Category Archives: Sleep

If you live in (most places in) the United States as well as many other countries, you have reset your clocks back by one hour last night (or last week). How will that affect you and other people?
One possibility is that you are less likely to suffer a heart attack tomorrow morning than on any other Monday of the year. Why? Let me try to explain in as simple way as possible (hoping that oversimplification will not lead to intolerable degrees of inaccuracy).
Almost all biochemical, physiological and behavioral parameters in almost all (at least multicellular) organisms display diurnal (daily) rhythms and most of those are directly driven by the circadian clock (or, more properly, by the circadian system). Here is an old and famous chart displaying some of the peaks (acrophases) of various physiological functions in the human:

It may be a little fuzzy, but you can see that most of the peaks associated with the cardiovascular function are located in the afternoon. The acrophases you see late at night are for things like “duration of systole” and “duration of diastole” which means that the Heart Rate is slow during the night. Likewise, blood pressure is low during the night while we are asleep.
Around dawn, heart rate and blood pressure gradually rise. This is a direct result of the circadian clock driving the gradual rise in plasma epinephrine and cortisol. All four of those parameters (HR, BP, Epinephrine and Cortisol) rise roughly simultaneously at dawn and reach a mini-peak in the morning, at the time when we spontaneously wake up:

This rise prepares the body for awakening. After waking up, the heart parameters level off somewhat and then very slowly rise throughout the day until reaching their peak in the late afternoon.
Since the four curves tend to be similar and simultaneus in most cases in healthy humans, let’s make it easier and clearer to observe changes by focusing only on the Cortisol curve in the morning, with the understanding that the heart will respond to this with the simultaneous rise in heart rate and blood pressure. . This is how it looks on a day when we allow ourselves to wake up spontaneously:

But many of us do not have the luxury of waking up spontaneously every day. We use alarm clocks instead. If we set the alarm clock every day to exactly the same time (even on weekends), our circadian system will, in most cases (more likely in urban than rural areas, though), entrain to the daily Zeitgeber – the ring of the alarm-clock – with a particular phase-relationship. This usually means that the rise in cardiovascular parameters will start before the alarm, but will not quite yet reach the peak as in spontaneous awakening:

The problem is, many of us do not set the alarm clocks during the weekend. We let ourselves awake spontaneously on Saturday and Sunday, which allows our circadian clock to start drifting – slowly phase-delaying (because for most of us the freerunning period is somewhat longer than 24 hours). Thus, on Monday, when the alarm clock rings, the gradual rise of cortisol, heart rate and blood pressure will not yet be as far along as the previous week. The ring of the alarm clock will start the process of resetting of the circadian clock – but that is the long-term effect (may take a couple of days to complete, or longer.).
The short term effect is more dramatic – the ring of the alarm clock is an environmental stressor. As a result, epinephrine and cortisol (the two stress hormones) will immediately and dramatically shoot up, resulting in an instantenuous sharp rise in blood pressure and heart rate. And this sharp rise in cardiovascular parameters, if the heart is already damaged, can lead to a heart attack. This explains two facts: 1) that heart attacks happen more often on Mondays than other days of the week, and 2) that heart attacks happen more often in the morning, at the time of waking up, than at other times of day:

Now let’s see what happens tomorrow, the day after the time-change. Over the weekend, while you were sleeping in, your circadian system drifted a little, phase delaying by about 20 minutes on average (keep in mind that this is an average – there is a vast variation in the numerical value of the human freerunning circadian period). Thus, your cardiovascular parameters start rising about 20 minutes later tomorrow morning than last week. But, your alarm clock will ring an entire hour later than last week – giving you an average of a 40-minute advantage. Your heart will be better prepared for the stress of hearing the ringing than on any other Monday during the year:

Now let’s fast-forward another six month to the Spring Forward weekend some time in March or April of next year. Your circadian system delays about 20 minutes during the weekend. On top of that, your alarm clock will ring an hour earlier on that Monday than the week before. Thus, your cardiovascular system is even further behind (80 minutes) than usual. The effect of the stress of the alarm will be thus greater – the rise in BP and HR will be even faster and larger than usual. Thus, if your heart is already damaged in some way, your chances of suffering an infarct are greater on that Monday than on any other day of the year:

This is what circadian theory sugests – the greater number of heart attacks on Mondays than other days of the week (lowest during the weekend), the greatest number of heart attacks on the Monday following the Spring Forward time-change compared to other Mondays, and the lowest incidence of heart attacks on the Monday following the Fall Back time-change compared to other Mondays.
A couple of days ago, a short paper appeared that tested that theoretical prediction and found it exactly correct (Imre Janszky and Rickard Ljung, October 30, 2008, Shifts to and from Daylight Saving Time and Incidence of Myocardial Infarction, The New England Journal of Medicine, Volume 359:1966-1968, Number 18.). The authors looked at a large dataset of heart attacks in Sweden over a large period of time and saw that (if you look at the numbers) the greatest number of heart attacks happens on Mondays compared to other days of the week (and yes, the numbers are lowest during the weekend), the greatest number of heart attacks occur on the Monday following the Spring Forward time-change compared to Mondays two weeks before and after, and the lowest incidence of heart attacks happens on the Monday following the Fall Back time-change compared to Mondays two weeks before and after:

Thus, the predictions from the circadian theory were completely and clearly correct. But I was jarred by the conclusions that the authors drew from the data. They write:

The most plausible explanation for our findings is the adverse effect of sleep deprivation on cardiovascular health. According to experimental studies, this adverse effect includes the predominance of sympathetic activity and an increase in proinflammatory cytokine levels.3,4 Our data suggest that vulnerable people might benefit from avoiding sudden changes in their biologic rhythms.
It has been postulated that people in Western societies are chronically sleep deprived, since the average sleep duration decreased from 9.0 to 7.5 hours during the 20th century.4 Therefore, it is important to examine whether we can achieve beneficial effects with prolonged sleep. The finding that the possibility of additional sleep seems to be protective on the first workday after the autumn shift is intriguing. Monday is the day of the week associated with the highest risk of acute myocardial infarction, with the mental stress of starting a new workweek and the increase in activity suggested as an explanation.5 Our results raise the possibility that there is another, sleep-related component in the excess incidence of acute myocardial infarction on Monday. Sleep-diary studies suggest that bedtimes and wake-up times are usually later on weekend days than on weekdays; the earlier wake-up times on the first workday of the week and the consequent minor sleep deprivation can be hypothesized to have an adverse cardiovascular effect in some people. This effect would be less pronounced with the transition out of daylight saving time, since it allows for additional sleep. Studies are warranted to examine the possibility that a more stable weekly pattern of waking up in the morning and going to sleep at night or a somewhat later wake-up time on Monday might prevent some acute myocardial infarctions.

And in the quotes in the press release they say the same thing, so it is not a coincidence:

“It’s always been thought that it’s mainly due to an increase in stress ahead of the new working week,” says Dr Janszky. “But perhaps it’s also got something to do with the sleep disruption caused by the change in diurnal rhythm at the weekend.”

Dr.Isis has already noted this and drew the correct conclusion. She then goes on to say something that is right on the mark:

And, of course, my first thought is, what about all the other times we are sleep deprived by, you know, one hour. Is waking up in the middle of the night to feed Baby Isis potentially going to cause Dr. Isis to meet her maker early? In that case Baby Isis can freakin’ starve. But, this is the New England Journal of Medicine and Dr. Isis appreciates the innate need that authors who publish here have to include some clinical applicability in their work.

The authors responded to Dr.Isis in the comments on her blog and said, among else:

We wonder whether you have ever tried to publish a research letter somewhere. The number of citations (maximum 5!) and the number of words are strictly limited. Of course we are familiar with studies on circadian rhythms and cardiovascular physiology. There was simply no space to talk more about biological rhythms than we actually did.

But what they wrote betrays that even if they are familiar with the circadian literature, they do not really understand it. Nobody with any circadian background ever speculates about people’s conscious expectations of a stressful week as a cause of heart attacks on Monday mornings. Let me try to explain why I disagree with them on two points they raise (one of which I disagree with more strongly than the other).
1) Sleep Deprivation. It is important to clearly distinguish between the acute and the chronic sleep deprivation. Sleepiness at any given time of day is determined by two processes: a homeostatic drive that depends on the amount of sleep one had over a previous time period, and a circadian gating of sleepiness, i.e., at which time of day is one most likely to fall asleep. Sleep deprivation affects only the homeostatic drive and has nothing to do with circadian timing.
Humans, like most other animals, are tremendously flexible and resilient concerning acute sleep deprivation. Most of us had done all-nighters studying for exams, or partying all night with non ill effects – you just sleep off the sleep debt the next day or the next weekend and you are fine. Dr.Isis is not going to die because her baby wakes her up several times during the night. This is all part of a normal human ecology, and human physiology had adapted to such day-to-day variations in opportunities for sleep.
The Chronic sleep deprivation is a different animal altogether. This means that you are getting less sleep than you need day after day, week after week, month after month, year after year, with rarely or never sleeping off your sleep debt (“catching up on sleep”). As a result, your cognitive functions suffer. If you are a student, you will have difficulties understanding and retaining the material. If you are a part of the “creative class”, you will be less creative. If you are a scientist, you may be less able to clearly think through all your experiments, your data, and your conclusions. No matter what job you do, you will make more errors. You may suffer microsleep episodes while driving and die in a car wreck. Your immune system will be compromised so you will constantly have sniffles and colds, and may be more susceptible to other diseases.
And yes, a long term chronic sleep deprivation may eventually damage your heart to the extent that you are more susceptible to a heart attack. This means that you are more likely to suffer a heart attack, but has no influence on the timing of the heart attack – it is the misalignment between the natural circadian rhythms of your body and the social rhythms imposed via a very harsh stressor – the alarm clock – that determines the timing. Being sleep deprived over many years means you are more likely to have a heart attack, but cannot determine when. Losing just one hour of sleep will certainly have no effect at all.
Thus, the data presented in the paper have nothing to say about sleep deprivation.
2) Cytokines. These are small molecules involved in intercellular signaling in the immune system. Like everything else, they are synthesized in a diurnal manner. But they act slowly. Maybe they play some small part in the gradual damage of the heart in certain conditions (prolonged inflammation, for instance), thus they may, perhaps, have a role in increasing risk of a heart attack. But they play no role in timing of it. Thus they cannot be a causal factor in the data presented in the paper which are ONLY about timing, not the underlying causes. The data say nothing as to who will suffer a heart attack and why, only when you will suffer one if you do.
If I was commissioned to write a comprehensive review of sleep deprivation, I may have to force myself to wade through the frustratingly complicated and ambiguous literature on cytokines in order to write a short paragraphs under a subheading somewhere on the 27th page of the review.
If I had a severe word-limit and needed to present the data they showed in this paper, I would not waste the space by mentioning the word “cytokine” at all (frankly, that would not even cross my mind to do) as it is way down the list of potential causes of heart attack in general and has nothing to do with the timing of heart attacks at all, thus irrelevant to this paper.
So, it is nice they did the study. It confirms and puts clear numbers on what “everybody already knew for decades” in the circadian community. But their interpretation of the data was incorrect. This was a purely chronobiological study, yet they chose to present it as a part of their own pet project instead and tried mightily to make some kind of a connection to their favourite molecules, the cytokines, although nothing warranted that connection. Nails: meet hammer.
The fake-insulted, haughty and inappropriate way/tone they responded to Dr.Isis is something that is important to me professionally, as is there misunderstanding of both the role and the tone of science blogs, so I will revisit that issue in a separate post later. I promise. It is important.
But back to Daylight Saving Time. First, let me ask you (again) to see Larry’s post from last year, where you will find a lot of useful information and links about it. What is important to keep in mind is that DST itself is not the problem – it is the time-changes twice a year that are really troubling.
Another important thing to keep in mind is that DST was instituted in the past at the time when the world looked very different. At the time when a tiny sliver of the population is still involved in (quite automated and mechanized) agriculture, when electricity is used much more for other things than illumination (not to mention that even the simple incandescent light bulbs today are much more energy efficient than they used to be in the past, not to mention all the newfangled super-efficient light-bulbs available today), when many more people are working second and third shifts than before, when many more people work according to their own schedules – the whole idea of DST makes no sense any more.
Even if initially DST saved the economy some energy (and that is questionable), it certainly does not do so any more. And the social cost of traffic accidents and heart attacks is now much greater than any energy savings that theoretically we may save.
Furthermore, it now seems that circadian clocks are harder to shift than we thought in the past. Even that one-hour change may take some weeks to adjust to, as it is not just a singular clock but a system – the main pacemaker in the SCN may shift in a couple of days, but the entire system will be un-synchronized for some time as it may take several weeks for the peripheral clocks in the liver and intestine to catch up – leading to greater potential for other disorders, e.g., stomach ulcers.
The social clues (including the alarm clocks) may not be as good entraining agents as we thought before either, especially in rural areas where the natural lighting still has a profound effect.
Finally, the two time-change days of the year hit especially hard people with Bipolar Disorder and with Seasonal Affective Disorder – not such a small minority put together, and certainly not worth whatever positives one may find in the concept of DST. We should pick one time and stick with it. It is the shifts that cost the society much more than any potential benefits of DST.
Related reading:
Daylight Saving Time
Daylight Savings Time worse than previously thought
Time
Sun Time is the Real Time
Seasonal Affective Disorder – The Basics
Lesson of the Day: Circadian Clocks are HARD to shift!
Lithium, Circadian Clocks and Bipolar Disorder
Everything You Always Wanted To Know About Sleep (But Were Too Afraid To Ask)