Antibiotics after wisdom teeth: when they help and when they do not

Most people reading this either have a prescription in their hand and want to know what it is for, or do not have one and are wondering why not. A few are reading it because something has changed since surgery. That last question goes first, because the situations that need action today are not the ones that hurt the most, and no tablet taken at home substitutes for being looked at.

A temperature at or above 100.4°F (38°C) together with facial swelling sits one step below that list. It needs assessment the same day rather than within the hour, and the same day means the same day. This practice does not open every day of the week, and an infection that is moving does not pause for opening hours, so where the office cannot be reached the correct destination is an urgent care centre or an emergency department.

There is a second urgent situation that specifically concerns antibiotics, and it is the one patients most often sit on. A course that is going to work generally declares itself early. If it is not working, waiting for the packet to finish is waiting through the part of the illness where the decision needed to be made.

Two other things are reported the same day and neither is an antibiotic question. Numbness, tingling or altered feeling in the lip, chin or tongue that is still present once the local anaesthetic should have worn off is reported to the office on the day you notice it, rather than held for a scheduled review, because the windows in which referral to a nerve injury service is most useful are time-limited. Bleeding that is still frankly running after two thirty-minute periods of firm, unbroken pressure is also a same-day call, at whatever hour, and after the first period rather than the second if you take an anticoagulant or antiplatelet medication.

One group of patients is moved up that list by history rather than by symptoms. If you are immunosuppressed, are having chemotherapy, have poorly controlled diabetes, or take a bone-modifying medication, a dental infection is assessed sooner and at a stage that would be an ordinary appointment for someone else. Say which medication you take when you call, because it changes what happens next.

An extraction leaves a wound in a part of the body that is never sterile and was never expected to be. The mouth carries a resident bacterial population at all times, and a healthy socket heals in the presence of it: the clot forms, tissue organises over it, and the body's own defences handle the rest. Sterility is not the goal and is not achievable. The question is whether bacteria establish an infection in the tissue, and after a straightforward removal in a healthy patient, they usually do not.

Routine antibiotics for third molar surgery have been studied repeatedly, and the summary cuts in both directions. Giving them to everyone does reduce the number of people who go on to develop a wound infection or a dry socket. It also means a large number of patients take a course in order that one of them avoids a problem, while every one of them carries the side effects, the interaction risk and the pressure on resistance. That trade is what sits behind the recommendation in most guidance against routine use, and it is why a prescription in an uncomplicated case is something to be argued for rather than assumed.

The harms are not theoretical. Nausea, loose stools, a metallic taste and thrush are the common ones. Allergic reactions run from a rash that settles to anaphylaxis, which is uncommon and immediate. Antibiotic-associated colitis is the serious one and can begin weeks after the course has ended. And antibiotics interact: with blood-thinning medication, with anti-rejection medication after a transplant, and, in one class in common dental use, with alcohol.

What actually lowers the chance of a wound infection after third molar surgery is mostly not pharmacological. It is how the soft tissue is handled, how thoroughly the site is irrigated before it is closed, whether existing infection is settled before an elective removal rather than operated through, and what happens in the first week: keeping the area clean once rinsing is permitted, not smoking, and eating enough to heal on. A prescription does not compensate for any of those and was never designed to.

There is a difference between an infection risk and an infection. The situations where antibiotics genuinely earn their place are ones where infection is already present and has moved beyond the site, where the patient's defences are impaired, or where the operation itself was long, contaminated, or performed through active disease. Those are specific findings rather than a general sense that the case was a difficult one.

Infection that has spread beyond the socket

A sore socket and an inflamed gum flap are local problems, and the treatment is mechanical: cleaning under the flap, irrigating the socket, removing what is driving it. Once swelling has crossed into the cheek, under the jaw, or up towards the eye, it has entered a tissue space and is no longer confined to where it started. That is the point at which a drug carried in the bloodstream has something useful to reach, and it is given alongside drainage rather than in place of it. Where pus has collected it is released as well. Relief tends to follow the drainage, which is the correct order of expectations to hold.

Systemic signs

Fever, feeling generally unwell, a rising heart rate, tender nodes in the neck, and jaw opening that is tightening are the body responding beyond the mouth. Any of them changes the assessment. A temperature at or above 100.4°F (38°C) with facial swelling is a same-day matter whether or not a course has already been started, and starting one at home does not convert it into something that can wait until the office next opens.

A compromised immune system

Patients on immunosuppressants, having chemotherapy, with poorly controlled diabetes, or taking bone-modifying medication are managed on a shorter fuse. The same infection moves faster and matters more, so cover is often used at a stage where an otherwise healthy patient would be watched. Transplant medication brings a second consideration: the macrolide and azole classes raise the blood level of tacrolimus and ciclosporin, so what is prescribed is chosen against that list and, where it matters, in conversation with the transplant team.

The operation itself

Some removals are not straightforward. A tooth taken out through an active infection that could not be settled first, a long procedure with substantial bone removal, a site that had to be reopened, or a coronectomy where roots are deliberately retained: each is judged individually. Whether antibiotic cover is used in those circumstances varies between surgeons and between cases, and it is a decision made against your medical history rather than by a fixed protocol.

A small group of patients are given an antibiotic to prevent an infection of the heart lining rather than to treat one in the mouth. Dental procedures that disturb the gum or the tissue around a root release oral bacteria into the bloodstream, which happens routinely and is normally cleared without consequence. In a heart with certain abnormalities those bacteria can settle on the affected surface, and the resulting infective endocarditis is serious enough that a single dose beforehand is judged worthwhile.

Two features of it are the ones most often misremembered. It is given before the procedure, in one dose, so that the drug is circulating at the moment the tissue is disturbed. And it is not a course afterwards. Continuing tablets for a week does nothing about a risk that existed for the few minutes the extraction took, and a course started the following day is not the same intervention under a different name.

The list of qualifying histories has narrowed considerably since the 2000s, and a good deal of what patients were told two decades ago no longer applies. As current American Heart Association guidance stands, it centres on a short set of conditions.

• A prosthetic heart valve, or a valve repaired using prosthetic material.
• A previous episode of infective endocarditis.
• Specific congenital heart conditions: unrepaired cyanotic disease; a repair using prosthetic material within the first six months after it was done; or a repair with a residual defect at or beside the prosthetic patch or device.
• A heart transplant recipient who has since developed a problem with a valve.

Several conditions patients expect to find on that list are not on it. Mitral valve prolapse, a bicuspid aortic valve, most murmurs, rheumatic heart disease without a prosthetic valve, coronary stents, bypass grafts, pacemakers and implanted defibrillators do not generally call for cover. Neither, on current guidance, does a replaced hip or knee: routine antibiotic cover before dental treatment is not recommended for most joint replacements. Guidance in both areas has moved more than once and may move again, so the determination for an individual patient rests with the cardiologist or the orthopaedic surgeon who knows the history. Where a surgeon has advised cover for a specific patient, that advice is taken up with them rather than overruled from a dental chair.

A meaningful share of requests for antibiotics after surgery are for things antibiotics have no effect on, which is understandable: the drug is reasonably associated with getting over an illness, and pain after an operation feels like something that ought to have a treatment attached to it. But prescribing for the wrong problem leaves the problem in place and adds the side effects on top of it.

Dry socket is the one worth being explicit about, because it produces some of the most severe pain of any ordinary complication and patients reasonably conclude that something that hurts that much must be an infection. It is not. It is loss of the blood clot from the socket, usually on day three, four or five, leaving bone exposed at the base. The signature is pain that had been settling and then sharply worsens, often radiating towards the ear, with a bad taste that rinsing does not shift. Treatment is an urgent appointment at which the socket is irrigated and dressed, and most patients report substantial relief within thirty to sixty minutes of the dressing going in. An antibiotic prescribed for it treats nothing and delays the appointment that does.

Pericoronitis confined to the gum flap is the other frequent misallocation. Antibiotics reach tissue through the bloodstream; they do not flush debris out of a pocket. Where the flap is sore but jaw opening is normal and there is no fever and no facial swelling, cleaning under the flap addresses the cause, and a course leaves the pocket exactly as it was. Many localised episodes quieten either way over several days, which is precisely why the course tends to get the credit and why the episode tends to come back. That is an argument for having the flap cleaned rather than an argument for waiting to see, because the episode that settles on its own leaves the anatomy that produced it untouched.

The written instructions you are given govern. They are specific to what was prescribed, to your medical history and to the operation that was performed, and where anything on this page differs from them, they win. Doses and timings are not published here for that reason: a general article is the wrong place for a number that has to match your prescription.

Start it when you are told to start it, which for surgical cover may mean before the procedure rather than after. Space the doses roughly evenly across the day rather than clustering them into the evening, because the point is to keep the drug present rather than to consume a certain number of tablets. Follow the instruction about food: some are taken with a meal to reduce stomach upset and some on an empty stomach because food interferes with absorption, and the two are not interchangeable.

Feeling substantially better after two days is what a working course looks like. It is not a reason to stop. The slogan about always finishing the course has been genuinely questioned in general medicine, and shorter courses are now appropriate for a number of infections, but that argument is about what length gets prescribed, not about a patient shortening one unilaterally at the point where the swelling has gone down. If you think the course looks longer than it needs to be, ask, and ask before you start it rather than on day three. Stopping early leaves whatever survived to regrow, and a second attempt is generally harder than the first.

• A missed dose is taken as soon as you remember, unless the next one is nearly due, in which case skip it. Two at once to catch up is not how that is corrected.
• Vomiting within an hour or so of a dose usually means it has not been absorbed. Ring and ask rather than guessing in either direction.
• The nitroimidazole class, of which metronidazole is the member in common dental use, produces a marked and unpleasant reaction with alcohol. Where the instruction says avoid alcohol, that is a specific warning rather than general caution.
• Do not stop a course because a side effect is mildly unpleasant, and do not push through a rash or any breathing symptom. The first is a phone call; the second is a reason to stop and be seen.

Three things not to do with antibiotics, which is where a surprising share of the avoidable harm comes from. Do not take a part course left over from something else: it masks the signs that an infection is spreading, does nothing about the source, and the infection still needs clearing afterwards. Do not take someone else's, whatever they were prescribed it for. And do not keep what is left at the end of a course for next time; take it back to a pharmacy.

Most side effects are less dramatic than that and are still worth reporting. Nausea, loose stools, a metallic taste, headache, and oral or vaginal thrush are the common ones. Some classes raise sensitivity to sunlight, which is not a trivial consideration in south Florida in the week after surgery. A rash with no breathing symptom is still a reason to ring before the next dose, because whether it is recorded as an allergy or as an intolerance changes what can be prescribed for you for the rest of your life.

Diarrhoea deserves its own paragraph. Mild loose stools during a course are common and settle when it finishes. What is not ordinary is diarrhoea that is watery and frequent, that comes with cramping abdominal pain or fever, or that contains blood or mucus, and it can begin during the course or weeks after it has ended. Antibiotic-associated colitis, most often caused by the organism Clostridioides difficile, is uncommon and serious. The reason it is worth knowing about in advance is that almost nobody connects a bowel problem three weeks from now to a dental prescription today, and that connection is what gets it treated promptly.

Interactions are the other reason the medication list you brought to the consultation matters. Some antibiotic classes potentiate warfarin and change how thin the blood is at exactly the point where that is relevant. The macrolides and azoles raise the blood level of tacrolimus and ciclosporin after a transplant. On hormonal contraception the position is narrower than the warning most people have heard: the class that genuinely reduces its effectiveness is the rifamycins, used mainly against tuberculosis, rather than anything ordinarily prescribed in dentistry. The real mechanism with a common antibiotic is simpler, in that vomiting or significant diarrhoea shortly after taking the pill can stop it being absorbed. Package leaflets and pharmacists still vary in what they advise, and taking the more cautious instruction for one cycle is a reasonable position to hold.

One last thing about allergy, because the consequences reach well beyond dentistry. Penicillin accounts for most recorded antibiotic allergy labels, and a large proportion of the people carrying one are not, on testing, allergic. The label frequently traces back to a rash during a childhood illness that was itself causing rashes, or to nausea, which is an intolerance rather than an allergy. Carrying an inaccurate label means second-choice drugs for decades, and those are often broader in their effect and less well matched to the infection. If yours came from something vague and long ago, it is worth asking your physician whether it can be tested and, where appropriate, removed from your record. Until it is, it stays on the record here and is treated as real.

Antibiotic stewardship is the practice of prescribing these drugs only where they do more good than harm, in the narrowest form that works, for the shortest length that works. The word sounds administrative. What it describes is the effort to keep the drugs working.

Resistance is usually explained as a public problem, which makes it easy to file under things other people ought to worry about. The personal version is more immediate. Every course you take applies selection pressure to the bacteria living on and in you: the ones the drug kills easily are removed, and the ones that survive it are the ones that repopulate. The bacterial population you carry afterwards is harder to treat, and it stays that way for months rather than days. Someone who has had several courses in a year is someone whose next infection has fewer options available to it.

Dental prescribing is part of that picture rather than outside it. The two situations in which it is easiest to reach for a drug without a finding to justify it are pain, and a localised problem whose actual treatment is a procedure. Neither is treated as an indication here. The standard applied is a simple one: a prescription follows a finding, and where a procedure is what the finding calls for, the procedure is what happens.

Which means that being told you do not need antibiotics is a clinical decision rather than an omission, and it is a decision that can be revisited. What is being said is that your findings today do not indicate them, not that they are unavailable to you if the picture changes. What is given alongside that here is the specifics: what would change the judgement, what to watch for, and who to call when it happens.

There is one pattern worth naming, because patients frequently do not recognise it as a pattern at all: repeated courses for the same tooth. Two or three courses in a year for the same wisdom tooth is not a treatment plan. The recurring infection is itself the finding, and it is telling you something about the anatomy that keeps producing it, which does not change between episodes. At that point the useful conversation is about what to do with the tooth, not about which antibiotic to try next.

None of these are awkward questions, and every one of them has a specific answer.

• If I am not being given antibiotics, what would change that, and what should I be watching for?
• If I am, what is this treating, when do I start it, and what should I expect to see by day two or three?
• What do I do about a missed dose, or if I vomit shortly after taking one?
• Does this interact with anything else I take, including blood thinners, hormonal contraception, or medication after a transplant?
• Which side effects mean ring the office, and which mean stop the course immediately?
• Where do I go if the office is closed or the line is not answered, and what counts as a reason to go there rather than wait?
• If I have been told I need cover for a heart condition, what document confirms it, and has it been checked against current guidance rather than against what I was told years ago?

The answers that apply to you are the ones written on your post-operative instructions, which are specific to what was done and what was prescribed. This page is general. Where the two differ, the sheet is the one to follow.