A woman in her late twenties is sitting on the edge of an exam table in an urgent care twenty-eight hours after a Saturday night she liked. Her throat is sore. She has a low headache that won't go, and a small specific patch of vision in her left eye that has been doing something strange since lunch. She is here because a friend made her come. The intake nurse asks if she has been assaulted. She says no, twice, because the question makes her angry and the honest version of the answer is complicated. Eventually she says the word choked, which is not the word for what she means, and the nurse, who has seen this before, writes it down without flinching and orders the imaging.

That is the framing this piece needs. Not the morning-show please don't, not the kink-blog here is how to do it safely. The exam table, twenty-eight hours later, with someone who likes her partner and is also having a small stroke. That happens. The literature on why it happens has moved a lot in the last two years, and almost none of the moving has reached the people doing the thing.

The honest opening

Sexual choking is one of the most prevalent intimate acts among adults under thirty-five, and one of the most poorly understood by everyone in the room when it's happening. Roughly half of young Australians and a comparable share of US college students report being choked during sex, women's exposure running highest. The cultural script says it's hot, normal, and (with the right precautions) safe. Two of those three are true.

The honest opener for an essay like this used to be we don't really know how risky it is. Between 2023 and 2025 the picture sharpened enough that we know it's worse than the field thought, in specific ways that aren't fixed by any of the "safer" rules the kink internet teaches. Pretending otherwise gets people killed slowly. The aim here is not to talk anyone out of anything. It is to make sure the picture in your head matches the picture in the literature, because right now, for most readers, it really doesn't.

What the 2024–2025 research actually shows

Three things, all of them inconveniently clear.

One: there is no safe pressure threshold. The Melbourne research group (Douglas, Sharman and colleagues) interviewed hundreds of young adults about how they keep choking "safe." The most common belief was that light pressure to the sides of the neck is fine. The same paper notes, dryly, that the difference between fatal and non-fatal strangulation involves very little variance in applied force, that unconsciousness can land between eight and eighteen seconds, and that there is no reliable way to know which side of that line you're on while you're on it. The "press the sides, not the front" rule comes from a misreading of where the carotids are versus where the trachea is. Both are vulnerable. Compressing the carotids is, in fact, how people lose consciousness, which is the mechanism doing the damage. (Douglas et al., 2025.)

Two: a single episode produces measurable brain-injury biomarkers. This is the part that has not made it into the mainstream conversation yet. In a 2025 randomized crossover study, women who had a single choking-involved sexual encounter showed an acute rise in neurofilament-light (NfL) compared to a non-choking control encounter. NfL is the standard blood marker for axonal injury; it goes up after concussion. It is not a marker that lies. Other work has found chronically elevated S100B and structural / functional brain differences (working memory, fMRI changes) in women who are choked frequently, even when none of the individual events ever produced loss of consciousness. (Acute biomarker study, Frontiers 2025.)

If you took the same data and labeled it amateur boxing, the public-health response would have been louder, faster, and unanimous.

Three: the most dangerous injuries are delayed. ER physicians have known this for two decades and the public still mostly doesn't. The thing that kills people from a strangulation event is rarely the event itself. It's the carotid artery dissection that develops over the next hours or days, throws a clot, and produces a stroke at a moment when the person having the stroke has stopped associating it with Saturday. Case series report symptomatic dissections days to weeks after the original compression, in patients who walked away feeling tired and bruised.

That is the woman on the exam table in paragraph one. Her stroke is twenty-eight hours late. Her partner is at home, scrolling, with no idea.

Why "but I'm doing it carefully" doesn't help as much as you think

Almost every "safer choking" guide teaches some version of the same logic: low pressure, short duration, sides not front, communication, consent, sobriety, a tap-out signal. Good ideas in general. None of them addresses the actual mechanism of injury.

The mechanism is that you're compressing the blood supply to a brain and, depending on hand position, the airway. The brain notices immediately. The vasculature notices on its own schedule, often after everyone has gone home. A short, light, consensual, sober, side-of-neck compression is still a compression of the carotid system. Less likely to kill you on the spot than the long, hard, drunken version. Not zero risk. There is no version of this that is zero risk, and the field has stopped pretending there is.

This is the part it's tempting to soften, because a reader doing this in good faith with a long-term partner is going to bristle. Bristle anyway. The 2024–2025 literature is unusually unified, and honest harm reduction does not lie to its audience to seem cool.

What actually reduces risk versus what just feels safer

There is still a difference between less likely to kill you tonight and equally likely to give you a delayed stroke. Worth being precise about which is which.

Things that genuinely reduce acute risk (the same-night kind):

  • Not being drunk or high. Substances impair the chokee's ability to signal in trouble and the choker's ability to read a face going wrong.
  • Hands, not ligature. A hand can be released instantly; a scarf, belt, or cord can knot or jam. The autoerotic deaths in coroner reports are almost always ligature, and almost always alone. Don't do this alone. Ever.
  • A pre-agreed tap-out that doesn't require speech. Speech is the first thing to go. Two taps anywhere reachable, both partners know what it means, you stop and check.
  • Stopping at the first sign of anything off. Color change, eyes losing focus, twitching, going limp, voice change after. Anything off, you're done for the night.
  • Not doing it on someone you don't know well, in a hookup, or at a play party. The skill is not in the choking; it's in reading the body of someone you know in detail.
  • Position. If the choker is on top, the chokee can't easily get away if something goes wrong. Configurations that let either person pull back fast are less bad.

Things that feel safer and don't really change the medical risk:

  • "Light" pressure. There is no calibrated light. The amount of force that produces unconsciousness in eight seconds is much less than people think.
  • "Sides of the neck only." This is where the carotids are. Compressing them is the dangerous part. The "sides not front" rule is anatomy misunderstood until it sounds like physiology.
  • "Just for a few seconds." Brain hypoxia is cumulative across an evening and across a relationship. A single short episode is enough to elevate injury biomarkers in a lab.
  • Consent. Necessary; not protective. Your carotid artery does not check whether you said yes before it tears.
  • Trust and communication. They make the event nicer. They do not make it medically safer.
  • Knowing CPR. CPR works on hearts that stop. It does not undo a stroke that develops over the next two days.

You'll notice that the second list contains most of what the kink internet calls "safe choking." That is the gap this essay exists to close.

The warning signs that mean go to the ER, not to bed

If you take only one thing from this post, take this list. Memorize it, tell anyone you do this with, tape it inside a cabinet.

Any of the following, in the hours or days after a choking episode, means urgent care or an ER, and means say the word strangulation, because the imaging protocol is different:

  • Any change in voice (hoarseness, weakness, the sense your voice is "not landing").
  • Difficulty or pain swallowing, especially with liquids.
  • Persistent neck pain or unusual neck swelling, especially asymmetric.
  • Any visual disturbance: blurring, a patch missing, double vision, transient blackout in one eye.
  • One-sided weakness or numbness, even briefly.
  • Slurred speech, word-finding trouble, sudden confusion.
  • Severe or new-pattern headache, especially with neck stiffness.
  • Pinpoint red dots (petechiae) on the face, scalp, or in the eyes.
  • Loss of consciousness at any point during the event, even for a second.
  • Loss of bladder or bowel control during the event.

The thing to flag for the clinician is not embarrassment. It is I had pressure applied to my neck. They've heard it. They have a protocol. The reason it matters is that the standard ER workup for "headache and sore throat" does not include the CT angiogram that catches a dissecting carotid before it throws a clot. Saying the word changes the workup. Not saying it can change the outcome.

Pregnant readers: any neck compression during pregnancy carries miscarriage risk on top of everything above. Take any of the warning signs to a clinician even faster.

The aftercare clause this changes

For anyone with a regular aftercare routine written down, choking belongs in its own category. Most aftercare scripts cover the first thirty minutes, the next twelve hours, the morning. The relevant window for strangulation injuries extends to three to seven days. The clause needs a check the symptom list item at the 24-hour and 72-hour mark, and an agreement that flagging any of those symptoms is not drama, not killing the mood, not ruining anything. It is the deal.

The pre-scene conversation needs the same upgrade. The usual negotiation script covers what's on and off the table; if choking is on the table, the addition is we both agree to take the warning-sign list seriously this week and to go to a doctor without arguing about it. Five seconds of negotiation, real protection against the failure mode that actually kills people.

This is the only place sober consent does meaningful work in this domain. Not as protection from injury, which it isn't. As protection from the more common failure: somebody noticing a symptom, not wanting to overreact, going to bed, and not waking up correctly.

The cultural shape of the conversation

One last thing, because the data sits inside it. The reason so many adults under thirty-five are doing this is not that they're reckless. The practice was normalized at scale by porn over roughly the last fifteen years and folded into the standard heterosexual repertoire without anyone running the safety review. The "safer ways" advice that filled in afterwards came from harm-reduction-minded kink educators doing their best with what was known at the time. They weren't wrong to try. The information they were working with has been overtaken by better data, and the better data is grim.

The just stop doing it messaging has not worked, either. It gets shouted at people who have already integrated the act into their sex life, often with partners they love. Telling them they're doing something dangerous and stupid does not get them to stop. It gets them to stop listening.

Better data calls for something quieter. Here is what we now know. Here are the rules that work and the rules that don't. Here is what to watch for afterwards, and here is what to say when you watch for it. Adults can take that and decide what to do with it. Most of the time, what they'll do is keep doing the thing, with more attention to what matters and less faith in what doesn't. That is what harm reduction is when it isn't performing.

The thing to leave with

The single sentence to carry out of this piece, if you carry only one: the worst injuries are delayed, and the window for catching them depends on you saying the word. Not on pressure, not on duration, not on consent, not on trust. On the willingness, twenty-four hours later, to put together a sore throat and a strange headache and a partner who liked the scene, and to call it what it was at the front desk.

The lights go down on Saturday. The bill, if there is one, arrives some weekday afternoon. Knowing when to open it is the entire game.