Can you prevent colds, and does cold weather really matter?

A cold starts when a virus, such as a rhinovirus, reaches your nose or throat⁸. Much of the best evidence comes from virus-challenge studies, in which quarantined volunteers are given nasal drops containing a cold virus and monitored.

Sleep. In 164 adults, those who slept under five hours or five to six hours a night (measured with a wrist sensor) were about four times as likely to develop a cold as those who slept more than seven; six to seven hours was not clearly different¹. An earlier study of 153 adults found that under seven hours of self-reported sleep carried nearly three times the risk of eight hours or more⁹. See sleep hygiene for ways to get there.

Stress and connection. Among 394 volunteers, rates of infection and of clinical colds rose steadily with psychological stress scores⁸. Among 276 others, people with only one to three kinds of social ties were 4.2 times as likely to catch a cold as people with six or more⁴.

Hand hygiene. A 2023 Cochrane review pooled nine trials (52,105 people) and found that hand hygiene lowered acute respiratory infections by 14 percent (relative risk 0.86), with moderate certainty; the effect on confirmed influenza was not significant².

Exercise. In 1,002 adults followed for 12 weeks, those who exercised aerobically five or more days a week had 43 percent fewer days with an upper respiratory infection than people who barely exercised³. This is an association, not proof.

Vitamin D. A 2017 analysis of individual data from 25 trials (11,321 people) found that daily or weekly vitamin D lowered infection risk (odds ratio 0.88), most of all in people with very low blood levels (below 25 nmol/L)⁶. A 2025 update of 40 trials (61,589 people) found no statistically significant protection overall⁷, and a UK trial of 6,200 people offered vitamin D testing and supplements saw no reduction in infections¹⁰. See vitamin D supplementation for dosing.

Cold weather. Cold and dry air are linked to more colds, but the reason is unclear. Among 892 military conscripts, each 1 °C drop in outdoor temperature went with a 4.3 percent higher risk of an upper respiratory infection, and drier air raised it too¹¹. In a randomized trial of 180 people, 13 of 90 whose feet were chilled reported a cold within five days versus 5 of 90 controls¹². Winter also means more time indoors near other people, which studies like these cannot fully separate. For zinc, vitamin C, and other treatments, see Natural Prevention and Recovery.

Sleep and colds. Prather and colleagues tracked 164 healthy adults aged 18 to 55 with a wrist motion sensor (actigraphy) for seven nights, then quarantined them, gave rhinovirus nose drops, and watched them for five days¹. Compared with people sleeping more than seven hours, those sleeping under five hours had an odds ratio of 4.50 (95% CI 1.08 to 18.69) for a clinical cold and those sleeping five to six hours 4.24 (1.08 to 16.71); six to seven hours was 1.66 (0.40 to 6.95). The pattern held after adjusting for antibody levels, season, body mass, and health habits, but sleep fragmentation was unrelated. Cohen's earlier study of 153 adults used 14 days of self-reported sleep: under seven hours carried an odds ratio of 2.94 (1.18 to 7.30) against eight or more, and sleep efficiency under 92 percent carried 5.50 (2.08 to 14.48) against 98 percent or more⁹. The confidence intervals are wide, both studies came from one research group, and a quarantined volunteer given a measured dose of virus is not the same as real-world exposure.

Stress and social ties. In the 1991 New England Journal of Medicine study, 394 volunteers completed stress questionnaires and were then given one of five respiratory viruses⁸. Infection rates ranged from roughly 74 to 90 percent and clinical colds from roughly 27 to 47 percent, rising in a dose-response way with stress; smoking, alcohol, exercise, diet, sleep quality, and white-cell counts did not explain the link. In the 1997 JAMA study, 276 volunteers reported 12 types of social ties before exposure to one of two rhinoviruses; those with more types of ties had fewer colds, made less mucus, cleared their noses better, and shed less virus (adjusted relative risk 4.2 for the fewest ties versus the most)⁴. These associations are strong and consistent, but they stay correlational because stress and isolation cannot be randomly assigned.

Hand hygiene. The 2023 Cochrane update covers 78 randomized trials of physical interventions, 19 of them on hand hygiene with usable data². For acute respiratory infection the relative risk was 0.86 (0.81 to 0.90; nine trials; 52,105 people; moderate certainty). For influenza-like illness it was 0.94 (0.81 to 1.09) and for laboratory-confirmed influenza 0.91 (0.63 to 1.30; eight trials), so there was no clear effect on those narrower outcomes. The reviewers flagged high or unclear risk of bias and low adherence in many trials.

Exercise. Nieman's 12-week study of 1,002 adults aged 18 to 85 found 43 percent fewer days with an upper respiratory infection in people exercising aerobically five or more days a week than in those exercising once a week or less, and 46 percent fewer when comparing the high with the low self-rated fitness group; symptom severity was 32 to 41 percent lower³. Everything was self-reported and observational.

Vitamin D: conflicting results. The 2017 individual-participant analysis (25 trials, 11,321 participants) reported an overall odds ratio of 0.88 (0.81 to 0.96), 0.81 for daily or weekly dosing without bolus doses, and 0.30 (0.17 to 0.53) for people whose baseline level was under 25 nmol/L; bolus doses showed no benefit⁶. The 2025 update of stratified data (40 studies, 61,589 participants) found an odds ratio of 0.94 (0.88 to 1.00; P = 0.057) and no effect modification by baseline vitamin D status, dosing frequency, or age; the funnel plot was asymmetric (Egger's test P = 0.002), which can signal small-study or publication bias⁷. CORONAVIT randomized 6,200 UK adults not already taking vitamin D to a test-and-treat offer (800 or 3,200 IU a day for six months) or no offer; an infection occurred in 4.6 percent with no offer, 5.7 percent at the lower dose (odds ratio 1.26; 0.96 to 1.66), and 5.0 percent at the higher dose (1.09; 0.82 to 1.46)¹⁰. The honest reading: protection in people who are severely deficient is plausible; for everyone else it is not established.

Cold exposure. Johnson and Eccles randomized 180 healthy volunteers to a foot-chilling procedure or a control procedure: 13 of 90 chilled people reported a cold in the next four to five days versus 5 of 90 controls (P = 0.047), and mean symptom scores over days 1 to 4 were higher (5.16 versus 2.89)¹². The outcome was self-reported symptoms without virus testing, and the authors note that earlier clinical research had failed to show an effect of cold exposure on infection. Mäkinen's population study of 892 conscripts counted 643 respiratory infection episodes; each 1 °C fall in temperature raised the estimated risk of upper respiratory infection by 4.3 percent, and each 1 g/m³ fall in absolute humidity by 10.0 percent¹¹. That study is observational.

Bottom line. Confidence 3 of 5. Sleep, social ties, and hand hygiene have consistent support and little downside; exercise and cold exposure are suggestive; vitamin D is disputed. None of this makes colds avoidable, and the overall effects are modest.