Study links shift work to increased risk of diabetes

“Type 2 diabetes is more common in people who work shifts, a large international study suggests,” BBC News reports.

The BBC reports on a review that searched the literature and found 12 studies including more than 225,000 people which looked at the link between shift work and diabetes.

When pooling the results the researchers found that overall, shift work was associated with a 9% increased risk of diabetes. The association was found to be stronger in men (37%) and for those working rotating shifts – such as two weeks on nights, two weeks on days (42%).

However, there are problems with concluding from these studies that there really is a link between shift work and developing diabetes. For example, it is difficult to establish cause and effect, because it’s not completely clear that people hadn’t already got diabetes at the time their shift work pattern was being assessed. It’s further unclear whether the apparent relationship may not just be caused because of other factors that are associated with both shift work and diabetes (such as diet and activity).

Also, none of the 12 studies were conducted in the UK, and half were from Japan. While the results may be applicable here, different cultures may have different a work ethic, environmental and health differences, meaning that they cannot so easily be generalised to all populations.

The identified relationship is undoubtedly worthy of further study, to see whether shift work could have direct biological effects on the body that lead to the development of diabetes.

Where did the story come from?

The study was carried out by researchers from Huazhong University of Science and Technology, Wuhan, Hubei, and Jiangxi Science and Technology Normal University, Nanchang, Jiangxi, both in China.

No sources of financial support are reported and the authors declare no conflicts of interest.

The study was published in the peer-reviewed journal of Occupational and Environmental Medicine. 

The UK media accurately reports the results of this study, and discusses the possible causes and hazards of shift working, such as disruption to the “body clock”, which seem plausible, if unproven. However, they do not highlight the study's limitations.

What kind of research was this?

This was a systematic review and meta-analysis.

Researchers searched the literature from across the world to find observational studies that have examined whether shift work may be associated with the risk of type 2 diabetes. The researchers then pooled the results of these studies.

Type 2 diabetes is a global health problem, and it is estimated that in just over 10 years time the number of cases could have increased by 65% to reach 380 million cases worldwide.

Diabetes is associated with considerable ill health and mortality. So identifying modifiable risk factors that may reduce risk of the disease developing is highly important. Being overweight or obese is the most well established modifiable risk factor for type 2 diabetes.

The researchers say that shift work, with its irregular working hours and rotating schedule has been demonstrated to have some effect on sleep patterns, tiredness, cognitive capacity, and digestion.

Some studies have even linked it with breast cancer and vascular disease. Therefore, this review aimed to look at the possible association with diabetes.

What did the research involve?

The researchers searched several literature bases for studies published up to April 2014 that have examined an association between shift work and diabetes, using relevant search terms, including impaired glucose tolerance and insulin resistance. Any study design or study population was eligible, but only studies in the English language were included. The researchers pooled observational studies that had directly examined the link between shift work as an exposure and diabetes as an outcome.

Shift work schedules were categorised as rotating, irregular and unspecified, night, mixed and evening.

Studies examining non-work-related nighttime activities/light exposure were excluded. Two researchers extracted data and assessed the quality of the studies.

What were the basic results?

Twelve studies met inclusion criteria: eight cohort studies (seven prospective, one retrospective), and four cross sectional studies, published between 1983 and 2013. The 12 studies included a total 226,652 people, with sample size in individual studies ranging between 475 and 107,915. There were a total 14,595 people with diabetes (6% of the total sample). Six studies came from Japan, two from the US, two from Sweden, one from Belgium and one from China. Eight of the studies included only men, two both sexes, and two women only.

The pooled results of all studies found that shift work was associated with a 9% increased risk of diabetes (odds ratio (OR) 1.09, 95% confidence interval (CI) 1.05 to 1.12).

The odds ratios when pooling only cohort studies and only cross sectional studies were fairly similar (slightly higher odds for cohort studies of 12% versus 6% for cross sectional).

They then carried out further sub-analyses to examine whether specific factors were associated. The odds of diabetes were much higher for men (37% increased risk) than for women (9%).

There was a significant association with diabetes for rotating shifts, irregular or unspecific shifts and night shifts; but no link for mixed or evening shifts. The largest association with diabetes was for rotating shifts (42% risk).

Sub-analyses of studies that had controlled for body mass index (BMI) in their models, and of studies that had controlled for physical activity in their models still found similar, significant links (7% increased odds of diabetes).

How did the researchers interpret the results?

The researchers conclude that “shift work is associated with an increased risk of diabetes. The increase was significantly higher among men and the rotating shift group, which warrants further studies”.

Conclusion

This systematic review finds an association between shift work and diabetes, overall the pooled results found that shift work was associated with a 9% risk of diabetes. The review has strengths in that it has reviewed the global literature and identified a reasonable sample of 12 observational studies including more than 225,000 people.

However, there are various important limitations that need to be taken into account before concluding that shift work directly increases the risk of diabetes.

Type 1 or type 2?

The main point under investigation was whether shift work could be a modifiable risk factor for the development of type 2 diabetes – the condition where people either produce less insulin, or their body is less sensitive to the effect of insulin – rather than the autoimmune condition of type 1 diabetes, which is not caused by lifestyle. However, all of the review’s results and reports of the individual studies just discuss “diabetes”. It is likely that most of these studies would have been looking at how shift work was associated with type 2 diabetes, but this is not clear.

Lack of clarity on diagnosis

It is not clear from the review whether these studies had definitely included a population of people who were all free from type 2 diabetes at the start, assessed their shift working pattern, and then looked at whether they developed type 2 diabetes during follow-up.

All we know is that the studies have looked at the association between shift work and diabetes. We don’t know whether all studies excluded diabetes at the start, and then used valid criteria to diagnose diabetes during follow-up.

If people already had diabetes (diagnosed or undiagnosed) at the time their shift working pattern was being assessed, then this doesn’t tell us anything about cause and effect. Given that four of the studies were cross sectional anyway, meaning that they are just snapshots in time, the fact that people have diabetes and are currently shift working doesn’t necessarily mean that shift working has caused the diabetes. 

Confounders

Because all the studies were observational (some cross-sectional), we cannot exclude the possibility that any association between diabetes and shift work is being influenced by confounding.

The review conducted sub-analyses only of studies that controlled for BMI in their models, and those that controlled for physical activity (though none seem to have controlled for both).

But aside from family history, no other mention of confounding factors is given, and it is unclear how well studies may have controlled for other factors.

Various sociodemographic, health and lifestyle factors may be associated with both doing shift work, and with risk of diabetes. This could therefore mean that it is not shift work that directly causes diabetes, but the various factors that are associated with shift work that cause diabetes.

Limited study population

No included studies came from the UK, with half coming from Japan. Though it may well be the case that the results from all of these studies can be applied to the UK, different cultures may have different work ethic, environmental and health differences meaning that they cannot so easily be generalised to all populations.

Also, the majority of studies, eight of the 12, included male only populations, therefore the results may have more applicability to men doing shift work than women.

No information given on occupations

Finally, we do not know whether any association between shift work and diabetes may be influenced by the type of work that people are actually doing (routine or professional). 

The identified relationship is undoubtedly worthy of further study to see whether shift work could have direct biological effects on the body that lead to the development of diabetes. As we are increasingly a 24/7 economy, many people are expected to work unsociable shifts, and the health effects of shift work may become more noticeable.

If there is a link between shift work and diabetes (or other chronic diseases), it is equally possible at this stage that it could still be due to confounding from various sociodemographic, health and lifestyle factors that are associated with both shift work and risk of diabetes.

Overall, it cannot be firmly concluded at this stage whether and how shift work may be associated with diabetes.