Infection after wisdom tooth surgery: signs and timing

Go to an emergency department now, and call 911, if you are having difficulty breathing or swallowing — including difficulty swallowing your own saliva — if your voice has changed in quality, if swelling has lifted the floor of your mouth so the tongue is pushed upward, if swelling has closed an eye or spread down into your neck, or if your mouth is opening less today than yesterday while you have a fever. Each of those means an infection has reached the tissue spaces around the airway, and that is a hospital problem from the moment it starts.

That instruction sits above everything else on this page because the practice telephone is a business cell phone rather than a staffed switchboard. It is not covered around the clock and it cannot be relied upon to be answered inside the few minutes an airway symptom allows. An emergency department can secure an airway, image the neck and admit you. A voicemail cannot do any of those things, and the half hour spent waiting for a call back is the half hour that mattered. For anything on the list above the order is hospital first, practice second.

Everything below is about the far more common situation. Something feels wrong, it is almost certainly not an airway emergency, and you are trying to work out whether what you have is ordinary healing behaving badly, a dry socket, or a genuine infection. On day three those three can look almost identical. By day six they behave completely differently, and the difference is in the shape of the change rather than in how much it hurts.

An infected extraction site is recognised by the shape of the trouble rather than by how severe it feels. Ordinary post-operative swelling rises for roughly forty-eight hours, holds for a day, then recedes a little each morning. Infection characteristically appears later — most often somewhere between day three and day seven — and it does the one thing ordinary inflammation never does: it gets worse after having got better. A face that was improving on Thursday and is tighter, hotter and more painful on Saturday is the pattern that matters, whatever the absolute amount of swelling happens to be.

Severity on its own is close to useless as a signal, and relying on it sends people to hospital who did not need to go and keeps people at home who did. A young patient after a difficult lower impacted removal can have a face that looks alarming on day two, sleep badly, struggle to eat, and be entirely on the expected curve. Another patient with a modest, unimpressive swelling, a temperature of 38.4°C and a jaw that will not open two fingers is in more trouble, and looks less unwell in a photograph.

Four things to track, and one question to ask each morning

• Swelling — is the face larger, the same, or smaller than it was at this time yesterday? Photograph it in the same light each morning rather than judging from memory.
• Pain — is the trend downward, flat, or upward? A dose of analgesia that covered you comfortably on day two and does not cover you on day four is a change in the underlying problem, not a change in the medicine.
• Mouth opening — measure it in finger-widths stacked vertically between the front teeth. Restricted opening is expected and is at its tightest around day two. Opening that is reducing after day three is the single most useful warning sign in the mouth.
• Temperature — take an actual reading rather than guessing from how you feel. A brief rise up to about 38°C in the first twenty-four hours is common inflammation. A temperature of 38°C or above appearing on day three or later is not.

The question to ask each morning is simply whether today is better, the same, or worse than yesterday. One worse day after day three is worth watching closely. Two consecutive worse days is a reason to be seen. The value of the question is that it forces you to compare against a fixed point instead of against the worst hour of last night, which is what most people unconsciously do at three in the morning.

Taste and discharge are the other useful signals, and they are more specific than pain. A foul taste that returns within a few minutes of rinsing, or a discharge you can express from the gum by pressing gently near the socket, points at a collection of pus. A bad taste on its own, with no swelling and no fever, is much more often trapped food or a dry socket than an abscess — which is precisely why the distinction below is worth getting right.

Post-operative infection after wisdom tooth removal most commonly declares between day three and day seven, with day four or five the usual point. Very little that appears in the first forty-eight hours is infection at all; that window belongs to inflammation, which is the repair response doing what it is supposed to do. A second and less familiar pattern also exists — delayed infection, surfacing somewhere between two and four weeks after surgery, nearly always in a lower socket, and often built around a retained bone fragment, a suture, or food that has packed into a socket still healing from the base upward.

The ordinary course runs roughly like this. Bleeding oozes for a few hours and stops. Swelling builds through the first and second day and reaches its maximum around forty-eight to seventy-two hours. Bruising, where it appears, comes out on day three to five and travels downward along the jaw and into the neck under gravity, which is unnerving and is not spread of infection. Jaw stiffness is tightest around day two or three and then eases. Pain is worst on day one and two, once the local anaesthetic has worn off, and reduces from there. By day seven most people are eating carefully but normally.

Several ordinary events in weeks two to four get reported as infection and are not. Small sharp fragments of bone working their way out through the gum are common after a surgical removal and are lifted out in a minute or two once they are loose. Dissolving sutures loosen and come away between day five and day fourteen, and a suture hanging by one end feels like something has gone wrong. Food packs into a socket for two to four weeks until the depth fills in, producing a bad taste after meals; an irrigating syringe used gently after eating deals with it.

What separates those from infection is that none of them reverses direction. The fragment does not produce a fever. The suture does not close your mouth. Food packing does not make the face swell. If a symptom in week two or three comes with swelling, temperature or reducing mouth opening, it is not a spicule of bone and it should be looked at.

Dry socket is not an infection. It is the loss or breakdown of the blood clot that should be sitting in the socket, which leaves bare bone exposed to air, saliva and food. Bone has no soft-tissue covering to shield it and very little capacity to protect itself, so it hurts — frequently more than the surgery did — but there is no collection of pus, no invasion of the surrounding tissue, and no fever. That is why antibiotics do not treat it, and why a course prescribed for it costs a week and delivers nothing.

It occurs in the region of two to five per cent of routine extractions and substantially more often after removal of impacted lower wisdom teeth, where published series report figures ranging into the twenties and occasionally higher. The risk rises with smoking or vaping in the first days, with vigorous rinsing, spitting or drinking through a straw before the clot has organised, with a surgically difficult removal, with a previous dry socket, and with oral contraceptive use. It occurs far more often in lower sockets than upper ones.

Treatment is mechanical rather than pharmacological. The socket is irrigated to flush out debris, and a medicated dressing — usually eugenol-based — is placed in contact with the exposed bone. Relief is typically within the hour and is one of the more dramatic changes in minor oral surgery. The dressing is renewed every two to four days until the granulation tissue covers the bone, which generally takes five to seven days in total. The reason to be seen rather than to manage it at home is that the dressing is the treatment, and nothing available over the counter substitutes for it.

The conflation with infection is understandable, and patients are not the only people who make it. Both produce pain that escalates after a period of improvement, in the same few days, in the same socket. The separating features are swelling, fever and trismus. A dry socket produces none of the three. If you have severe pain and one of those three is present, the diagnosis is not a straightforward dry socket, and it wants an examination rather than a dressing.

There are four possible responses to a post-operative symptom, and mixing them up causes harm in both directions. People go to an emergency department at midnight with a bruise, and people wait for a call back with a jaw that is closing. The rungs are not graded by how frightened you are. They are graded by how fast the situation can change.

1. Watch, and reassess in the morning. For anything on the expected curve — swelling still rising inside the first forty-eight hours, bruising appearing on day three, stiffness, discomfort that responds to analgesia at proper doses.
2. Telephone the practice. For things that need advice, an irrigation, or a routine appointment in the next few days: food packing, a loose suture, a bone fragment, a bad taste with no swelling or fever, or a question about medication.
3. Be seen within twenty-four hours, in person. For the patterns that mean something has changed and cannot be assessed down a phone line: swelling increasing after day three, fever from day three, expressible pus, severe pain from day two to four with an empty-looking socket.
4. Emergency department now, and call 911 for anything involving breathing or swallowing. For airway or floor-of-mouth involvement, an eye closed by swelling, swelling spreading down the neck, fever with a mouth that is closing, or bleeding not controlled by twenty minutes of pressure. Telephone the practice afterwards.

The rung that gets misused is the second one. Patients with a genuine third-rung problem telephone, leave a message, and then treat the wait for a call back as though it were treatment. It is not. If you have decided that something needs to be seen within a day, the decision is to be seen — through the practice if it answers promptly, through an urgent dental service, a hospital dental department or an emergency department if it does not. The route matters much less than the timing.

Most people do not need antibiotics after wisdom tooth surgery, and an uncomplicated removal in a healthy patient is routinely managed without them. Swelling, stiffness and soreness in the first two or three days are inflammation, which is the repair response and not something an antibacterial drug acts on. Antibiotics work against bacteria that have invaded tissue. They do not reduce ordinary post-operative swelling, they do nothing whatever for a dry socket, and they do not empty a collection of pus once one has formed.

The evidence on giving them routinely before surgery is real and modest, and it is worth stating in numbers rather than in impressions. Systematic reviews of prophylaxis for third molar removal find a genuine reduction in infection, on the order of nineteen patients treated to prevent one infection against a background rate of a few per cent. In the same reviews roughly one patient in ten treated experiences an adverse effect. Most are nuisances — nausea, diarrhoea, thrush, a rash. The uncommon ones are not: Clostridioides difficile colitis, and anaphylaxis, which is the reason a drug allergy history is asked about twice.

When a course is genuinely indicated

• Infection that has spread beyond the socket into the surrounding soft tissue: a swelling with a defined edge that is enlarging, with warmth and redness of the overlying skin.
• Systemic signs — a temperature of 38°C or above from day three, feeling generally unwell, a raised heart rate, or tender lymph nodes under the jaw.
• A compromised immune system: chemotherapy, high-dose or long-term steroids, immunosuppressants after a transplant, or poorly controlled diabetes. The same infection moves faster in those patients and is treated earlier.
• Involvement of the deeper tissue spaces — in which case the antibiotic goes alongside drainage, frequently intravenously and in hospital, and never instead of drainage.
• Specific cardiac and prosthetic-joint histories, where a single dose is given before the surgery rather than a course afterwards. That is a different decision made for a different reason, and it is settled at the consultation, not on day five.

Restraint here is not thrift and it is not indifference. Every course prescribed to close an uncomfortable conversation selects for organisms that will be harder to treat in somebody, possibly in you, and adds a one-in-ten chance of a side effect to a patient who was going to recover anyway. If antibiotics are not offered, the useful question is not why they are being withheld but what would change the answer — and the reply should be specific enough to act on. A temperature above this number. A swelling that reaches this landmark. A mouth opening tighter than yesterday.

If you have been given a course, take it exactly as directed and note the clock. An antibiotic that has produced no improvement within forty-eight to seventy-two hours of the first dose usually means one of two things: there is a collection that needs draining, or the diagnosis was wrong and the problem was never bacterial. Both of those need a re-examination rather than a second prescription. Continuing to swallow tablets while nothing improves is the commonest way a treatable socket infection becomes a hospital admission.

The examination is the treatment decision. The socket is inspected and gently probed, the swelling is felt for whether it is firm and diffuse or soft and pointing, mouth opening is measured, the temperature is taken, and the lymph nodes under the jaw and in the neck are checked. A radiograph is taken where a retained root fragment or a bone sequestrum is suspected, because a foreign body keeps producing infection until it is removed, and no length of antibiotic course changes that.

Where pus has collected, it is released. In a socket that often means irrigating the socket thoroughly under local anaesthesia and clearing debris and granulation tissue from it. Where the collection sits in the soft tissue, a small incision is made and the cavity is opened, sometimes with a drain left in place for a day or two. It is worth knowing in advance that local anaesthetic works less predictably in acutely inflamed tissue, so the numbness may be less complete than it was on the day of surgery, and that is a property of infected tissue rather than a failure of technique.

Some infections belong in a hospital rather than in a dental chair, and saying so early is part of the job. Involvement of the submandibular, sublingual or pharyngeal spaces, trismus that prevents adequate examination, a patient who is systemically unwell, or any concern about the airway means transfer for imaging, intravenous antibiotics and drainage — which for a deep space collection is generally performed under general anesthesia in an operating theatre. That decision is made on the spot and it is not a failure of the original surgery. It is the reason the fourth rung of the ladder exists.

Upper wisdom teeth have their own version of this. The roots of an upper third molar frequently sit against the floor of the maxillary sinus, and a communication between mouth and sinus is a recognised outcome of removing one. Where infection develops there, the signs are different from the classic swollen cheek: fluid or air passing between the mouth and the nose, a one-sided nasal discharge, pain across the cheekbone, and a taste change when bending forward. Those are also worth reporting rather than watching, because a small communication managed early behaves very differently from one that has been open for a month.

The day-three-to-seven pattern describes a healthy adult. Several groups run a different course, and for them the thresholds on this page are too slow. If you are having chemotherapy, take immunosuppressants after a transplant, are on long-term or high-dose steroids, have poorly controlled diabetes, or have a condition affecting white cells, an infection can advance without producing the swelling and fever that ordinarily announce it. In those situations the instruction is to make contact at the first change rather than to wait for a second worse day, and to say plainly on the call which of these applies to you.

Bone-modifying medication is its own category. Bisphosphonates and denosumab, particularly at oncology doses, are associated with medication-related osteonecrosis of the jaw, in which bone at an extraction site fails to heal and stays exposed. Exposed bone that persists beyond eight weeks after surgery is not a slow dry socket and it is not an ordinary infection, although it can become secondarily infected. It is a distinct diagnosis managed differently, and it needs assessment by someone who deals with it regularly rather than repeated courses of antibiotics. The same is true of anyone who has had radiotherapy to the head and neck.

Smoking and vaping change the odds rather than the diagnosis. Both reduce the blood supply at the wound margin and both are strongly associated with dry socket and with slower closure. Older patients, and anyone whose lower tooth was deeply impacted and needed bone removal, tend to swell more and to recover over a longer arc, which shifts every date on this page a few days to the right without changing the shape of the curve.

It is also worth being direct about what waiting costs. Most of the time, waiting an extra week with an infected socket costs a week of avoidable pain that a ten-minute irrigation would have ended. Occasionally it costs considerably more: an infection that stays in the socket for four days can be in the muscle spaces on the fifth, and the treatment then involves a hospital admission, general anesthesia and a drain rather than a syringe of saline. The distance between those two outcomes is usually a day or two of hoping it settles.

Whoever assesses you is trying to establish direction, and a call that supplies it gets a faster and more accurate answer. Have the facts ready rather than describing feelings, because the same sentence — "it is really painful" — is said by the patient with a dry socket and by the patient with a spreading infection, and it does not separate them.

• The date of the surgery and which tooth or teeth were removed, upper or lower, left or right.
• Your actual temperature, as a number, taken in the last hour.
• How wide your mouth opens today in stacked finger-widths, and what it was yesterday.
• Whether the swelling is larger, the same or smaller than yesterday, and where it now reaches — cheek, under the jaw, toward the eye, into the neck.
• Whether there is a foul taste, a discharge, or pus you can express from the gum.
• Every medication taken since surgery, with the dose and the times, including anything bought over the counter.
• Any medical condition or medication that affects healing or immunity, and any drug allergy.

When you are seen, the questions that produce the most useful consultation are the ones that ask for a finding rather than a reassurance. What did you see in the socket. Is there exposed bone, a collection of pus, or a retained fragment. What is the diagnosis, in one word. If an antibiotic is being prescribed, what specifically is it treating, and what happens if it does not work. If it is not being prescribed, what should make me come back, expressed as a number or a landmark rather than as "if it gets worse".

Ask what the next forty-eight hours should look like if the treatment is working. A specific answer — the pain should drop within an hour of the dressing, the swelling should stop increasing by tomorrow evening, mouth opening should be wider by Wednesday — gives you a test you can apply yourself at home. A vague answer gives you nothing to measure, which puts you back to guessing at three in the morning, which is where this whole page started.