Sleep and catching a cold: what the virus-challenge studies show

Most advice about sleep and colds rests on a plausible guess: tired people get sick more. A handful of studies went further and tested it in a way that is unusual in medicine. They took healthy volunteers, put a cold virus into their noses, kept them in quarantine, and checked who actually got ill. What the volunteers had done with their sleep beforehand was already on record.

This is informational material, not medical advice. If symptoms are severe, getting worse, or not going away, that is a question for a doctor. So are sleep problems that keep coming back.

The design: sleep first, virus second

Deliberately infecting volunteers sounds odd, and it is also what makes the evidence clean. In an ordinary study of "who catches colds", people meet viruses at different times, in different doses, in different places. Here every volunteer got the same rhinovirus (a common cold virus) as nasal drops, so exposure was the same for everyone. Researchers then watched for a "clinical cold": confirmed infection plus objective signs of illness, not just a volunteer saying they felt rough.

The same research group ran this design twice, six years apart, and measured sleep differently each time. The difference matters.

Study one: sleep diaries for two weeks

In the 2009 study, 153 healthy adults aged 21 to 55 reported their sleep every day for 14 days. They gave sleep duration and sleep efficiency, which is the share of time in bed that you actually spend asleep. Then they were quarantined and given the virus (Cohen et al., 2009).

People who had averaged less than 7 hours were 2.94 times more likely to develop a cold than people who had averaged 8 hours or more (95% confidence interval 1.18 to 7.30). People with sleep efficiency under 92% were 5.50 times more likely than those at 98% or higher (95% CI 2.08 to 14.48). The authors checked whether pre-existing antibodies, age and other demographics, season, body mass, socioeconomic status, psychological variables and health habits could explain it. They did not.

One detail is easy to skim past. How rested people felt was not associated with who got a cold. The share of days they rated themselves as rested told the researchers nothing, while the hours and the efficiency did.

Study two: a wrist sensor instead of memory

The weak point of study one is that it relied on self-report, and people are poor judges of their own nights. The 2015 study addressed that. Researchers had 164 healthy adults aged 18 to 55 wear a wrist actigraph (a movement sensor that estimates when you are asleep) and keep a sleep diary for 7 consecutive days. Then came quarantine, rhinovirus drops, and five days of monitoring (Prather et al., 2015).

Compared with people who slept more than 7 hours a night, the odds of a clinical cold were:

• under 5 hours: 4.50 times higher (95% CI 1.08 to 18.69)
• 5 to 6 hours: 4.24 times higher (95% CI 1.08 to 16.71)
• 6 to 7 hours: 1.66 times higher (95% CI 0.40 to 6.95), which the authors did not count as a greater risk

The estimates held after adjusting for antibodies before the challenge, demographics, season, body mass index, psychological variables and health practices.

How much to trust the numbers

Look at those confidence intervals, the ranges in brackets. The one for under 5 hours runs from 1.08 to 18.69. In plain words, the data are compatible with almost no effect and with one roughly four times larger than the headline figure. With 164 people, that is what you get. The direction is consistent across both studies, and it is the direction, more than any single ratio, that deserves attention.

The 6-to-7-hour group deserves its own sentence. Its interval (0.40 to 6.95) comfortably includes "no difference", and it also includes a sizeable one. The study could not tell. So "6 hours is fine" and "6 hours is risky" are both claims that this data does not make.

There is also an unresolved difference between the two studies. In the diary study, sleep efficiency stood out. In the actigraphy study, sleep fragmentation, meaning how broken up the night was, was unrelated to susceptibility. A sensor and a diary may be measuring different things under the same name. Which of them is closer to what the immune system responds to is not settled.

Finally, these studies observed sleep habits that people already had. Nobody was assigned to sleep five hours or eight. They did not test whether sleeping more lowers anyone's risk, and nothing here supports a rule like "sleep X hours to avoid colds". The people who sleep less may also differ in ways the researchers did not or could not measure.

A second line of evidence: vaccines

If short sleep changes how the immune system responds, it should show up outside colds too. A 2012 study followed 125 healthy adults aged 40 to 60 through the standard three-dose hepatitis B vaccine series, measuring sleep with actigraphy and electronic diaries (Prather et al., 2012).

Shorter sleep measured by actigraphy was associated with a lower secondary antibody response. It was also linked to a lower likelihood of being clinically protected six months after the final dose. Sleep efficiency and subjective sleep quality were not significant predictors. The authors wrote that short sleep "may negatively affect" antibody responses.

The evidence type here is observational: healthy people, vaccine, sleep measured alongside. This is a finding about one vaccine in one age group, not guidance about vaccination. It matters for this article because it is one possible explanation the authors offer for the cold results. If short sleep dulls part of the immune response, the cold finding has a mechanism to hang on. It is a hypothesis that fits, not a demonstrated pathway in these cold studies.

The arrow also points the other way

So far sleep comes first and illness second. A review of the field describes a two-way link: sleep affects innate and adaptive immunity and is associated with lower infection risk, while immune activation from an infection changes sleep (Besedovsky et al., 2019). Being ill can increase how long and how deeply you sleep, but it can also disrupt sleep. The review adds that prolonged sleep deficiency can lead to chronic low-grade inflammation.

This has a practical reading for anyone looking at their own numbers. Say you have a rough week of sleep in October and a cold in the second week. The virus-challenge design rules this confusion out, because the sleep was measured before the virus was given. Your life has no such guarantee. A bad stretch might precede an illness, or it might be the first part of one that was already under way. Looking back from the day you feel sick, you usually can't tell which.

What this means for your own data

None of this gives you a number to hit, and none of it gives you a score. Sleeping less than usual is not a sentence, and a good night is not a shield. What the studies offer is a reason for curiosity about your own pattern across a season.

MeteoHealth reads sleep, HRV and resting heart rate from Apple Health. It logs the local outdoor weather (pressure, temperature, relative humidity, wind speed and UV index) next to your symptom entries, and it looks for relationships in your own records using Pearson correlation and ANOVA, with false-discovery-rate correction so that chance patterns are less likely to look like findings. The app tracks nine fixed symptoms, including fatigue, drowsiness and insomnia. It has no entries for cough, sore throat or fever, and it does not detect, diagnose or predict infections.

So the realistic use is modest. Over weeks, you can look at your sleep trend next to the days you logged fatigue, and see whether they move together for you. That is observation, not a risk score, and when the trend and an illness do coincide, the review above is the reminder that the order of events may not be the one you assume. For a stretch of poor sleep that does not lift, or one that comes with symptoms, bring the records to a doctor. The winter article covers the weather side of the season, and sleep debt covers why a single hours figure is a weak summary of a week.