Tuesday April 5 2011
The results need to be confirmed in the wider population
“Working long days could increase heart disease risk,” according to the Daily Mirror, which says that “extra hours can increase the risk of heart attack by 67%”.
This news is based on a study that followed 7,095 British civil servants for over 10 years, examining how their working hours related to their risk of having a heart attack. Over the course of the study, 192 experienced a heart attack, with those working over 11 hours per day being 67% more likely to experience one than people working 7 to 8 hours. When the researchers used working hour data to modify an established model for predicting heart attacks, the predictive accuracy of the process was also improved.
This was a well-conducted study but it was carried out in only a single group of low-risk workers, meaning its findings do not apply to the British population as a whole. Also, it is not clear exactly how long working hours might increase the risk of heart attack, as it could be a result of numerous factors such as stress, associated unhealthy lifestyle choices or even working long hours themselves. This technique has shown some merit, but further research will need to test it in other groups and to explore why long hours might lead to a heart attack.
Where did the story come from?
The study was carried out by researchers from University College London and received funding from a number of sources, including the Medical Research Council, the British Heart Foundation and The Wellcome Trust.
The study was published in the peer-reviewed medical journal Annals of Internal Medicine.
This research was covered accurately by newspapers.
What kind of research was this?
This was a cohort study that followed a group of civil servants who were free from heart disease to see whether working long hours was associated with the development of new heart disease, which for the purposes of this study was defined as non-fatal heart attack or death from heart disease. Beyond this, the study also aimed to see how information on working hours might improve risk models currently used to predict coronary heart disease in a low-risk, employed population.
What did the research involve?
The study followed a cohort of people who were participating in a large research project called the Whitehall II study. This study was established to follow British civil servants to help identify how work environment, health-related behaviours and socioeconomic status are associated with clinical disease.
Working hours were measured by a questionnaire given to participants between 1991 and 1993. At this time, the researchers excluded participants from the cohort who already had coronary heart disease, part-time employees and people for whom they could not collect data on working hours. Daily working hours were classified as:
- 7 to 8 (“ normal working hours”)
- 9 (“1 hour of overtime work per day”)
- 10 (“2 hours of overtime work per day”)
- 11 or more hours (“more than three hours of overtime work”)
The final cohort consisted of 7,095 participants aged between 39 and 62 (2,109 women and 4,986 men). Individual participants were followed until 2002 to 2004. During the follow-up period the researchers recorded the number of cases of non-fatal heart attack and death following heart attack.
At the start of the study (the baseline) the researchers also measured and recorded known risk factors for heart disease, such as age, sex, cholesterol levels, blood pressure and smoking habits. They also asked about whether people were taking blood pressure medication, antiplatelet agents (to prevent blood clots) or lipid-lowering medications (such as statins).
The relative influence of each of these risk factors can be clinically used to calculate a person’s risk using a statistical model called the “Framingham risk score”. The researchers calculated the 10-year risk of coronary heart disease using the standard Framingham risk score, assessed the influence of working hours on risk and finally developed a new model that incorporated this data on working hours.
What were the basic results?
On average, participants in the study were followed up for 12.3 years, during which time 192 of the 7,095 participants had a non-fatal heart attack or died from heart disease. A further 171 died of other causes.
As the researchers had followed individuals for differing lengths of time they calculated the incidence of heart attack per ‘person-years’, a measure calculated by multiplying the number of participants by the length of time each was followed. In total, the study provided 80,411 years of participant data. The researchers then calculated that the rate of heart attacks was 23.9 per 10,000 person-years.
The researchers found that just over half of the people worked a normal 7- to 8-hour day (54%) whereas 10.4% worked 11 hours or more.
The researchers then used working time data to adjusted the risk factors included in the Framingham risk score and calculated that, relative to a person who worked a 7- to 8-hour-day, people who worked over 11 hours had a 67% increased risk of heart attack (hazard ratio HR 1.67, 95% confidence interval, 1.10 to 2.55). They found no difference in risk of heart attack of people who worked 9- or 10-hour days relative to people who worked 7 to 8 hours (HR 0.90, 95% CI, 0.60 to 1.35 and HR 1.45, 95% CI, 0.99 to 2.12, respectively).
The researchers found that adding working hours into the Framingham risk score model improved the sensitivity of the model to identify people who would later go on to develop coronary heart disease. They found that by adding working hours into the model, 4.7% of people were correctly reclassified as being at risk.
How did the researchers interpret the results?
The researchers say that long working hours influence the likelihood that a person will go on to have a heart attack and that adding this risk factor to the Framingham risk score improves the ability of the model to predict risk of heart attack in a low-risk, employed population. They say that “a potential advantage of using working hours as a risk marker is that ascertaining this factor in a clinical interview is simple, quick and almost cost free”.
This research has demonstrated that in a group of employed individuals who did not have heart disease, working long hours (over 11 per day) was associated with an increased risk of subsequent heart attack compared to people who worked normal 7- to 8-hour-days. This research is of importance and may help to improve models for predicting heart attack risk through the addition of a single, simple measure. However, the researchers rightly pointed out several limitations to their study:
- The researchers modelled the risk only in a single population comprising civil servants, and did not validate their results in a second population. However, they say they did perform statistical tests and simulations to test the validity of their model and that these suggest that the level of improvement of the Framingham risk score model that they calculated is not overly optimistic.
- The researchers only measured risk factors and medication usage once, at the start of the study. Therefore, their data did not account for any changes that may have occurred over the several years of follow-up.
- The cohort was comprised of low-risk persons, who were free from heart problems at baseline and did not include people with lower socioeconomic status. Therefore the findings may not be generalisable to higher risk groups in the general population.
- All participants were drawn from the civil service, therefore their work behaviours and environment may not be typical of those seen in other workplaces or professions.
- The research did not look at why long working hours may be associated with higher risk for heart attack and could not establish whether long work hours in itself caused the increased risk or whether it may be caused by unmeasured confounders. For example, long work hours may be associated with stress and disrupted eating, sleeping and exercise opportunities.
Further, the researchers only looked at outcomes of fatal or non-fatal heart attack, and did not look at changes in other disease markers or risk factors (e.g. changes in cholesterol, blood sugar etc), or look at people who had evidence of heart disease but did not develop a heart attack. These things may have helped to see how and why long hours may be contributing to the development of heart disease. More research is needed to assess why there is an association.
Overall, this was a well-conducted study that has highlighted another easily measured risk factor for heart attacks. Further research is now needed to understand why this may be the case, and to validate the model in more diverse populations.