The denominator first. Controlled human trials of water deprivation to death: 0. None were found, and none can ethically be run. Every measured range below therefore comes from one of three places: end-of-life care where patients chose to stop drinking, athletic and occupational physiology where body-mass loss was measured directly, or single clinical cases. Any figure presented as a clean "humans survive N days without water" is not traceable to a controlled measurement, and the widely repeated "three days" is a survival-folklore heuristic, not a measured result.
B1. The only large human dataset on complete fluid abstention: VSED
| Field | Content |
|---|---|
| Grade | B+ (human cohort via physician survey, 99 cases) |
| Population | Primary care patients in the Netherlands who hastened death by voluntarily stopping eating and drinking |
| Identifier | DOI 10.1370/afm.1814. PMID not shown on the page I opened. |
| Citation | Bolt EE, Hagens M, Willems D, Onwuteaka-Philipsen BD. "Primary Care Patients Hastening Death by Voluntarily Stopping Eating and Drinking." The Annals of Family Medicine 2015;13(5):421-428. |
MEASURED:
- Median time from the start of stopping eating and drinking until death: 7 days
- Sample for that measure: 86 patients. Ten responding physicians did not remember and 3 responses were missing (13.1%).
- Total patients who hastened death by VSED: 99
- Life expectancy at the time of VSED, n = 99: under 1 week 2%; 1 to 4 weeks 32%; 1 to 12 months 41%; over 1 year 26%
- Symptoms in the last 3 days before death, n = 96: 36% had symptoms, 42% had none, 10% had none but received palliative sedation, 11% did not know
- "Duration too long" was cited as the reason for an incomplete match to the patient's wish in 11% of patients
READ. The range around the 7-day median is not reported in that paper. The only verbatim sentence on the page carrying the survival figure is a figure caption: > "Cumulative survival curve for duration until death after start of VSED."
Limitation stated plainly and it is large. These were patients at the end of life, the majority already with a life expectancy under 12 months, receiving palliative care and in many cases mouth care. A median of 7 days in that population is not a healthy-adult survival figure and must never be presented as one. It is the closest measured human dataset on complete fluid abstention that exists, and it is drawn from a population already dying.
B2. What the physiology does across the span, by percentage of body mass lost
| Field | Content |
|---|---|
| Grade | A (professional body position statement with graded recommendations) |
| Population | Physically active people. Thresholds are stated as general human physiology. |
| Identifier | DOI 10.4085/1062-6050-52.9.02 |
| Citation | McDermott BP, Anderson SA, Armstrong LE, Casa DJ, Cheuvront SN, Cooper L, Kenney WL, O'Connor FG, Roberts WO. "National Athletic Trainers' Association Position Statement: Fluid Replacement for the Physically Active." Journal of Athletic Training 2017;52(9):877-895. |
MEASURED thresholds, read off the statement:
| Body mass lost | What happens | Verbatim where available |
|---|---|---|
| above 1% | Thermoregulation compromised | "Thermoregulation is compromised by hypohydration greater than 1%" (Rec. 4, SOR B) |
| about 2% | Thirst rises | "Thirst typically increases at about 2% hypohydration" |
| 2% or more | Endurance performance reduced | "performance was consistently reduced when hypohydration met or exceeded 2% body mass loss" |
| above 3% | Exertional heat illness risk rises | "the risk of exertional heat illness increases at moderate to severe levels (greater than 3%)" (Rec. 4, SOR B) |
| 3% to 4% | Anaerobic performance decrements in most studies | |
| above 5% | Consistently impaired performance, extreme thirst | "In athletes, a deficit greater than 5% is consistently associated with impaired performance, extreme thirst" |
| above 10% | Severe clinical category | Clinical bands given as mild about 1% to 5%, moderate about 5% to 10%, severe above 10% |
Also MEASURED from that statement: core temperature rises roughly 0.15°C to 0.20°C for every 1% of body weight lost during activity, described as "a generally linear increase in core temperature." Hypohydration "leads to increased cardiovascular stress during exercise" (Rec. 3, SOR A).
READ, verbatim, Recommendation 2, SOR B: > "Both severe clinical hypohydration and hyperhydration can degrade athletic performance and are potentially fatal."
Instrument caveat recorded. The converted text of that PDF was garbled in two places: at the clinical severity bands and at the core-temperature coefficient. Those two figures are flagged as needing confirmation against the original PDF before they are published on any customer surface. Every other figure in the table above came through clean.
During exercise, thirst does not protect you. READ from the same statement: at body mass losses approaching 2% at rest thirst increases, but during exercise thirst signals at similar levels "appear to be absent or ignored."
B3. What withholding fluid actually costs, measured as a requirement
| Field | Content |
|---|---|
| Grade | B+ (prospective longitudinal observational, 22 workers, 66 subject-days) |
| Population | Male expatriate construction workers in the UAE: carpenters, steel fixers, general labourers |
| Identifier | DOI 10.1186/1745-6673-3-21. PMID not shown on the page I opened. |
| Citation | Bates GP, Schneider J. "Hydration status and physiological workload of UAE construction workers: A prospective longitudinal observational study." Journal of Occupational Medicine and Toxicology 2008;3:21. |
MEASURED:
- 22 male workers, three groups, each monitored over 3 consecutive 12-hour shifts, giving 66 subject-day records over 9 study days
- Average fluid intake 5.44 L per 12-hour shift
- Average urine specific gravity over the 3 days 1.012
- WBGT consistently exceeded 27.5°C, the recommended limit for moderate work, especially at midday
- Aural temperature showed no change between day 1 and day 3 or between shift start and finish
- Average working heart rate showed no significant change across the shift or between days
- Fluid intake correlated positively with heart rate (r = 0.719) and negatively with urine SG (r = −0.519)
READ, verbatim: > "Most subjects commenced work euhydrated and maintained this status over a 12-hour shift."
> "people can work, without adverse physiological effects, in hot conditions if they are provided with the appropriate fluids"
INFERRED (one hop, marked as inference). Hop 1: men doing manual labour in WBGT above 27.5°C drank 5.44 L per 12-hour shift to stay euhydrated (MEASURED). A dry fast sets that intake to zero for the same hours. No study has measured what happens to those workers at zero intake, so the size of that deficit is not measured and is not stated as a number here. The requirement is measured. The consequence of removing it is inference.
Study's own limitation, READ: the authors acknowledge that 22 subjects is too few to establish safety conclusively, and they note an internal date inconsistency (Results describe May 2006, the Environmental Assessment section refers to a 10-day assessment in June).
dryfasting.com has been published since 2014. This evidence record was compiled on 22 November 2026. Every row names the fasting modality the study actually ran, the system it was measured in and the grade. Where no source carries a DOI or PMID the row says so rather than supplying one. Nothing is sold on this site.