Water

TL;DR. The eight-glasses-a-day rule has no scientific origin anyone can find, and about 20 to 30 percent of your fluid arrives in food anyway. Thirst is a good guide for most healthy adults; the groups where it is not are older people, whose thirst sensation genuinely declines, and infants, athletes, and anyone ill. Tap water in a country with regulated supply is one of the most tested substances you will ever consume, and bottled water is not measurably safer. The real water hazards are the ones you cannot taste: lead from old plumbing, and, in a small number of places, contamination events.

1. What water does

Your body is 50 to 60 percent water by weight (more in infants, less in older adults and in people with more body fat, since fat tissue holds less water than muscle).

Daily balance, approximate for a sedentary adult:

Note the third row: you manufacture roughly 250 to 350 mL of water a day as a by-product of metabolism. This is why hibernating and desert animals can go long periods without drinking, and why fat oxidation produces water.

2. How much do you need?

The official figures (EFSA, total water from all sources including food):

• Men: 2.5 L/day
• Women: 2.0 L/day
• Roughly 20 to 30 percent of that comes from food

The US Institute of Medicine sets 3.7 L for men and 2.7 L for women, also total, which is why different sources appear to disagree.

"Eight glasses of water a day" has no identifiable evidence base. The most-cited origin is a 1945 US Food and Nutrition Board recommendation of about 2.5 L a day, whose very next sentence noted that most of this is contained in prepared foods, a sentence that dropped out of subsequent retellings. Heinz Valtin's 2002 review in the American Journal of Physiology searched for the evidence and found none.

What actually determines your needs: body size, activity, ambient temperature and humidity, altitude, diet (more protein and salt means more water for excretion), pregnancy and breastfeeding, illness (fever, vomiting, diarrhoea), and medications (diuretics).

Thirst works for most healthy adults. It is triggered by roughly a 1 to 2 percent rise in plasma osmolality, well before dehydration becomes a problem. The groups where thirst is an unreliable guide are the ones that matter:

• Older adults: thirst sensation declines with age and kidney concentrating ability falls. Dehydration is a common and under-recognised cause of confusion, falls, constipation, and hospital admission in older people.
• Infants and young children: higher surface-area-to-mass ratio, cannot ask.
• Athletes and manual workers in heat: sweat losses can reach 1 to 2 L/hour and thirst lags behind.
• Anyone with vomiting or diarrhoea.

The urine colour check is the most practical self-assessment: pale straw is fine, dark amber means drink more, colourless throughout the day may mean you are drinking more than you need. B vitamins turn urine bright yellow and invalidate the test for a few hours.

3. Dehydration and overhydration

Dehydration effects appear early:

The cognitive effects at 1 to 2 percent are real and replicated: attention, short-term memory, and mood measurably decline, and this is the most practically relevant finding for ordinary life.

Overhydration is rarer and genuinely dangerous. Exercise-associated hyponatraemia occurs when large volumes of plain water dilute blood sodium: confusion, nausea, headache, seizures, cerebral oedema, and death. It has killed marathon runners, most often slower participants who drank at every station, and it has killed people in water-drinking contests and in some hazing incidents. Kidneys can excrete roughly 0.8 to 1.0 L per hour at most; exceeding that sustained is how it happens.

The practical rule for endurance events changed as a result: drink to thirst, not to a schedule, and use electrolyte drinks for events over about 90 minutes.

4. Tap water

In countries with regulated supply, tap water is among the most tested substances you consume. UK and EU regulations set limits on around 50 parameters, with water companies conducting millions of tests a year and public reporting. US supplies are regulated under the Safe Drinking Water Act.

Treatment typically involves screening, coagulation and flocculation, sedimentation, filtration, and disinfection.

Chlorine is the disinfectant that ended waterborne epidemic disease in industrialised countries, and its introduction is among the highest-impact public health interventions in history, alongside sanitation. The taste is the price. It dissipates if you leave water in an open jug in the fridge for an hour, or use a carbon filter.

Disinfection by-products (trihalomethanes, haloacetic acids) form when chlorine reacts with organic matter, and are regulated with limits. The risk trade-off is not close: the harms of untreated water vastly exceed the by-product risk.

Fluoridation. Around 0.7 to 1.0 mg/L, added in some supplies, naturally present in others. It reduces dental caries substantially, and the topical effect on enamel is now understood to matter more than ingestion. Excess during tooth development causes dental fluorosis, usually mild white flecking. Recent attention to possible neurodevelopmental effects concerns concentrations several times higher than those used in fluoridated supplies, mostly from regions with high natural fluoride, and the interpretation is contested (Chapter 16).

Hard versus soft water. Hardness is dissolved calcium and magnesium. It is harmless, it contributes a small amount of both minerals, and some ecological studies have associated hard water areas with slightly lower cardiovascular mortality. It causes limescale, which is a plumbing issue rather than a health one. Water softeners exchange calcium for sodium, which is worth knowing if you are on a sodium-restricted diet; most installations leave one unsoftened tap for drinking.

The genuine tap water risks:

• Lead, from lead service pipes and lead solder in houses built before roughly 1970 in the UK and 1986 in the US. There is no safe level of lead, and the effects on children's neurodevelopment are established (Chapter 92). The Flint, Michigan crisis from 2014 is the reference case, where a change in water source without corrosion control stripped protective scale from lead pipes. Practical steps if you have old plumbing: run the tap for 30 seconds to two minutes if the water has stood, use only cold water for drinking and cooking (hot water dissolves more lead and sits in a tank), and have the water tested. Many water companies test on request.
• Private wells and boreholes, which are not covered by public supply regulation and need independent testing for nitrate, bacteria, arsenic, and lead.
• Old buildings with copper and galvanised pipes.
• PFAS, covered in Chapter 93.

5. Bottled water

Types:

• Natural mineral water: from a protected underground source, bottled at source, cannot be treated except for limited processes, must have a stable mineral composition.
• Spring water: from an underground source, may be treated.
• Purified/table water: often treated municipal supply. A large share of bottled water sold globally is exactly this.

Is it safer than tap? In countries with regulated supply, no. Bottled water is regulated as a food and is generally tested less frequently than public supply. Recalls for bacterial and chemical contamination occur.

Mineral content varies enormously and is the one genuine reason to choose a specific brand. Some European mineral waters supply meaningful calcium (up to 500 mg/L) and magnesium; others are essentially nothing. Some are high in sodium, which matters on a restricted diet.

The costs: bottled water is typically 500 to 1,000 times the price of tap, and its environmental footprint is orders of magnitude higher, from plastic production, transport, and disposal. Microplastics are found in bottled water at higher concentrations than tap, and a 2024 study using improved detection methods found around 240,000 particles per litre in bottled water, mostly nanoplastics. What that means for health is genuinely unknown (Chapter 93).

Where bottled water is the right answer: places without safe supply, during boil-water notices, in travel to countries where supply is unreliable, and for people with specific immunocompromise where a sterile source is advised.

6. Filters, and what they actually do

Practical notes: carbon jug filters are for taste, mainly. Change cartridges on schedule, because an exhausted filter becomes a bacterial substrate. Keep filtered water refrigerated, because removing chlorine removes the residual disinfection.

7. Other drinks and hydration

All drinks hydrate, including tea and coffee. The idea that caffeinated drinks dehydrate you is wrong at ordinary intakes: the mild diuretic effect is more than offset by the water in the drink, and tolerance develops. Controlled studies find no difference in hydration status between coffee and water at typical consumption (Chapter 60).

Alcohol genuinely does dehydrate, by suppressing vasopressin, which is a real contributor to hangover (Chapter 61).

The Beverage Hydration Index, from a 2016 study by Maughan and colleagues, ranked drinks by how much fluid was retained four hours later. Milk and oral rehydration solution outperformed water, because their sodium, lactose, and protein slow gastric emptying and promote retention. This is why milk is a reasonable recovery drink and why ORS works.

Oral rehydration solution deserves its place here: the discovery that glucose-coupled sodium absorption continues even in an inflamed gut (Chapter 10) made ORS possible, and it has saved tens of millions of lives from diarrhoeal disease. A home version is roughly 6 level teaspoons of sugar and half a level teaspoon of salt in 1 litre of clean water, though commercial sachets are cheap, accurate, and preferable.

Water-rich foods contribute meaningfully: cucumber, lettuce, watermelon, tomatoes, soups, and yoghurt are all over 80 percent water.

8. The bottom line

• Eight glasses a day is not a scientific recommendation. Total needs are around 2 to 2.5 L a day from all sources, roughly a quarter of which comes from food, and you manufacture some yourself.
• Thirst is adequate for most healthy adults. It is not adequate for older people, infants, athletes in heat, or anyone ill, and dehydration in older adults is a common under-recognised cause of confusion and falls.
• Mild dehydration measurably impairs concentration and mood at 1 to 2 percent of body weight.
• Overhydration is real and can kill. In endurance events, drink to thirst rather than to a schedule.
• Tap water in a regulated supply is among the most tested things you consume and is not less safe than bottled. The hazard that matters is lead from old plumbing: run the tap, use cold water only, and get it tested.
• All drinks hydrate, including coffee and tea. Alcohol is the exception.

Sources and notes

Body water fractions, daily balance, and metabolic water production follow standard physiology texts. Reference intakes follow EFSA's 2010 Dietary Reference Values for water and the US Institute of Medicine's 2004 report. The absence of evidence for "eight glasses a day" is Valtin, American Journal of Physiology, 2002, which traced the claim to a 1945 Food and Nutrition Board recommendation whose qualifying sentence was dropped. Cognitive effects of mild dehydration follow Ganio et al. and Armstrong et al.'s controlled studies. Exercise-associated hyponatraemia follows Almond et al., New England Journal of Medicine, 2005, on the Boston Marathon, and the international consensus statements recommending drinking to thirst. Dehydration in older adults follows Hooper's Cochrane work on hydration assessment. Drinking water regulation follows the UK Water Supply (Water Quality) Regulations, the EU Drinking Water Directive, and the US Safe Drinking Water Act. Lead in water and the Flint crisis follow Hanna-Attisha et al., American Journal of Public Health, 2016. Fluoridation evidence follows the Cochrane review by Iheozor-Ejiofor et al. The Beverage Hydration Index is Maughan et al., American Journal of Clinical Nutrition, 2016. Microplastics in bottled water follows Qian et al., PNAS, 2024.

Open questions. Optimal fluid intake for health outcomes, as opposed to for avoiding dehydration, has never been established by trial. What microplastic exposure from bottled water does to health is entirely unknown.

👉 Next: coffee and tea.