What should you eat and drink when you are sick?

Fluids first

Fever, vomiting and diarrhea increase the water your body loses, so fluids come first^1,10. Yet no randomized trial has tested whether extra fluids shorten a respiratory infection, and some observational reports suggest over-drinking may harm in chest infections. For a cold or flu, drink to thirst and keep sipping¹.

Stomach bugs are different. European pediatric guidelines name oral rehydration solution (a measured mix of water, salt and sugar) as the main treatment, to be started as soon as possible, with regular feeding, including milk and breastfeeding, continuing¹⁰. In 647 young children with mild gastroenteritis, half-strength apple juice followed by preferred drinks led to fewer treatment failures than a branded electrolyte drink¹¹. That applies to mildly dehydrated children; guidelines still favor oral rehydration solution¹⁰. See diarrhea and stomach bugs.

Warm drinks and soup

In 30 people with colds or flu, a hot fruit drink relieved runny nose, cough, sneezing, sore throat, chilliness and tiredness, while the same drink at room temperature helped only three of those symptoms². In test-tube work, chicken soup and each of its ingredients slowed neutrophils, white blood cells that drive inflammation⁴. Nobody has shown that soup shortens a cold, so think of it as a hydrating comfort food.

Honey

For children aged 12 months and over, honey probably eases cough more than no treatment or placebo, based on six trials, mostly after one night's use⁸. It must never be given to babies under 12 months, because it can carry the spores that cause infant botulism⁹. See what calms a cough.

Feed a cold, starve a fever?

In mouse models of bacterial sepsis, not eating protected and feeding harmed; in influenza and viral sepsis, feeding protected⁵. In human volunteers, a short report found that eating raised interferon-gamma and fasting raised interleukin-4, immune markers rather than illness outcomes³. A 2024 review found too little human data to advise by type of infection⁶. Let your appetite lead: small amounts of what appeals, and skip what does not.

Milk and mucus

In 60 volunteers given a cold virus, milk and dairy intake was not linked to more nasal secretions or congestion⁷. If milk feels unpleasant while you are ill, skip it, but there is no evidence you need to.

Watch for dehydration

For feverish children under five, NICE advises parents to look for a sunken soft spot, dry mouth, sunken eyes and no tears, and to seek advice if they appear¹². See Fever: Friend or Foe?. For gentle meal ideas, the recipe library has a "When You're Sick" collection.

Fluids

The Cochrane review by Guppy and colleagues searched five databases up to late 2010 for randomized trials of advising people with acute respiratory infections to drink more. It found none, so there is no trial evidence either way. The authors noted that some observational studies of lower respiratory tract infections linked higher fluid intake to harm, and that antidiuretic hormone, which makes the body hold on to water, is raised in some chest infections¹. For gastroenteritis in children, the 2014 European guideline puts early oral rehydration with a hypoosmolar solution first and says feeding should continue¹⁰. The Toronto trial randomized 647 children aged 6 to 60 months with mild gastroenteritis and minimal dehydration (68% had no sign of dehydration). Children given half-strength apple juice and then their preferred fluids (n=323) had treatment failure within seven days less often than those given apple-flavored electrolyte solution (n=324): 16.7% versus 25.0%. Intravenous rehydration was needed in 2.5% versus 9.0%; hospitalization and the frequency of diarrhea and vomiting did not differ¹¹. It was single-blind, at one emergency department in a high-income country, so it does not overturn standard advice for moderate or severe dehydration.

Hot drinks and soup

Sanu and Eccles compared a hot fruit drink with the same drink at room temperature in 30 people with colds or flu. The hot drink did not change objectively measured nasal airflow, but it improved the sense of airflow and gave immediate, sustained relief from six symptoms; the cooler drink helped three. The authors discuss the effects in terms of placebo and of changes in salivation and airway secretions². Rennard and colleagues found that a traditional chicken soup inhibited neutrophil migration in a chemotaxis chamber in a concentration-dependent manner; the chicken and every vegetable did so on their own, and commercial soups differed greatly⁴. The drink trial measured symptoms in a single session and the soup work was a laboratory assay: neither measured how long a cold lasts.

Feeding and fasting

Wang and colleagues showed in mice that the effect of eating depends on the infection. In bacterial sepsis models, anorexia protected and nutritional supplementation was harmful, with glucose necessary and sufficient for those effects. In influenza and viral sepsis models the opposite held: supplementation protected, and blocking glucose use was lethal⁵. In a two-page report, van den Brink and colleagues found that activated T cells from human volunteers made more interferon-gamma after food intake and more interleukin-4 after food deprivation, shifting the balance between T-helper 1 and 2 responses; these were immune markers, not illness outcomes³. A 2024 review concluded that clinically applicable human data are limited, that needs appear to depend on the specific organism rather than a simple bacteria-versus-virus split, and that age, baseline nutrition, illness phase and other conditions all matter⁶. Nothing here tells a person with a cold what to put on a plate.

Milk and honey

Pinnock and colleagues followed 51 of 60 volunteers infected with rhinovirus-2 for 10 days (510 person-days), recording milk intake (0 to 11 glasses a day, mean 2.7) and weighing used tissues. Milk and dairy intake was not associated with nasal secretion weight or congestion symptoms. People who believed that milk makes mucus reported more cough and congestion but did not produce more secretions⁷. This was a study in healthy adults with one virus, and intake was observed, not assigned. The honey Cochrane review pooled six trials with 899 children aged 12 months to 18 years. Against no treatment, honey lowered cough-frequency scores by a mean difference of -1.05 points on a 7-point scale (95% CI -1.48 to -0.62; 154 children; moderate certainty); against placebo the difference was -1.62 (95% CI -3.02 to -0.22; 402 children). Honey and dextromethorphan did not differ clearly (low certainty), and beyond three days honey had no advantage over placebo or salbutamol. Most children were treated for one night, and the authors saw no strong evidence for or against honey⁸. The infant botulism warning rests on repeated epidemiologic findings and positive honey samples, and the CDC advises against honey for babies under 12 months⁹.

Where the evidence is thin or conflicting

Advice to drink more has never been tested in a trial for colds or flu, and observational signals for chest infections point the other way¹. The hot-drink benefit may largely reflect expectation², and the soup effect is shown only in a test tube⁴. Mouse studies show feeding helps in one kind of infection and hurts in another, and the human report measured only immune markers, so the old saying may be half right for the wrong reasons^3,5,6. The milk study cannot rule out effects at higher intakes or in people with a milk allergy⁷.

Bottom line

Confidence is low (2 of 5). Fluids and oral rehydration are well supported for children's diarrhea and vomiting^10,11, and honey has moderate-certainty support for night cough in children over one year⁸. Hot drinks, soup and the feed-or-fast rule are comfort-level advice built on a small trial, lab work, animals or immune markers^2,3,4,5,6. No food or diet has been shown to shorten a cold or flu.