Most sore throats, coughs and sinus infections are viral. What antibiotics actually treat, and what to do instead.
Being told you do not need antibiotics is unsatisfying. You feel dreadful, you have taken time to see someone, and you leave with advice about paracetamol and fluids.
It is worth explaining what is actually behind that answer, because it is not caution or rationing. Antibiotics kill bacteria. They have no effect whatsoever on viruses — not a reduced effect, none. And most of the infections people want them for are viral.
Entirely viral. An antibiotic will not shorten them by an hour.
The large majority are viral. Even for the minority caused by streptococcus, antibiotics typically shorten symptoms by about a day, and most cases resolve on their own regardless.
They are worth it in specific circumstances — which is what the FeverPAIN and Centor scores are for, and which is a judgement made from the specific features of your illness, not from how unwell you feel. See sore throat and tonsillitis.
Acute bronchitis is usually viral, and the cough commonly lasts three to four weeks whether or not you take anything. Pneumonia is different and does need antibiotics — distinguishing the two is precisely what the assessment is for.
Around 98% of sinusitis is viral. Antibiotics are generally reserved for symptoms lasting beyond ten days without improvement, or a distinctive pattern of getting better then markedly worse.
The majority resolve within a few days without treatment. Guidance supports a delayed prescription or watchful waiting for many children, with clear exceptions for the very young and the very unwell.
This deserves its own heading, because it is the single most persistent myth in general practice.
Colour tells you nothing about whether an infection is bacterial. Green mucus comes from an enzyme released by your own white blood cells doing their job. It is as common in viral infections as bacterial ones. It is not evidence of anything requiring antibiotics.
Antimicrobial resistance is not a distant policy concern. Drug-resistant infections are associated with a very large number of deaths worldwide each year, and the number is rising.
But there is a more immediate, personal argument, and it usually lands better.
Antibiotics have real side effects. Around one in ten people gets diarrhoea. Thrush is common. Rashes are common. Clostridioides difficile infection, which can be severe, is a recognised complication.
They disrupt your gut microbiome, and recovery takes months.
They interact with other medicines, including warfarin, methotrexate and some statins.
And resistance is personal, not only societal. Taking antibiotics unnecessarily makes your own bacteria more resistant, for months afterwards. So the next time you have a UTI, the first-line treatment is more likely to fail.
Taking an antibiotic you do not need is not a neutral act with a small chance of benefit. It is a small chance of harm with no chance of benefit.
Worth mentioning because it affects a surprising number of people.
Around 10% of people report a penicillin allergy. When formally tested, more than 90% of them are not allergic. Most were labelled in childhood after a rash during an illness — a rash that was very often the virus itself, or the classic amoxicillin rash of glandular fever.
This matters because the alternatives are generally broader-spectrum, less effective for some infections, and more likely to cause resistance and C. difficile.
If your allergy label came from a childhood rash you cannot remember, it is worth having it properly reviewed.
A useful middle option, and under-used.
You are given a prescription with instructions to hold it for two or three days and use it only if things worsen or fail to improve. Studies consistently show most people never fill it — and outcomes are the same as taking it immediately.
It removes the anxiety of being left with nothing, without committing you to a drug you probably do not need. If it would help you, ask.
Paracetamol and ibuprofen together, staggered, are more effective than either alone for fever and pain — and most people take too little, too irregularly.
Fluids, enough that your urine stays pale.
For sore throat: salt-water gargles, lozenges, and anaesthetic sprays.
For cough: honey has as much evidence as most over-the-counter cough medicines, and is cheaper. Not for children under one.
For blocked nose: steam, saline rinses, and short-term decongestants — no more than a week, or you get rebound congestion.
Rest. Genuinely.
Call 999 or go to A&E for: difficulty breathing or breathlessness at rest; a rash that does not fade under a glass; confusion or unusual drowsiness; a stiff neck with light sensitivity; blue lips; not passing urine; or any child who is floppy, unresponsive or breathing rapidly.
Seek same-day assessment for: feeling significantly worse after initially improving; fever lasting more than five days; coughing up blood; chest pain on breathing; being unable to swallow saliva; spreading redness with severe pain; or any infection in someone immunosuppressed, having chemotherapy, or without a spleen.
The people who need a lower threshold entirely: babies under three months, anyone immunosuppressed, anyone with chronic lung disease, and pregnant women. In those groups the calculation changes and the advice above does not straightforwardly apply.
We prescribe antibiotics when they will help. We say so plainly when they will not, and we explain why rather than simply refusing.
We would also rather tell you that during a £40 consultation than have you leave without an explanation. Sometimes the most useful outcome of an appointment is being told, with reasons, that this will get better on its own — and being told exactly what would change that.
A same-day consultation can assess an infection properly, prescribe where it is warranted, and tell you honestly when it is not.

Clinically reviewed by Dr Mohammad Zubair Khan, GMC 7563469
Last reviewed
August 23, 2026
Ipsum dolor sit amet consectetur commodo aliquam augue duis aliquet ipsum donec tempus ac interdum enim.