nurses' Health Study

Key Points Question Is rotating night shift work prospectively associated with healthy aging that simultaneously considers chronic diseases, cognitive and physical function, and mental health? Findings In a large cohort study of 46 318 female nurses, long-term rotating night shift work was associated with modestly decreased odds of healthy aging after 24 years of follow-up. Meaning In addition to the existing literature suggesting that shift work is associated with increased mortality, the findings of this study further suggest that shift work is also associated with worse overall health among women who survive to older ages.

Abbreviations: HA: healthy aging; a MV-based PS weight: propensity weight was based on: age at baseline (continuous), education (registered nurse, bachelor, or graduate), marital status (married, widowed, or separated/divorced), household income (quintiles); baseline hypertension and high cholesterol (yes or no); family history of cancer, MI and diabetes (yes or no); menopausal status and hormone use (premenopausal, postmenopausal never users, postmenopausal past users, postmenopausal current users), aspirin use (regular use or not); lifestyle factors including smoking history (never, former smoker, current smoker), alcohol intake (none, 1-14.9, ≥15 g/d), total energy intake (quintiles), diet quality (Alternate Healthy Eating Index score, in quintiles), physical activity (metabolic equivalent task-hours per week, in quintiles), standing and sitting time (in quintiles); sleep duration (5, 6, 7, 8, 9h); b MV+BMI-based PS weight: propensity weight was additionally based on: body mass index at baseline (BMI <18.5, 18.5-24.9, 25-29.9, ≥30 kg/m 2 ). b Lifestyle score included five lifestyle factors: diet quality, smoking, physical activity, alcohol consumption, and BMI. We defined a healthy diet as an AHEI score in the top 40% of our cohort. For smoking, we defined the low-risk as never smoking. We classified low-risk of physical activity as more than 30 minutes a day of moderate or vigorous activities (including brisk walking) that require at least the expenditure of 3 metabolic equivalents (METs) or more per hour. We defined low-risk alcohol consumption as moderate alcohol consumption, i.e. 5-15 g/day. Low-risk body weight was defined as BMI in the range of 18.5-24.9 kg/m 2 . For each factor, the participant received a score of 1 if she met the criterion for low-risk and 0 otherwise (high-risk factor). The sum of these five scores provided a final low-risk score of 0, 1, 2, 3, 4 or 5 with higher scores indicating a healthier lifestyle. Adjusted for same confounders as the main analysis except for lifestyle factors, i.e., age at baseline (continuous), education (registered nurse, bachelor, or graduate), marital status (married, widowed, or separated/divorced), household income (quintiles); baseline hypertension and high cholesterol (yes or no); family history of cancer, MI and diabetes (yes or no); menopausal status and hormone use (premenopausal, postmenopausal never users, postmenopausal past users, postmenopausal current users), aspirin use (regular use or not); sleep duration (5, 6, 7, 8, 9h).  Self-reported disease was confirmed by professional staff through medical record or pathology report review, telephone interview, or supplementary questionnaire inquiries, and has been shown valid in this cohort.
Diabetes was confirmed using a validated supplementary questionnaire. Myocardial infarction and stroke were adjudicated using medical records, so were cancer, Parkinson disease, and multiple sclerosis. Coronary artery bypass graft surgery, congestive heart failure, kidney failure, amyotrophic lateral sclerosis, and chronic obstructive pulmonary disease (COPD) are based on self-reports.

eMethods 2. Assessment of Covariates
Information on a broad range of covariates were obtained, including demographic characteristics such as marital status, race, education, and household income; lifestyle factors (physical activity, sleep behavior, and dietary data); family history of cancer, myocardial infarction, and diabetes; clinical diagnoses of hypertension and high cholesterol; use of supplemental multivitamin and aspirin; menopausal status and postmenopausal hormone use.
Median annual household income and home value were estimated from the census tract of participant's residence, geocoded to the 1994 U.S. Census. Weight and height were collected on the 1976 questionnaire, and weight was further requested every two years thereafter. Self-reported weight was highly correlated (r=0.96) with measured weight in a previous validation study. 40 Dietary data were collected using a validated semi-quantitative food frequency questionnaire, 41-43 based on which we derived total energy intake, alcohol and coffee consumption, and the Alternate Healthy Eating Index (AHEI). Women reported the average total hours of sleep in 24 hours and frequency of snoring in 1986. Specifically, we defined regular snorers as those who reported snoring regularly, and non-snorers as those who only occasionally or almost never snored.
Physical activity was assessed based on participants' reports regarding the following recreational activities: walking, jogging (>10 minutes/mile), running (10 minutes/mile), bicycling, tennis/squash/ racquetball, swimming, calisthenics/ aerobics/ aerobic dance/rowing machine, yoga/stretching/toning, and lawn mowing. Based on this information and intensity of each physical activity as expressed by the metabolic equivalent of task (MET), total weekly MET-hours were calculated.
Participants were dichotomized into being physically active and inactive, using a threshold of 7.5 MET-h/wk, which corresponds to the minimum current physical activity recommendations. 44 The reproducibility and validity of this questionnaire in our cohorts has been demonstrated previously. 45,46 eMethods 3. Stratified Analysis (Methods and Results) In the primary outcome analysis, we conducted stratified analysis by age in 1988 (46-55, ≥56 years), BMI (<25, ≥25 kg/m 2 ), physical activity (<7.5, ≥7.5 MET-hours/wk), smoking (never, former smoker, current smoker), alcohol consumption (none, 1-14.9, ≥15 g/d), regular snoring (yes, no) and sleep duration (6, 7, 8). In order to evaluate whether an overall healthy lifestyle can attenuate the adverse effects of rotating night shifts on healthy aging, we also conducted a stratified analysis according to a lifestyle score based on five factors, 51 including a high diet quality score (upper 40%), never smoking, 30 min/d of moderate to vigorous physical activity, moderate alcohol consumption, and BMI of 18.5 to 24.9 kg/m 2 . Statistical interactions were assessed by likelihood ratio tests comparing the models with and without the multiplicative interaction terms.

Stratified analysis (results)
Overall, we did not observe statistically significant differences in the association by several pre-specified factors (P for interaction >0.05; eTable5), except for physical activity. The inverse association of rotating night shift work with healthy aging was statistically significant among physically inactive women and no association was observed for physically active women (P for interaction =0.02). Moreover, the inverse association between rotating night shift work and healthy aging was more evident among older women although P for interaction =0.10: among those aged 46-55 years at baseline (survived to 70-79 years in 2012), the MV-adjusted OR comparing those with 10 years vs. those with no rotating night shift work was 0.84 (95% CI: 0.73 to 0.98); by contrast, among participants aged 56-68 years at baseline (survived to 80-92 years in 2012), the corresponding OR was 0.57 (95% CI: 0.40 to 0.82).