He came home without it again. Second one this month, and the last one had a tracker clipped to it that cost more than the collar did.
So now you're holding the replacement, and there's a small slider sitting on the clasp with weight settings printed beside it. The support page tells you exactly what to do when a collar keeps coming off. Move the slider up a setting, so it takes more force to let go.
You are being asked to decide, by hand, how hard your cat has to pull before his collar releases him. Nobody in the aisle is going to walk you through that one, so here's the information you need to decide.
Are collars safe for cats?
Mostly yes. The numbers are a lot less frightening than the stories that get passed around.
The largest survey of the question asked more than 5,000 European owners about every cat they had ever had. Around 4% of cats had a collar incident that caused injury, and 0.6% died. Small numbers, but behind one is somebody's cat, so they shouldn't be waved off.
Now set them beside the risks he already runs every time he goes out the door. In that same survey, 70.5% of the cats hit by a car died. Vet clinic records tell the same story from the other side: collar injuries came to somewhere between 0.17% and 0.34% of cat cases, while road trauma ran to 4.2%.
The collar is what you're worried about. The road is what gets him.
Then why do vets keep warning about them?
Because some of the people who know cats best have looked at the same evidence and landed somewhere else.
Cats Protection, the largest cat charity in Britain, will not tell you to put one on. Their guidance points you at a microchip first and describes collar accidents that can cause strangulation, deep skin lacerations and jaw damage. An emergency vet group has written that its clinics have seen scores of cats hurt by their collars.
They've got good reason. Emergency clinics and rescue charities meet the cats it went wrong for. That's the whole population they deal with, and a lifetime of that shapes a view.
Closer to home, the BC SPCA takes the other line and recommends a breakaway collar, fitted properly, alongside permanent ID.
So the researchers say the risk is small and the welfare bodies say they've seen enough. The published injury figures run from under half a percent to twenty-six percent depending on who you ask.
Why do the numbers disagree so badly?
Four pieces of evidence, four very different answers.
The controlled trial. 538 pet cats were randomly assigned one of three collar types and followed for six months: a standard plastic buckle, a breakaway plastic buckle, and an elastic stretch collar. Just under 73% wore it for the full six months. Not one cat in the study was injured or killed by his collar. That's the reassuring end of the range, and it's the only experimental study anyone has run.
Look at the sample, though. About three quarters of those cats lived indoors only. An indoor cat has a lot fewer branches and fence posts to get hung up on.
The vet's-eye view. Researchers interviewed 107 veterinarians and worked out that a vet sees roughly one collar injury every 2.3 years of practice. At a single clinic, collar injuries came to 0.33% of 4,460 cat cases over three years.
The owner surveys. Sixteen percent of 393 self-selected owners reported a cat injured or killed in connection with a collar. In a British survey of 113 owners, 26% had taken a collar off after an injury or signs of distress.
The company survey. A pet tracker maker surveyed around 2,000 owners and turned up nine collar incidents between them. That's 0.47%, and none of the nine needed a vet.
Don't split the difference. They're not measuring the same thing.
The 16% counts owners rather than cats, over a whole lifetime of owning them instead of six months. Only about a third of those owners even had a cat in a collar. It also came from a survey people opted into, and the people with a story are the ones who click. The 0.47% came from a company that sells GPS collars. The vet figures only ever see the cats hurt badly enough that somebody booked an appointment.
All four are real numbers. None of them tells you the risk to your cat.
What actually goes wrong?
The 5,000-owner survey collected 652 collar problems. Sorted by what happened:
- Coat and skin trouble - 188 of the 652
- Caught on something, no injury - 286
- Caught, minor injury, no vet needed - 95
- Caught, injury that needed veterinary treatment - 30
- The cat died - 14
Those are shares of the problems people reported, not shares of cats. Most collar trouble is a collar hooking on a fence and a cat wriggling out of it in a temper.
The bad ones follow a pattern, and it's not the one you'd expect. They usually involve the whole front leg going through the collar rather than just the paw. That pulls the strap up into the armpit and leaves a wound there that will not close. The cats who died mostly strangled after the collar caught on a branch or a fence.
That's what the slider on the clasp is protecting him from.
Why does a monitoring collar make this harder?
Because it only works while it's on him. That's the whole proposition. It counts his steps, watches his sleep, in some cases reads his pulse, and every one of those needs an unbroken run of hours on his neck. Our guide to how pet health wearables work goes through what the sensors are doing in there.
A cat safety collar is built to do the opposite. Its job is to let go.
The veterinary literature says this out loud. A peer-reviewed review notes that a collar which releases easily is what's recommended for unsupervised use. That same feature can stop the collar doing the job it was put on for.
To its credit, one tracker company spells this out on its own support page. If the collar breaks away the tracker goes with it, because the mechanism puts the cat's safety ahead of keeping the device attached.
The same page then tells you a collar releasing during play is normal. If that happens more than once or twice a week, tighten the setting.
And this is not a niche problem any more. A 2024 survey asked central European owners who had collared a cat for the first time in the previous two years what was hanging off it. A tracker was on nearly one collar in five. More common than a bell.
Is elastic as safe as breakaway?
Elastic collars were designed on a sensible idea: give the collar enough stretch and a snagged cat can wriggle out. Cats Protection now advises against them, and the reason is what the cat does with that stretch. He gets a paw up to push the thing off, the elastic gives, and now the leg is through it too. Front leg through the collar is the injury that produces the armpit wound.
The research points the same way. In that 538-cat trial the elastic collars were the only ones that snagged on objects, 2.3% against zero for the other two types. When the 107 vets could remember what kind of collar was involved in an injury, elasticated collars accounted for 39% of the cases.
Some trackers ship with an elastic collar and call it anti-strangulation. It's the older design, and the charity that runs the rehoming centres has moved off it.
Go for a quick-release buckle, the kind that snaps open under load.
Does a breakaway collar solve it?
Not completely, and that's the part the packaging leaves out.
The review that recommends breakaway collars also concludes that safety collars do not fully prevent severe incidents. Rare. Not zero.
And in the 5,000-owner survey the researchers could not show breakaway collars made the incidents end any better. How much force a breakaway needs also varies between makers, so two collars both sold as safety collars can behave nothing alike.
Use one anyway. The recommendation stands and it's the right one. In that same 2024 survey, about one newly fitted cat collar in seven had no breakaway at all. That's a lot of cats in the design the whole field moved away from.
Just don't treat it as a guarantee, or let it distract you from the thing that actually keeps him safe.
That thing is fit. Two fingers between the collar and his neck, no more and no less. Too loose and a leg gets through, too tight and it chafes. Look the strap over for wear while you're in there, because a frayed collar will fail somewhere other than the clasp.
Then check it again ten minutes later. A cat who has just been collared hunches and flattens himself. The fit you measured on a tense cat is not the fit he'll be wearing once he relaxes and stalks off.
And watch him properly for the first stretch after any new collar goes on. In that six-month trial, ten of the 538 cats got the collar caught in the mouth, and six of those ten managed it inside the first five days. The trouble shows up early.
Does he have to like it?
He does, more than the research on any of this would suggest.
The strongest predictors of whether a collar was still on after six months were the owner's own expectations, and how many times the collar got put back on. Not the cat's age. Not the collar type. The owner.
Which means the first fortnight is the part that decides it. Put the collar on and immediately do something good: dinner, treats, the brush he likes. Take it off before he starts working at it, and build up.
Watch for the signs the researchers use to pull a cat out of a study: scratching hard at his neck, over-grooming, going off his food. Those mean back up a step, not push on.
What about the harm nobody talks about?
Coat and skin problems were the biggest reported category after simple snagging, and they get almost no airtime. Matting under the strap, hair loss, a red patch, a cat scratching at his neck more than he used to.
On a monitoring collar that has a second edge. Scratching is a behaviour these devices count. A collar that makes him itch is a collar reporting a problem it caused, and you'll watch that line climb with no way to tell which it is.
Then there's whatever you hang off it. A bell or a disc or a tracker pod is an object, and an object can wedge where a flat collar wouldn't. Cats Protection names fence panels as the classic trap. A cat can hook a claw in a bell too.
The veterinary guidance is to keep large hooks, rings and loops off a cat's collar, and a tracker is a bigger object than a name tag.
Of the injury cases those 107 vets could put a collar type to, flea collars were the biggest group at 42%. More than elastic. Far more than ordinary buckle collars. Cats Protection also rates flea collars poorly at the actual job of killing fleas. If he's wearing one out of habit, that's a conversation for your next vet visit.
So what should you actually do?
Your best setup is a tracker that clips onto a collar you chose yourself. That way the safety decision is yours and the device just goes along for the ride. One maker says so outright: its cat tracker ships with a safety collar, and you're free to use a different one. Pick a quick-release buckle, fit it at two fingers, and you get the monitoring without giving up the safety release.
Unfortunately, some devices don't let you. Integrated collars build the tracker into their own strap, and at least one maker tells buyers that swapping in a smaller collar interferes with how the device works. If that's the product you want, you're accepting their release setting, so find out what it is before you buy rather than after.
Whatever you pick, back the ID up with permanent identification. A collar built to come off takes his name tag with it when it goes. Nearly every microchip in that six-month trial was still reading correctly at the end, which no collar tag can promise. In Canada a registered tattoo counts too.
And here's the step people skip: register it, then keep it current. A chip nobody can trace to a phone number you still answer is a number under his skin and nothing else. Moved recently or changed your number? Log in and update it.
If he's indoors and it's really his health you're watching, ask whether it needs to go round his neck at all. A litter tray that weighs him every visit, or a bowl that logs which cat drank, gets you the early warning with none of this on the table. We go through those in monitoring a cat without a collar.
What about putting the tracker on a harness?
Everybody lands here eventually. Don't do it.
A harness spreads the load across his chest instead of his throat, and it holds a device just fine. Some are sold with a pocket for exactly that. Trouble is, a harness has no breakaway. That's the point of a harness.
So a harnessed cat who snags on a fence has nothing to release him. The advice from behaviour guides and from the tracker companies themselves is the same: never leave a harness on an unsupervised cat, indoors or out.
A harness is for walks you're on the other end of. Anything else and it stays off him.
What if he's already sick?
Then the calculation shifts, and it shifts toward the collar coming off more easily rather than less. A cat managing kidney disease or IBD is a cat you want data on. He is also, quite often, a cat who has lost weight since you last checked the fit.
A collar that fit him in March may be riding loose by July, and loose is the setting that lets a leg through. If his weight is moving, the fit check moves with it.
Our guide to monitoring a pet with a chronic condition has the signs no collar can pick up: the ones that mean stop reading graphs and start phoning.
So the next time he strolls in with a bare neck and no explanation, that isn't the collar failing you. That's the bit of plastic on the clasp doing the exact job you bought it to do.
Has your cat lost more collars than you'd like to admit to? Tell us in the comments, we'd like to know what the going rate is.
Part of our guides to how the technology works. Start with how pet health wearables work for what the sensors are reading, and smart collar battery life for how often the thing has to come off anyway.