Do you need to finish the antibiotics?

Transcript

Welcome to the sofidex Health Drive podcast series. Today we are talking about antibiotics, and about the one rule almost everybody can recite: whatever happens, you finish the course. Maybe you are on day three of a five-day box right now, and you feel fine. The cough has eased, the fever has gone, and there are still tablets in the packet. Somewhere in the back of your head a voice is saying that if you stop now, you are breeding superbugs.

And then somebody tells you that rule was scrapped years ago, and that you can stop whenever you feel better. That is not quite true either, and it is a risky thing to half-hear. So we want to ask a better question than finish it or stop it. Who decides how long a course should be, what that decision rests on, and why the answer keeps getting shorter.

We will go through the reason most of us were given for finishing, and why it does not hold up. We will look at the trials that tested shorter courses head to head, and at how the courses written on prescriptions have changed because of them. And we will get to the infections where the long course still matters, and to the half-empty packet in the bathroom cabinet.

If you are taking antibiotics right now, nothing in this episode is a reason to stop them on your own. The person to ask is whoever prescribed them, or the pharmacist who handed them over. One thing before we start. This is health information, not medical advice, and nothing in it is a substitute for talking to your own doctor about your own treatment. What we can tell you is where all of it came from — 13 published sources, all listed with the episode. I want to do this one through a single person, if that's alright. Not a real patient. Somebody we're making up, so we can follow her the whole way through. Call her Ruth. She's sixty-one. Last week she had a cough that wouldn't shift, and a temperature. Her GP listened to her chest and said it was a mild pneumonia. Amoxicillin, three times a day, five days. And this morning, day three, she woke up feeling fine.

And that's the moment I want to spend this episode on. Ruth is standing in her kitchen with two days of tablets left. And two voices in her head. One is her mother's. Her mother was a ward nurse, and the one rule she drilled into Ruth was that you finish the course. Every tablet. Or you make the bugs stronger. The other voice is her son's. He sent her a link last night saying that rule was scrapped years ago. So she wants to know which of them is right. And the short answer is neither of them. Not quite. Which is the interesting part.

Let's start with the mother, then. Because I think most of us were raised on exactly that rule. Finish the course, or you breed superbugs. It's almost a moral thing. Where does it actually come from?

It comes from a picture that feels completely obvious. You've got an infection. Millions of bacteria. The antibiotic kills the weak ones first. So if you stop early, the theory goes, the tough ones survive. They multiply. And now you've got an infection the drug can't touch. It's a lovely story. It's got a villain. It's got a moral. And there's a clear instruction at the end. It's the reason generations of people have swallowed the last two tablets while feeling perfectly well. But it's a story about evolution inside one infection. And it skips over where resistance mostly shows up. Which isn't where Ruth would expect.

Okay, so where does it show up? Because I'd have said exactly where she thinks. In the infection she's treating.

In everything else she's carrying. That's the bit the picture leaves out. Ruth's antibiotic doesn't know it's meant for her chest. It goes everywhere. Her gut. Her skin. The back of her nose. Her bladder. All of those are full of bacteria that were never the problem. And every day she takes the tablets, those bacteria are being sorted. The ones that can survive the drug do well. The ones that can't, don't. So the longer the course, the longer that sorting runs. Which turns the mother's rule upside down. If resistance is about exposure, extra days aren't protection. They're more exposure.

That's a big claim, though. It sounds right when you say it. But is that just a nicer story, or has anybody actually measured it in real people?

It's been measured. And it's worth being precise about what was measured. There's a systematic review in the BMJ from 2010 that pulled together studies of people given antibiotics by their GP. And they looked at whether those people were carrying resistant bacteria afterwards. They were. So that's the first finding. Sobering on its own. But the part that matters for Ruth is this. Studies reporting the quantity of antibiotic prescribed found that longer duration and multiple courses were associated with higher rates of resistance. Longer. Not shorter. And the effect was greatest in the month immediately after treatment, but could persist for up to 12 months.

Associated, though. You chose that word quite carefully. Which makes me think it's not the whole story, and that you're about to tell me why.

I did choose it. Because those were mostly observational studies. People who get longer courses might be sicker, or older, or in hospital more. So it's a link, not proof. But there's a trial that gets closer. It was done in the Dominican Republic, about twenty-five years ago. Nearly eight hundred small children with respiratory infections, randomised to one of two amoxicillin courses. And at the day 28 visit, the share carrying resistant pneumococcus was 24 per cent in the short-course, high-dose group, compared with 32 per cent in the standard-course group. Now, there's a catch. The short course was also a higher dose. So two things changed at once. Which means it can't tell you which of them did the work. But nothing in it says the longer course was the safer one.

So the mother's reason doesn't hold. Does that mean the son's right? Ruth just stops when she feels better, and puts the box away?

No. And this is where I'd slow Ruth right down. Because the son's version is the one that can actually hurt her. What the evidence tested was never "stop when you feel like it". It tested shorter courses. Fixed ones. Decided by a doctor, with a defined end. And sometimes with a doctor checking the patient first. But feeling fine is a poor test of whether an infection's gone. The fever drops. The energy comes back. And the bacteria can still be there. So the son's link has taken a real change and turned it into an instruction nobody tested. That's what I'd want Ruth to hear. The rule that changed is about how long courses should be. It isn't permission to invent your own.

Okay. So tell me about the change itself. When you say the evidence tested shorter courses, what did that actually look like? Who was in these trials?

The cleanest version is a trial from the Netherlands, published in 2006. Adults in hospital with pneumonia. Mild to moderately severe. And everybody got three days of amoxicillin into a vein. Then the ones who'd substantially improved were randomised. Some got five more days of amoxicillin tablets. The rest got identical-looking dummy tablets. Nobody knew which. And the result was that discontinuing amoxicillin treatment after three days wasn't inferior to discontinuing it after eight days. At day 10, the clinical success rate was 93 per cent in both groups. Not similar. The same. And look how carefully that was set up. These weren't people deciding for themselves. Doctors decided they'd improved enough. Then chance decided the rest.

And did the people who stopped early do worse in any way at all? Because I'd expect there to be some catch hiding in there somewhere.

Not that they measured. Both groups had similar resolution of symptoms. And here's a twist on it. Adverse events were reported by 21 per cent of the group that kept taking amoxicillin, against 11 per cent of the group on dummy tablets. Now — careful. That's a small trial. And that gap could have been chance. So I wouldn't lean on it. But it points the direction you'd expect. Every extra day of an antibiotic is another day for it to upset your stomach. Which is the bit the old rule never counted. It treated the extra days as free. They aren't.

That's one trial, though. Dutch hospitals, nearly twenty years ago, and a fairly small number of people. Could it just have been a fluke?

It's held up. Repeatedly. And in places that matter for listeners. There's a large British trial in children called CAP-IT. 824 children with pneumonia, sent home from hospital on amoxicillin. Some got three days. Some got seven. And the thing they measured was whether a child needed antibiotics again within four weeks. That happened in 12.5 per cent with 3-day treatment versus 12.5 per cent with 7-day treatment. Identical again. But here's what I like about it. It's honest. And the children on three days coughed for a little longer. The median cough lasted 12 days on the short course against 10 days on the long one. So shorter doesn't cost nothing. The cost was a couple of days of cough. And the saving was four days of antibiotics.

So it's not just pneumonia? Because I'd assumed pneumonia was the serious one. The one where you'd want the long course most, if anywhere.

That's a common assumption. And no, it's not just pneumonia. Somebody did the sensible thing and gathered up the systematic reviews themselves. Across the infections people mostly get treated for at home. They included 9 reviews. Sore throats. Ear infections. Sinus infections. Bladder and kidney infections. Chest infections. In adults and in children. And for infection after infection, short and long courses cured people about equally. Their conclusion was that short course antibiotics are as effective as longer courses for most common infections managed in ambulatory care. Ambulatory just means outside hospital. Most. Not all. We'll come back to the exceptions, because they matter. But that's the centre of gravity now. The burden of proof flipped. It used to sit on the short course. Now it sits on the long one.

Hang on, though. If all of this has been known for years, why is Ruth's mother's rule still the one everybody can recite?

Partly because it's a good rule to remember. Simple. Moral. Easy to say at a pharmacy counter. And partly because changing a slogan is slow. Back in 2017, a group of British infection specialists put it in the BMJ, under a title that's more or less the whole argument: the antibiotic course has had its day. But there's a real problem with announcing a change like that. People hear "the rule's changed" and stop listening to the rule entirely. Which, as we said, is the son's mistake. So the slogan's still around. And meanwhile the courses themselves were changing quietly. On the prescription. Where most people never notice.

Changed how, though? If Ruth got her box out of the cupboard next to an old one from ten years ago, what would actually be different on the label?

The number of days. So take Ruth's own diagnosis. The current UK guidance for adults with pneumonia says to stop antibiotic treatment after 5 days unless results suggest a longer course is needed or the person isn't clinically stable. Five days. And for children with a straightforward pneumonia, the same guidance says to offer a 3 day course. Then bladder infections. For a woman who isn't pregnant, the first-choice tablets in the UK guidance run for 3 days. Three. And sitting over all of it, the UK guidance on antibiotic use tells prescribers to follow guidelines on prescribing the shortest effective course. That phrase is the whole philosophy now. Not the longest course that's safe. The shortest one that works.

So hang on. Ruth's five days isn't the old long course she's thinking of cutting. It's already the short version.

It is. And that changes her question completely. She's sitting there thinking her GP gave her the old-fashioned long course, and she's weighing whether to trim it. She hasn't got that. She's got the modern short course, set short on purpose. And yes, the Dutch trial stopped at three days. But those patients were in hospital. A doctor checked they'd substantially improved before anybody stopped anything. Ruth's got a kitchen and a thermometer. That's all. Nobody's examining her on day three. So the two days she's eyeing up aren't the old padding. They're part of a course that's already been cut down. That's the difference.

And the guidance says unless she isn't clinically stable. That sounds like hospital language. What does it actually mean for someone at home, in her terms?

It means five days is where the course stops for somebody who's genuinely getting better. It isn't a promise that five is always enough. The guidance gives examples of what not stable looks like. One is having had a fever in the past 48 hours. And the others are things her doctor measures. Heart rate. Breathing rate. Blood pressure. Oxygen. So for Ruth, there are two ways five days might not be the end of it. Her results might show something that needs longer. Or she might not be well, and the fever keeps coming back. Neither of those is for Ruth to decide alone. But if she's on day five and still having temperatures, that's not a reason to quietly take Alan's leftover tablets. It's a reason to ring the surgery.

Wait. Alan? Who's Alan, and why has he got tablets? You've just brought a whole new person into this.

Her husband. I should introduce him properly, because he lives in the bathroom cabinet. Last spring Alan had a dental infection. He got a course, felt better after four days, and stopped. So there's a half-used strip sitting behind the plasters. And Ruth's been half thinking that if her cough comes back, she's got a backup. Which is one of the most ordinary ways this whole subject goes wrong in real houses. Not the person who stops on day three. The person who saves what's left for next time.

Okay, I want to come back to Alan's strip. But first, you said there were exceptions. Infections where the long course really does matter. So Ruth's mother wasn't wrong about everything?

She wasn't. And that's important, because the people most at risk from the son's link are people with one of these. The clearest example is tuberculosis. TB is treated with antibiotics, usually for at least 6 months. Six months. And the NHS is completely unambiguous about it. It says it's important to take your antibiotics correctly and until you've completed the course, even if you feel better. If you stop your treatment early, TB could come back. That's the mother's rule, word for word. And for TB it's exactly right. Every dose, however well you feel. And TB isn't the only infection where a long course is standard. Those are usually run by specialists, with the length chosen for a reason. Nobody's suggesting cutting those short. And nobody should stop them because they feel better.

But TB is a specialist thing, and Ruth's chest infection isn't TB. Is there anything closer to home, something a GP treats every week, where longer still wins?

There is. And it's a good one, because it shows the evidence isn't one-directional. Sore throat. Antibiotics aren't routinely used for sore throats any more. But when a doctor does prescribe penicillin for one, the UK guidance gives a range. For a sore throat that needs penicillin, the guidance says it's for 5 to 10 days, and that 5 days may be enough for symptomatic cure, but a 10-day course may increase the chance of microbiological cure. So that's the honest trade. Five days may well be enough to make you feel better. But ten days may give a better chance of clearing the bug. And for some people that difference matters. So "always shorter" isn't the new rule either. The new rule is the right length for that infection, for that person. That's a judgement. Which is why it's on the prescription and not in your head.

And the bladder infections? You said three days there, and that sounded very short to me. Is there a trade like that too?

Yes. And it's been measured well. There's a Cochrane review of bladder infections in women that pooled dozens of trials. Three days of antibiotic therapy is similar to 5-10 days in achieving symptomatic cure, while the longer treatment is more effective in obtaining bacteriological cure. Same pattern as the sore throat. The symptoms? Just as good either way. The bacteria are cleared a bit more often with the longer course. But the review also found adverse effects were significantly more common in the 5-10 day treatment group. That's the bit that gets lost. Longer isn't free. Every extra day is another day of feeling sick, bloating, diarrhoea. So somebody has to weigh it. And the UK guidance has come down on three days for women who aren't pregnant. The review itself says a longer course could be considered for women where clearing the bacteria is important.

So we've got Ruth on day three of five, feeling fine. Her mother's reason doesn't hold, and her son's instruction is wrong. What's the actual answer for her, this morning?

Take the last two days. And I want to be clear about why, because it isn't the mother's reason. It's not that stopping now would breed superbugs. It's that five days is already the short course the current guidance sets for what she's got. The trials that went shorter did it with doctors checking people first. So the honest version of the advice is take the course you were prescribed. For common infections, it's often already been cut down. And if you think it's longer than it needs to be, that's a question for the person who wrote it. Not a decision for the kitchen.

And what if she forgets one? She's been feeling fine, she's busy, and suddenly it's lunchtime and the breakfast tablet's still sitting there in the packet.

That happens to everybody. And the NHS page on antibiotics covers exactly this. First, check the leaflet that came with the medicine. But in most cases, you take the dose you missed as soon as you remember, and then carry on with the course as normal. If it's almost time for the next dose anyway, you skip the missed one and carry on as usual. And the one thing not to do is take a double dose to make up for it. So a missed tablet isn't a disaster. It's a thing to sort out calmly. And if she's not sure, that's a question a pharmacist answers all day long.

But what if she does ring and ask about the length, too? Is that a reasonable thing to do, or is she going to be told off for wasting somebody's time?

It's completely reasonable. And the obvious person is the pharmacist, who's often the quickest person to reach. She can say: I'm on day three of five and I feel fine, is five days the right length for me? She'll probably be told yes. And she'll be told why, which is worth having. If the answer's ever different, it's because somebody who knows the infection and her history looked at it. That's what matters. The decision got made by somebody who could actually make it. That's the whole shift, really. The question moved from the patient's willpower to the prescriber's judgement. And patients are allowed to ask the prescriber to use it.

And what does the NHS actually tell people now, on its own website? Because I'd assumed it's still the mother's line, word for word.

It's quieter than that. The NHS's main page on antibiotics says to take antibiotics as directed on the packet or the patient information leaflet, or as instructed by your GP or pharmacist. That's the wording. It's careful. That page doesn't say feel free to stop. And it doesn't tell you to finish every tablet whatever happens, either. The TB page does that. For TB. So the general line is simply: follow the instructions you were given. And that's the version that survives everything we've talked about. It works for Ruth's five days. It works for a three-day bladder infection. It works for six months of TB. Because the length lives on the label now. And the label is where the evidence went.

Okay. Now Alan's strip. The half-pack behind the plasters. It feels sort of thrifty to keep it. What's actually wrong with that?

Almost everything. Start with the obvious one. It wasn't prescribed for Ruth. And the NHS is plain about this. Only ever take antibiotics prescribed for you, and never borrow them from a friend or family member. Alan's tablets were picked for a tooth. A chest infection might need a different drug altogether, or a different dose. Then there's the amount. It's half a course. So if Ruth takes it, she gets a wrong-length course of a possibly wrong drug, for an illness nobody's diagnosed. And the third problem is the sneakiest. It makes her less likely to see anybody. If the cough comes back and it's something that needs a doctor's ears and a different treatment, a strip in the cabinet is a reason to put that off. And that delay is the real risk.

And why are there leftovers at all? If everyone's supposed to take exactly the course they're given, where do all these half-strips actually come from?

Two places, mostly. One is Alan. People stop early without asking, and the rest goes in a drawer. The other is more boring, and it isn't anybody's fault. The box doesn't always match the course. If the packet holds more tablets than the days you were prescribed, there's a remainder before you've done anything wrong. And that's been noticed. The UK guidance on antibiotic use asks the people who run the system to consider supplying antimicrobials in pack sizes that correspond to the recommended course lengths. Which is a small, unglamorous fix. But it's a fix for exactly the strip in Alan's cabinet.

So what should she actually do with Alan's strip, then? Keep it just in case, or get rid of it somehow?

Not keep it. Ask the pharmacist what to do with it. They'll know what's done locally. And then there's the bigger point, which is about Alan rather than Ruth. Alan's half-strip exists because he stopped on day four of his course without asking anybody. Under the mother's rule, that was a sin. Under the son's rule, that was smart. Under the actual evidence, it was neither. It was a guess. Maybe four days was enough for his tooth. Maybe it wasn't. Nobody checked. And the leftover tablets are the evidence that the decision got made in the wrong place.

There's something slightly uncomfortable in all this, though. People were told one thing for decades, and now the reason turns out to be wrong. Does that make them trust the next piece of advice less?

It might. And I don't think the answer is to pretend the old rule was clever. Its reason didn't survive being checked. But I'd put it like this. The behaviour it produced was mostly fine. Sticking to what the doctor wrote. It was the reason that was wrong. And the reason matters, because the wrong reason is what makes people think the son's link is a licence. If you think finishing was about killing every last bug, then the rule being dropped sounds like "stop whenever". If you understand it was about getting the length right, the change sounds like what it was. Somebody went and measured the length. And found that for most common infections, courses could be shorter. That's the kind of correction you'd want medicine to make.

Let me push on the resistance thing once more. Because people hear "antibiotic resistance" and think it's about their personal tablets. Is Ruth's two days going to make any real difference to the superbug problem?

For Ruth, today, probably not much either way. And I think it's healthy to say that out loud. Resistance is a population problem built out of millions of courses. The bits one person controls are mostly about not taking antibiotics that weren't needed in the first place. Not pushing for them for a cold. Not taking somebody else's. Not keeping a stash. The NHS's own page makes the point that many mild bacterial infections get better on their own without using antibiotics. So if Ruth wants to do something about resistance, the big lever isn't her last two days. It's the next time she's tempted to ask for antibiotics for something that would clear by itself. Or the next time she's about to reach into the cabinet.

So bring it back to her. It's still this morning, she's still in the kitchen, and the box is on the side. What does she actually do?

She takes her tablet with breakfast. She finishes the five days, because that's already the short course for what she's got. If she wants to be sure, she rings the pharmacy and asks whether five days is right for her. She'll probably hear yes, along with the reason. Then she watches for the things that would change the plan. A fever that comes back. Breathlessness getting worse instead of better. Not improving at all. Any of those, she rings the surgery rather than raiding the cabinet. And this weekend she asks the pharmacist what to do with Alan's strip. Then there's the conversation with her son. He wasn't wrong that something changed. He was wrong about what. The length got shorter. The decision didn't move.

Before we finish, what don't we know? Because I think people will hear all this and assume the shortest possible course is now settled, for everything.

It isn't. The first gap is resistance itself. The overview of reviews we mentioned found good evidence that short courses cure people as well. But its authors also wrote, in so many words, that they found inadequate evidence about the effect on antibiotic resistance. So the case that shorter is better for resistance rests mostly on observational links and on mechanism. Plus a handful of trials like the one in the Dominican Republic, where the dose changed too. That's a decent case. It's not a closed one. Gap two. The infections that haven't been tested. Most of the trials are in the common things. Chests, throats, bladders. For plenty of rarer infections, nobody's run the comparison, and the long course stays the default because nobody's shown it can be cut. And gap three is the one I find most interesting. Almost all of this compares one fixed length with another fixed length. What nobody's shown is whether you could tailor the stop to each person safely — without a doctor at the bedside — and what that would even look like. That's the version Ruth's son thinks already exists. It doesn't. And until somebody tests it, the length stays on the label.

If Ruth takes one thing from this whole conversation, what's the one thing you'd want it to be?

That her mother and her son were both half right. The courses really have got shorter. And the old reason for finishing them doesn't hold up. But the answer isn't to stop when she feels better. It's to take what she was prescribed, because it's probably already short. And to ask, if she thinks it isn't.

And three things she could take to her doctor, or her pharmacist, next time she's handed a box of these.

First. Is this the shortest course that will work for my infection? It's a perfectly normal question. And it tells the person you'll take what they decide seriously. Second. What should make me ring you before the course is finished, and what should make me ring you after it? Because that's the safety net. And it's different for a chest than for a throat or a bladder. And third. What do I do with any tablets left over? Ask that every single time. Because a strip in the cabinet is how one person's course turns into somebody else's guess. It's the dullest question of the three. It's also the one that would have kept Alan's strip out of the cabinet.