Abstract
Long working hours, defined as working >48 hours per week, continues to be a reality for many worldwide. Based on the most recent data published by the International Labor Organization (1), 43.9% of the workforce worldwide and 23.7% of the workforce in the Americas fit into this category. This is especially true for those in health professions. Despite an ongoing concern about potential adverse effects of long working hours on overall health, clear evidence of which health aspects are affected has been difficult to ascertain. Although there are many studies examining the physiological and psychological effects of long working hours, the study by Lee and colleagues is the first study to examine the association between thyroid disease and long working hours (2).
Many of the studies on the adverse effects of long working hours have originated in Japan, where it was estimated that 22% of the workforce worked >48 hours per week in 2007 (3). The data are mixed, with frequent conflicting findings. Several meta-analyses have attempted to clarify the risks associated with long working hours. Bannai and Tamakoshi (4) showed no effect of long working hours on all-cause mortality, but a likely increased risk for the development of heart disease, depression, anxiety, and disordered sleep. Wong et al. (5) showed an increased risk of heart disease, metabolic syndrome, sleep disorders, fatigue, and injuries, and Kivimaki et al. (6) showed an increased risk for stroke. Diabetes was reported to be increased in shift workers (3) and in low-socioeconomic status workers (7). At this point, the only clear consensus is that long working hours lead to stress, disordered sleep, and fatigue.
This study is a cross-sectional study that identified 2059 adults from the Korea National Health and Nutrition Examination Survey (KNHANES) conducted from 2013 to 2015 who were currently working full time, did not have established thyroid disease, and who had a full set of thyroid function tests. The authors were able to divide the group based on their work history and the hours that they worked on a weekly basis. They identified 45 individuals who had an increased thyrotropin (TSH) and 56 who had a decreased TSH. Importantly, only 1 individual with an increased TSH had a low free thyroxine (fT4) and only 4 who had a low TSH also had an increased fT4, so this is essentially a group of individuals with subclinical thyroid disease. Overall, the average number of hours that the individuals with an increased TSH worked was significantly higher than the euthyroid group, with 56% of these individuals in the highest tertile in hours worked and 76% in the top two tertiles. The unadjusted odds ratio for having an increased TSH was 1.38 (95% confidence interval [CI] 1.11–1.73) for every 10 hours increase in weekly work hours over the lowest group (36–42 hours). This increased odds ratio was persistent when stratified according to sex, occupation, income level, and education. When limited to a TSH >7 mIU/L, the odds ratio increased to 1.42 (95% CI 1.08–1.86). While 46% of the individuals with a low TSH also were in the highest tertile and 78% were in the top two tertiles, the average number of hours worked did not achieve statistical significance. Based on these data, the authors suggest that long working hours are associated with hypothyroidism.
The authors do not suggest a causality in their findings. However, whether long working hours are a risk factor for developing intrinsic thyroid dysfunction is the important question raised by this study. A strength of this study is that patients with pre-existing autoimmune thyroid disease as evidenced by positive thyroid peroxidase antibodies were excluded, so the results do not simply reflect the natural history of autoimmune thyroid disease. Since only 1 patient with an increased TSH and 4 with a decreased TSH had abnormal T4 levels, the vast majority of patients had subclinical thyroid disease. Thus, this may be a phenomenon limited to abnormal TSH levels irrespective of overall thyroid function. As is well known, TSH levels are affected by a wide range of nonthyroidal factors (8).
As already noted, it is clear from the literature that long work hours are associated with increased stress, disordered sleep, and fatigue (4,5). All of these factors may play an inhibitory role in TSH secretion and lead to decrease in TSH levels (8). Thus, a decreased TSH would have been expected to be the most common abnormal thyroid parameter due to the stress of working long hours. Indeed, more workers in the study had a low TSH (56) than an increased TSH (45), with an odds ratio of ∼1.2 risk for every 10 hours increase in work week. However, the overall trend was not significant and is not pursued by the authors.
In contrast to a low TSH, nonthyroidal factors play a much more limited role in increasing TSH levels. Outside of the recovery phase of nonthyroidal illness, an increased TSH is almost always a result of the action of drugs decreasing secretion of thyroid hormone from the gland (8). Thus, it is difficult to attribute a role for stress, disordered sleep, and fatigue in the development of the increased TSH observed in this study.
Hypothyroidism has been associated with obesity, metabolic syndrome, and decreased physical activity, but more as a cause of these physiological and metabolic parameters and not as a result. Indeed, none of these parameters appear to be playing a role as a cause of the subclinical hypothyroidism identified in this study. The body mass index in all groups in this study was equivalent and within the normal range (24–25), and the odds ratio for hypothyroidism in manual laborers was similar to that of office workers. Thus, the etiology of the abnormal TSH levels remains unclear.
As the authors note, this is a cross-sectional observational study, so a causal factor cannot be determined. Also, the thyroid function tests were obtained at one point in time. Other than reporting perceived stress, there is no data on symptoms of clinical thyroid dysfunction, so there is no way of determining any adverse effects of the abnormal TSH levels. While the authors suggest that long work hours are more likely to be causal of the increased TSH levels, I would pose that the opposite may be true, as it may take workers with an increased TSH longer to complete their work. However, neither of these causal factors can be determined with the current data.
As the authors suggest, if there is a causal association between long working hours and subclinical hypothyroidism, it would be an important public health issue. Increasing hypothyroidism in the workforce would likely adversely impact worker productivity, but this study is long away from these conclusions. Screening for hypothyroidism is controversial even in targeted populations, such as during pregnancy (9). While the American Thyroid Association recommends screening for thyroid disease in all adults beginning at age 35 years with follow-up testing every 5 years, the U.S. Preventive Services Task Force states that there is insufficient evidence for or against screening. Failing additional convincing data, there is no indication for screening the general workforce for thyroid dysfunction or even targeting those individuals working >48 hours per week.
Similar to screening, treatment of subclinical hypothyroidism is also controversial (10). Even in clinical situations wherein there is general agreement in treating, such as during pregnancy, evidenced-based beneficial effects have been difficult to demonstrate. Thus, treating mildly increased TSH levels in individuals working long hours without any other clinical indication would not be indicated.
In summary, this is an interesting and provocative study that does warrant further investigations. Possible areas of investigation include prospectively evaluating thyroid function in targeted groups at risk for long working hours, such as medical students, physicians, managers, and CEOs.