Cold Water
Four separate things happen to a body in cold water, and the one everybody can name arrives last.

In a swimming flume at the University of Portsmouth, ten volunteers were asked to swim breaststroke against a current for ninety minutes, or until they could no longer keep their feet off the floor of the tank. They did it three times over, in water held at 25 degrees Celsius, then 18, then 10. All ten finished the warm swim and eight finished at 18, but at 10 degrees only five lasted the distance: one failed outright at sixty-one minutes, and four more were lifted out as they closed on it. Their deep body temperatures then read 35 degrees, the outermost edge of the definition of hypothermia and nothing like a lethal one. Their blood was barely cold, and their arms and legs had quit anyway.
Michael Tipton, Clare Eglin, Mikael Gennser and Frank Golden published that in the Lancet in August 1999, and their conclusion pointed away from the obvious culprit: treatment of immersion victims should be aimed at the consequences of near-drowning rather than at severe hypothermia. Golden and Tipton, at Haslar and later Portsmouth, had spent decades building the framework the experiment sits inside. Cold water does four distinct things to a person, in sequence, and each one can kill on its own: cold shock, swim failure, hypothermia, and collapse at the moment of rescue. Cold here begins around 15 degrees Celsius, warmer than most people's idea of the word and warmer than the sea off almost any temperate coast for most of the year.
The first of the four arrives inside thirty seconds and is finished in two or three minutes. Skin temperature drops away and the body answers with an involuntary gasp, then hyperventilation at six to ten times normal, while breath-hold time falls to a quarter of its usual span. If the head is under the surface when the gasp comes, water goes into the lungs, and no amount of swimming ability prevents it. The vessels near the skin clamp shut, the heart rate climbs, blood pressure surges, and a Danish study in 2007 found blood flow to the brain falling sharply on sudden immersion, one route by which a strong swimmer blacks out in the first minute. Tipton and Michael Shattock described a further mechanism in 2012 and called it autonomic conflict: the sympathetic nerves carry an order from the chilled torso to speed the heart up, the chilled face with the breath held triggers the diving reflex and sends the opposite order down the parasympathetic nerves, and a heart receiving both at once can lose its rhythm.
Everything the survivor is tempted to do in that first minute makes it worse. The drill the Royal National Lifeboat Institution now teaches, on the strength of the Portsmouth work, is to do almost nothing: fight the urge to thrash, lean back, spread the arms and legs, and float for about ninety seconds until the breathing comes back under command. Only then swim, or shout, or reach for something solid. Float to Live is the whole of it, and it is aimed at the stage that produces most of the bodies.
The second stage is the one the flume caught. As water strips heat from the arms and legs, the superficial nerves and muscles cool, and coordination fails before strength does. The stroke shortens, the rate rises to make up for it, and the swimmer's body angle climbs steadily toward the vertical, the tell a lifeguard on shore can read at a distance. Golden and P. T. Hardcastle had reported the effect in 1982; the flume measured it. In water near freezing the progression can run out inside ten minutes, and somebody with no flotation drowns at the end of it carrying a core temperature a coroner would record as normal. That is much of why cold-water deaths appear in the statistics as drownings.

Hypothermia proper is the third stage and the slowest. It starts when deep body temperature falls under 35 degrees, which in water between 5 and 10 degrees takes one to two hours. Shivering comes first, then clumsiness, then failing judgment; lower still, the shivering stops and consciousness goes with it. This is the stage that owns the public imagination and the one a person in the water has the most time to survive. The Canadian Safe Boating Council teaches the first three as a mnemonic, one-ten-one: roughly a minute to get the breathing back, ten minutes of useful movement, an hour before the cold reaches the core. The numbers are rounded hard, and the first two do most of the work.
The fourth stage sits outside the mnemonic and catches out the rescuers. Golden, G. R. Hervey and Tipton set it down in the Journal of the Royal Naval Medical Service in 1991 as circum-rescue collapse, having collected survivors who were talking when the line reached them and dead soon after. Water standing around a body presses on the legs and abdomen and holds blood where the heart can use it. Lift someone vertically out of that, in a helicopter strop or over the side of a boat, and the support vanishes in seconds; the authors recorded aortic and central venous pressure falling during exactly that manoeuvre, leaving a heart already chilled and unstable to absorb the loss. The instruction that follows is counterintuitive enough that it has to be taught: someone who has been in cold water a long time should come out lying down.

All of it sits in the operating instructions of the world's lifesaving services. The International Life Saving Federation's position statement on cold water, approved in 2003 and revised in 2016, puts two of the consequences in language meant to be followed under pressure.
3All immersion victims should be quickly rescued. Those victims who have been in cold water for a considerable time, and whose airways are not being threatened by wave splash, should be rescued in a near horizontal attitude, if possible, to prevent a potentially adverse fall in blood pressure. Those whose airways are threatened should be rescued by the quickest method regardless of body attitude.
6In those who are apparently dead, avoid a hasty diagnosis. The maxim “no one is dead until they are warm and dead” is appropriate provided you are not too remote from medical support. Follow resuscitation protocols and arrange rapid transport to hospital. People, especially small children, who have been submerged for up to an hour in ice-cold water have been successfully resuscitated.
Medical Position Statement 12, Cold Water Immersion, statements 3 and 6.
Volunteers dropped into ice water on camera, with the four stages narrated as they happen. Watch what the hands do around the eight-minute mark of immersion: the subjects are still talking and still lucid, and they can no longer work a buckle or hold a rope. That gap between a clear head and a useless hand is the thing the mnemonic is trying to warn people about.
On the afternoon of 11 February 2011 a school dragon boat capsized in Præstø Fjord in southern Denmark, air at 4 degrees and water at 2. Fifteen people went in, median age sixteen, and one drowned in the capsize. Seven of the rest were pulled out in circulatory arrest, pulseless, pupils fixed and dilated, at a median deep body temperature of 18.4 degrees, the coldest at 15.5. All seven were put on extracorporeal circuits within four hours and rewarmed from the inside outward, and all seven lived, six with mild to moderate cognitive damage and one with severe. The maxim the lifesavers print gets passed around as sea-going folklore, and Præstø Fjord is why it appears in the protocols instead.