This book is written about mainland China. Where it says national, domestic or the state, it means China, not your country.

How the body tells the time, and why night shifts harm people

This is the long version of Section 2, item 39 (the longer you work night shifts, the higher the cardiovascular risk). That item in the main text only does the sums. This piece explains the reasoning behind them.

The question to answer is this: can the body really tell what time it is? If it can, why is night work not something you get used to after a few more nights awake?

1. There really is an organ in the body that keeps time

Deep in the human brain, in the hypothalamus, there is a small cluster of neurons called the suprachiasmatic nucleus, about 20,000 cells. It is the master clock of the whole body.

It keeps time through a set of "clock genes" that inhibit one another. Two proteins, CLOCK and BMAL1, switch on the transcription of two other genes, PER and CRY. After the PER and CRY proteins build up, they turn back and switch off the pair of proteins that started them. When those proteins have broken down, the cycle begins again. One full turn takes about 24 hours.

So this clock does not work out the time by watching the outside world. If a person is shut in a room with no light, no clocks and no social cues at all, this rhythm keeps running all the same. Only the period is not exactly 24 hours, and it slowly drifts from day to day. The 2017 Nobel Prize in Physiology or Medicine went to the three researchers who discovered this molecular mechanism.

2. The clock is set by light, and not through the line used for "seeing things"

In the retina, besides the cells that form images, there is a special kind of light-sensing cell. They contain a pigment called melanopsin. They take no part in seeing things clearly. They send the strength of the light straight to the suprachiasmatic nucleus.

So whether light can set the body clock has no direct link to whether the eyes can see. Some completely blind people still have body clocks that light can set.

This is also why Section 3, item 2 (get up at a fixed time, also at weekends) has to be combined with light in the morning.

3. The key point: there is more than one clock

The liver, the gut, the heart and fat tissue each have their own set of peripheral clocks.

The master clock is set by light. The peripheral clocks are set, to a large extent, by "when you eat".

On a normal routine the two agree. You see light in the day and eat in the day, so the master clock and the peripheral clocks point to the same time. That is why in ordinary life you do not notice they are two separate systems at all.

4. The problem with night shifts is not "staying up". It is that the two clocks are pulled apart

A person on night shifts still sees light in the day and still lives a daytime social life. So the master clock generally stays locked to the outside day and night cycle.

But you eat, work and run your metabolism in the biological middle of the night. The peripheral clocks are pulled to the other side.

This state is called circadian misalignment. The problem is not that you sleep a few hours less. It is that the clocks all over the body do not match each other.

5. And you cannot adapt

Some people think that a person who works a fixed night shift for a long time will surely adapt.

One review pooled six studies that used the rhythm of melatonin secretion to test people on long-term fixed night shifts. The melatonin rhythm is the most reliable indicator of the state of the internal body clock.

The result: fewer than 3% of people reached "complete" adaptation. Fewer than a quarter reached "substantial" adaptation.

It made no difference how many years they had done it. The reason is in the previous section: you still have to see light in the day and live a daytime social life, and these cues keep pulling the master clock back.

Source: Folkard S (2008). Do permanent night workers show circadian adjustment? A review based on the endogenous melatonin rhythm. Chronobiology International. https://doi.org/10.1080/07420520802106835

6. How do we know the misalignment itself does the harm, and not just less sleep

This is the question most worth asking in the whole matter. People on night shifts sleep less, eat in a disordered way, are under more stress and smoke more. Any of these could be the real cause.

Fortunately this can be isolated in a laboratory.

The 2009 study by Scheer and colleagues: 10 adults lived in a laboratory for 10 days. The researchers artificially set the "day" to 28 hours, which forced their eating and sleeping to pass through every point of the body clock. Four meals a day, with fixed calories.

When behavior and the body clock were out of step by about 12 hours, the results were:

Leptin fell by 17% (leptin is the hormone that makes a person feel full)
Blood glucose rose by 6%, and this happened while insulin had already risen by 22%
The daily rhythm of cortisol was completely reversed
Mean arterial pressure rose by 3%
Sleep efficiency fell by 20%

The most important point: of the 8 people with enough data, 3 had after-meal blood glucose that entered the prediabetic range, in only 10 days.

Because calories were fixed and the experiment was controlled in a laboratory, these changes cannot be put down to "less sleep" or "eating more". Misalignment alone is enough to cause these changes.

Source: Scheer FA, Hilton MF, Mantzoros CS, Shea SA (2009). Adverse metabolic and cardiovascular consequences of circadian misalignment. PNAS, 106(11), 4453-4458. https://doi.org/10.1073/pnas.0808180106

7. The first thing you can change: move eating to the daytime

Since the peripheral clocks are set by when you eat, there is one direction you can try: work the night shift as usual, but do not eat at night.

A randomized controlled trial did this (registration number NCT02291952). Twenty healthy people simulated night work and were randomly divided into two groups. One group ate at night and in the day, following the common habit of night workers. The other group ate only in the day.

The report on the primary outcome said: in the group that ate at night, the central rhythm and the peripheral glucose rhythm were out of step, and glucose tolerance got worse. The group that ate only in the day showed neither of these.

The cardiovascular report (a secondary analysis of the same trial) said: in the group that ate at night, pNN50, an index of cardiac vagal regulation, fell by 25.7% and RMSSD fell by 14.3%; LF/HF, which reflects the balance between sympathetic and vagal activity, rose by 5.5%; and PAI-1, a pro-thrombotic factor, rose by 23.9%. In the group that ate only in the day, these changes did not appear.

But this was only 20 people, and it was simulated night work, not real night workers.

A systematic review that looked specifically for this kind of randomized trial searched six databases and found only 4 trials. The results were not tidy either: some trials saw improvement in fasting insulin and insulin resistance, and some did not; for high-density lipoprotein, none of the four trials saw a change; and for after-meal blood glucose, the approach of fasting for 10.75 hours at night was indeed better than the approach of eating meals and a late-night snack at night. The review's authors said themselves that there were too few trials to generalize.

So this is all that can be said at present: during night shifts, concentrating eating in the daytime as far as you can has support from randomized trials, but the evidence is still thin. It costs nothing and is not hard to do, so try it if you are willing. There is a second thing you can try, in the next section: use light to move the whole body clock over.

Sources: Chellappa SL et al. (2021). Daytime eating prevents internal circadian misalignment and glucose intolerance in night work. Science Advances. https://doi.org/10.1126/sciadv.abg9910; Chellappa SL et al. (2025). Daytime eating during simulated night work mitigates changes in cardiovascular risk factors: secondary analyses of a randomized controlled trial. Nature Communications. https://doi.org/10.1038/s41467-025-57846-y; Wulandari F et al. (2026). Cardiometabolic status among shift workers under meal time regulation: A systematic review of randomized controlled trials. Nutrition and Health. https://doi.org/10.1177/02601060261464874

8. Can you move the clock over

Everything so far says "misalignment is harmful". Then the other way round: can you move the whole clock to the night-shift side?

In theory yes, and it has been done. But the light it needs is an order of magnitude brighter than most people think.

The 1990 study by Czeisler and colleagues: 8 young men worked a week of night shifts, and two conditions were compared.

In the control condition, the light at night was ordinary indoor lighting, about 150 lux. Lux is the unit of illuminance, and a room with ordinary fluorescent lights is around one or two hundred. The result was that over six nights in a row, the lowest point of body temperature still fell at about half past three in the morning. The lowest point of body temperature is the standard indicator of where the body clock points. This shows that the body clock did not move at all.

In the treatment condition, they were given bright light of 7,000 to 12,000 lux at night, and slept in the day in a nearly completely dark place. After four days, the lowest point of body temperature had moved to nearly 3 in the afternoon. That is, the whole clock was turned upside down.

Source: Czeisler CA, Johnson MP, Duffy JF, Brown EN, Ronda JM, Kronauer RE (1990). Exposure to bright light and darkness to treat physiologic maladaptation to night work. New England Journal of Medicine, 322(18), 1253-1259. https://doi.org/10.1056/NEJM199005033221801

First, the light at night must be bright enough. Working under ordinary fluorescent lights is the same as the control group above, in which "nothing happened". To have an effect you need a dedicated light therapy box.

Second, sleep in the day must be strictly shaded from light. Blackout curtains and an eye mask. The experiment above did "bright light at night" and "nearly complete dark in the day" together. It was not a choice between the two.

Third, wear sunglasses on the way home from work. This one is the easiest to leave out. In a 1994 experiment with 50 people, bright light and sunglasses each worked on their own, and doing both gave the best effect; without sunglasses, the body clock moved only in the direction of advancing, and could not move in the direction needed for sleeping in the day. For the source, see the sources column of Section 2, item 39 (if you can move to another post, move early from night shifts).

To see why the third matters so much, look at how light acts: the same light, falling at different times, pushes the body clock in opposite directions. Light before the lowest point of body temperature pushes the clock later; light after it pulls the clock earlier.

When your night shift ends and you walk home, it is just after the lowest point of body temperature. What this morning light does is in the opposite direction to what you did by sitting in front of the light box all night.

The laboratory can do it, daily life cannot

The few people in the laboratory did not go outside for a week. In real life, on your days off you live a daytime life: you pick up the children, spend time with family, run errands. Once you do that, the clock is pulled back. This is where the figure in Section 5 comes from: fewer than 3% of long-term fixed night workers can adapt completely.

So in the real world the aim is usually not to turn the clock fully over, but to move only part of the way, so that the clock stops at a middle position of "awake enough at work, and not completely ruined on days off".

How much you can get in practice

A meta-analysis pooling 11 studies: light therapy gave shift workers about 32.5 minutes more total sleep time, and raised sleep efficiency by about 2.91. The effective dose was 900 to 6,000 lux, for at least 1 hour each night.

Another laboratory study used intermittent bright light. In the treatment group, melatonin secretion overlapped the scheduled sleep period by 4.90 hours; in the control group it was only 2.62 hours (P=0.002). The decline in cognitive performance that the control group showed on the last night shift was held down in the treatment group. The authors state plainly in their conclusion that this approach still needs to be tested in real night workers.

Sources: Zhao C, Li N, Miao W, He Y, Lin Y (2025). A systematic review and meta-analysis on light therapy for sleep disorders in shift workers. Scientific Reports. https://doi.org/10.1038/s41598-024-83789-3; Lammers-van der Holst HM et al. (2021). Efficacy of intermittent exposure to bright light for treating maladaptation to night work on a counterclockwise shift work rotation. Scandinavian Journal of Work, Environment & Health. https://doi.org/10.5271/sjweh.3953

One thing nobody has measured

"If the clock is moved the right way, does cardiovascular risk fall with it" -- no trial has measured this.

What is known is that misalignment is harmful, which the experiment in Section 6 showed. But that does not mean the reverse, "moved correctly means safe". And whether it is good or bad to go back and forth like this, with the clock moved back again on days off, has likewise not been measured by anyone.

The endpoints of the two measures above are, for one, metabolic and cardiovascular indicators, and for the other, sleep. Neither reached illness and death.

9. What this piece does not cover

How much higher night shifts make cardiovascular risk, and how it is counted by years, is in Section 2, item 39 (the longer you work night shifts, the higher the cardiovascular risk).

How to make up sleep after staying up late is in Section 2, item 38 (make up sleep the night after staying up late).

How long to sleep at night and whether your routine is regular is in Section 2, item 13 (sleep about 7 hours a night, on a fixed routine).

Light in the morning and a fixed time to get up are in Section 3, item 2 (get up at a fixed time, also at weekends).