Daylight Savings and Heart Attacks

With the recent passing in the Senate of the Sunshine Protection Act, which seeks to move the US to permanent Daylight Savings Time (DST), there’s been wider discussion on the benefits and downsides of changing our clocks twice a year. Marco Rubio, the main sponsor of the bill, provides a list of potential benefits in a one pager on his website, making the case that permanent DST will reduce car crashes, reduce health risk, cut robberies by 27% (!), stimulate economic activity, reduce childhood obesity, benefit farmers, and reduce energy usage. There’s plenty of discussion to be had on each of these points, and this post will focus specifically on the relation between clock changes and health risk (specifically cardiac arrest, as this point has received more media coverage) – but first we need to address Rubio’s claim of a 27% robbery reduction from permanent DST. To support this point, he references a 2015 study by the Brookings Institute – but in the second sentence of the paper abstract the reduction is called out as 7%, not 27%. It seems Rubio (or more likely, his staffers) instead looked at the reduction in crime only during the single evening hour which gained extra sunlight, where the observed reduction was indeed 27%. It’s unclear whether this fairly egregious mischaracterization was intentional, but with the large number of grammatical mistakes in the one pager (a press release for a US Senator!) I’m inclined to give Rubio the benefit of the doubt (though it’s tempting to dispel once and for all this fiction that Senator Rubio doesn’t know what he is doing…). While this post will focus on the health risk point, it’s clear that further verification is required for all the benefits listed.

Reading through the media coverage of the health risks of clock changes, it’s easy to get the idea that it’s a massive problem (particularly the loss of an hour in the Spring). Business Insider states that “Daylight-savings time is literally killing us”, the New York Post asks “Does daylight savings time cause ‘biological clock shock’ to your heart?” (with the article answering strongly in the affirmative), and Diagnostic and Interventional Cardiology shares that “Springing forward may have drawbacks for heart health”. The basic idea all these articles (and others) present is that that sleep is important for good health, and clock changes impact sleep. All these articles cite studies to support the point, with significant overlap – in fact, all three (and many others) primarily look at the same Michigan study on heart attack rates, which, as they all call out, found a 24% increase in the rate of heart attacks the Monday following the switch to DST. This fact, by itself, is arguably sufficient to justify the headlines, but the counterpoint either omitted or glossed over by the articles is that the study found a 21% reduction in heart attacks the Tuesday after the switch to Standard Time in the Fall (where we gain an hour of sleep). Additionally, no significant difference in heart attacks was observed when evaluating impact at the weekly level – essentially, only the distribution within the week was changed. 

These findings point to a negligible overall impact of clock changes on heart attack risk – as stated in the study (emphasis mine): 

Our data argue that DST could potentially accelerate events that were likely to occur in particularly vulnerable patients and does not impact overall incidence. There is considerable controversy over the health and economic benefits of DST, and some authorities have argued that this practice should be abandoned. Although we are unable to comment on the merits of these arguments, our data suggest that while such a move might change the temporal fluctuations in AMI [acute myocardial infarction], it is unlikely to impact the total number of MIs in the broader population.

Sandhu A, Seth M, Gurm HS. Daylight savings time and myocardial infarction. Open Heart 2014;1:e000019. doi:10.1136/openhrt-2013-000019

It seems this is a case of media selectively filtering the facts; the study itself certainly doesn’t support the provocative headlines. However, the study does still leave some open questions – for example, why does the Spring time change lead to more heart attacks on Monday, but the Fall time change leads to fewer on Tuesday? As the study was limited to a subset of hospitals in Michigan, the aberrations may simply be due to limited data – the median number of heart attacks was only 31 per day, and only a few years were included, meaning natural variance could affect the findings, particularly when analyzing the impact at the daily level. 

To get around these limitations, ideally we’d like to look across multiple studies, and conveniently a meta-analysis was published in 2019 by Manfredini et al. This meta-analysis looks at seven different studies, including the one from Michigan. These studies and their high-level findings are summarized below:

In all, five of the seven found a significant increase in heart attacks the week after the Spring clock change, though the magnitude of the increase varied (particularly in the studies with smaller sample sizes). The meta-analysis of these studies pointed to an overall impact of 5% (see below).

In the Fall, the results were far less conclusive, with most studies seeing insignificant decreases in the rate of heart attacks (the Culic study represents a significant outlier, potentially due to the small sample size), and the meta-analysis suggesting no significant impact.

Based on the findings from the meta-analysis, it seems that, contrary to the findings of the Michigan paper (which had the largest sample size in the meta-analysis), there may indeed be an increase in heart attack risk after the Spring clock change, and it’s not apparent that this increase is cancelled out by a decrease in the Fall. The Swedish data in particular (used for both studies by Janszky) points to a significant impact and represents 50%+ of the weighting for both Fall and Spring. However, a couple issues remain:

• The daily findings vary significantly (with peak increases in risk suggested on Sunday, Monday, Tuesday, Wednesday, and Thursday across different studies), suggesting an incomplete picture of the relationship between sleep loss and health risk.
• The largest study points to no change in risk, and the second largest study suggests the risks essentially cancel out, meaning the meta-analysis results may be more a result of the weighting approach.

Neither of these issues is large enough to fully discount the results, and so for now, it seems reasonable to assume clock changes do cause some health impacts – but it also seems prudent to conduct further analysis (especially within the US) before leaning on the claims to drive policy decisions.