Olecranon bursitis (elbow hygroma): septic or not?

Why does the elbow bursa get infected more often than the others?

A bursa is a gliding cushion. The olecranon bursa works in conditions the others never meet: under thin skin, over a prominent bone, and without a millimetre of muscle to absorb the load. That anatomy explains both how common a hygroma is and its infection risk.

A bursa sitting under the skin, not buried in the tissues

The olecranon bursa is a potential space, lined with synovium, interposed between the posterior surface of the olecranon and the skin. Normally it holds a film of fluid of the order of a millilitre, clinically invisible. Its function is to let the skin glide over the bone during flexion and extension of the elbow, a movement in which the posterior skin travels several centimetres.

What sets it apart is its position: it is superficial. Between the outside world and the bursal cavity there is nothing but dermis and a very thin layer of subcutaneous tissue. Compare it with the psoas bursa, the trochanteric bursa or the pes anserine bursa: all are covered by thick, vascularised muscular or aponeurotic layers, which form a mechanical and immunological barrier. The olecranon bursa has none.

Two clinical consequences, not one

The first is mechanical. The whole load of leaning on the elbow passes through this cushion. Leaning on the elbows for long stretches, working on the floor on the elbows, crawling, resting the forearm on a hard armrest: repeated microtrauma is the commonest cause of olecranon bursitis¹³. The old names for the condition (student's elbow, miner's elbow, plumber's elbow), all describe the same mechanism: prolonged, repeated pressure.

The second is infectious, and it is the one that structures this article. A bursa separated from the outside world by a few millimetres of thin skin, over a bony prominence that grazes at the slightest friction, offers a skin staphylococcus a short route in. That is why Staphylococcus aureus accounts for about 80 % of superficial septic bursitis¹⁵, and for 88 % of culture-positive cases in the Calgary series⁸: the organism that infects the bursa is the one living on the skin above it.

That short route also explains the place of trauma in the history. In the Calgary cohort, a preceding injury was found in 53 % of cases⁸. In other words: when you look for a point of entry, you find one every other time. When you do not find one, that rules nothing out, because a minor breach of the skin may have gone unnoticed or already healed.

Trauma is the common cause of mechanical bursitis and of septic bursitis. It therefore does not tell them apart: it precedes both.

Stell was already making this point in his 1996 review: trauma “may cause either”⁷, and local erythema may be present in both. A history that turns up a fall on the elbow three days earlier is not reassuring, it opens both hypotheses.

What this changes compared with the other bursitis sites

At the other sites covered on this site, the dominant question is that of the procedure: whether to aspirate, whether to operate, whether to inject. At the prepatellar bursa, also superficial, the logic is very close to that of the elbow, and the two conditions are in fact studied together in the literature⁹. At the pes anserinus or the greater trochanter, deep bursae, infection is a rare possibility and the debate is about the associated tendinopathy.

At the elbow, the infection question comes first, and not only in frequency. It comes first because it is the only one where the error costs dearly and quickly: a neglected mechanical bursitis remains a nuisance; a neglected septic bursitis can progress to soft-tissue spread, osteitis of the olecranon, and exceptionally to systemic infection.

How do you recognise an olecranon bursitis in the clinic?

The positive diagnosis of olecranon bursitis is one of the simplest in elbow pathology. It is the diagnosis of severity that is difficult. This chapter sets out the picture; the next tackles the infection question.

A visual picture before a palpatory one

Olecranon bursitis presents as a rounded, well-defined swelling centred on the tip of the olecranon. Patients often describe it as having appeared quickly, sometimes overnight, and its size surprises them: a distended bursa can reach the volume of a golf ball. The consistency is tense, fluctuant when the effusion is copious.

Two clinical features carry strong orienting value, and both bear on what the bursitis is not :

• Passive elbow movement is preserved, or barely limited, and above all it is not painful through mid-range. A distended bursa mechanically gets in the way of terminal flexion, because the posterior skin tightens over the mass, but extension and mid-range stay free. That is what clinically separates bursitis from joint involvement: an arthritis of the elbow, septic or not, gives an elbow that is painful throughout range and held in an antalgic flexed position.
• The swelling is posterior and superficial, it can be moved under the fingers relative to the bone, and it does not fill the para-olecranon gutters. A joint effusion of the elbow, by contrast, shows as a filling of the lateral triangle between olecranon, epicondyle and radial head.

This distinction is not academic. In the historical series of Ho, Tice and Kaplan, the authors explicitly noted that septic bursitis was not associated with septic arthritis and could easily be distinguished from it by the characteristic bursal swelling and the joint examination³. The two conditions rarely coexist, and a mobile elbow is a strong argument in favour of the bursa.

The history, and what really needs looking for in it

Beyond the mechanism, four elements of the history change what you do:

1. The time course of onset. A swelling that builds over a few hours to two days, painful, points to infection or to a crystal flare. A swelling that builds over weeks, with little or no pain, in someone who leans on their elbows, points to a chronic mechanical hygroma.
2. The pain, and above all how out of proportion it is. A chronic mechanical hygroma is classically barely painful: the patient comes about the lump, not about the pain. A frankly painful bursitis at rest, one that wakes the patient at night, falls outside the mechanical picture.
3. The point of entry through the skin. A graze, a sting, a bite, a gardening wound, a skin condition at the elbow. Found in 53 % of septic bursitis cases⁸, and therefore absent in nearly one in two.
4. The background. Known gout or rheumatoid arthritis, immunosuppression, corticosteroid treatment, renal failure, alcohol misuse. These backgrounds change both the probability of infection and how it presents, often more insidiously.

What imaging brings, and what it does not

Ultrasound readily confirms the bursal effusion, measures its volume, shows synovial thickening and hyperaemia on power Doppler, and distinguishes a bursitis from a cellulitis¹³. It is also useful for guiding aspiration¹⁵.

What it does not say is whether the fluid is infected. This is a point on which the review by Lormeau and colleagues is explicit: ultrasound helps in diagnosing bursitis and in guiding the procedure, while telling aseptic from septic bursitis usually goes through examination of the aspirated fluid¹⁵. Thick, septated fluid with Doppler hyperaemia is compatible with infection; it is equally compatible with an advanced crystal-induced or haemorrhagic bursitis. Imaging does not replace aspiration when the question asked is that of sepsis.

Which signs point to an infection, and what are they really worth?

This is the heart of the article, and it is also where the literature is most often misquoted. The percentages circulating in reviews and quick-reference guides almost all describe cohorts of patients who are already infected. A frequency within the disease is not a discriminating power.

The methodological trap that has to be named first

The largest published series of septic olecranon bursitis is the Calgary one: 118 consecutive cases managed as outpatients over 21 months⁸. It gives the frequency of the signs with rare precision. And it is precisely that precision that misleads: every patient in that cohort is septic. There is no comparison group.

Saying “erythema is present in 92 % of septic bursitis” is accurate. Concluding that erythema allows infection to be affirmed is wrong: one would still need to know how many non septic bursitis cases show erythema. Yet Stell pointed out that local erythema can be seen in both forms⁷. A sign that is very sensitive but poorly specific does not help to confirm; it helps, when it is absent, to doubt.

The most counter-intuitive finding: fever is the exception

In that same cohort, fever defined objectively as a temperature of 37.8 °C or above was present in only 20 % of patients⁸. The difference with the 45 % who reported fever or chills on questioning is instructive: the symptom as declared is twice as frequent as the sign as measured.

Four patients in five with confirmed septic olecranon bursitis are afebrile at the time of examination. Taking the temperature reassures wrongly.

Local warmth, readily quoted as a cardinal sign of infection, was noted in only 36 % of cases. Here again, judging warmth by hand is a poorly reproducible test, and the figure reflects the examination method as much as the underlying biology. The next chapter shows that an instrumental measurement of the same parameter does far better.

What the few comparative series actually establish

Only two studies compare septic and non-septic bursitis from the same recruitment, and both are old and small.

Ho and Tice, 1979, on 30 olecranon and prepatellar bursitis cases of which 10 were septic: fever, tenderness, peri-bursal cellulitis and skin involvement overlying the bursa were more frequent in the septic cases; a high fluid white cell count, a low bursal-to-serum glucose ratio and a positive Gram stain distinguished septic from non-septic forms². The authors give no sensitivity or specificity for these signs.

Smith and colleagues, 1989, on 46 consecutive olecranon bursitis cases of which 11 were septic, provide the only quantified diagnostic value available, and it does not concern a classic clinical sign but an instrumental measurement: the skin temperature differential¹. That is the subject of the next chapter.

What is clinical examination alone worth for deciding septic or not?

The honest answer fits in one sentence: not enough to decide on its own, and that is precisely what justifies aspiration. But one measurable parameter does markedly better than clinical impression, and it has remained oddly little known.

The temperature differential: the only quantified diagnostic value available

In 1989, Smith and colleagues prospectively analysed 46 consecutive olecranon bursitis cases over one year, of which 11 proved septic and 35 non-septic. Alongside fluid analysis, they used a surface probe to measure skin temperature over the affected bursa and over the contralateral olecranon serving as a control¹.

The results are clear-cut. The mean differential was 0.7 °C in the non-septic cases against 3.7 °C in the septic ones. In all the septic cases, the differential reached or exceeded 2.2 °C, with a range of 2.2 to 5.1 °C. Taking that 2.2 °C threshold, the measurement proved 100 % sensitive and 94 % specific. The authors conclude that it appears more useful than the bursal fluid white cell count, than the predominant cell type or than the Gram stain for telling the two forms apart early¹.

Why this finding has not replaced aspiration

A sensitivity of 100 % is impressive, and that is where one has to be precise rather than enthusiastic. It rests on 11 septic cases. A sensitivity of 11 out of 11 is compatible, at 95 % confidence, with a true sensitivity going down to around 72 %: the interval is wide because the numbers are small. The figure does not say that the measurement never misses an infection, it says that it missed none out of eleven.

Three further limits are worth setting out:

• No replication. This 1989 finding has, to our knowledge, never been reproduced by an independent team on larger numbers. It has, on the other hand, been taken up as a criterion in the algorithm of Baumbach and colleagues, which keeps a pre-bursal differential above 2.2 °C among the clinical features suggesting a septic form⁹. Being taken up in an algorithm is not external validation.
• An instrument few people own. The measurement requires a contact probe applied under comparable conditions on both sides. Comparative palpation by hand is nothing like it in terms of reproducibility, and it is probably what explains the modest 36 % of “warmth” recorded in the Calgary cohort.
• A specificity of 94 % is still imperfect. Out of 35 non-septic cases, two crossed the threshold. A crystal-induced bursitis in acute flare is hot: the parameter measures inflammation, not infection.

The temperature differential is the best documented clinical sign, and it rests on eleven infected patients in a single 1989 study. That is both far better than nothing and far less than proof.

The practical conclusion: examination orients, it does not conclude

No source supports the idea that a cluster of clinical signs would allow the sample to be dispensed with when infection is plausible. The literature's constant position, from Stell in 1996⁷ to Lormeau in 2019¹⁵, is that the distinction goes through examination of the fluid.

The reviews that propose a structured care pathway say nothing different: Blackwell and colleagues build theirs from the history and the examination, with particular emphasis on the warning signs, but it is indeed fluid analysis that governs the infective branch¹¹.

What clinical examination does do well, on the other hand, is sort patients into three groups: those in whom infection is highly unlikely and where a procedure would be more risky than useful; those in whom it is possible and where a sample is called for; and those in whom it is probable and where the sample must not delay treatment. That is exactly the role a physiotherapist can hold on the front line, and it is already a great deal.

What is the place of aspiration, and when should it be avoided?

Aspiration is both the test that settles the question and a procedure with documented risk. That tension explains recommendations that appear contradictory. They are not: they simply do not address the same clinical situation.

What the fluid brings, and what the thresholds are worth

The aspirated fluid should go for direct microscopy after Gram staining, cell count and differential, crystal analysis, and aerobic and anaerobic culture. Purulent fluid, a positive Gram stain or a positive culture confirm infection.

The difficulty starts with the cell count. The proposed thresholds do not agree, and the gap between them is not marginal: it runs from about 2,000 to 50,000 white cells per microlitre depending on the source. The algorithm of Baumbach and colleagues, which is the most quoted synthesis, keeps 3,000 cells per microlitre with more than 50 % polymorphs, combined with a bursal-to-serum glucose ratio below 50 %⁹.

Two practical consequences follow from this spread. First, fluid below the threshold is not wholly reassuring when the clinical picture is worrying. Second, culture takes precedence over the cell count: it is culture that confirms, and cell-rich fluid with a negative Gram stain remains suspect.

Gram stain and culture miss a share of the infections

This is a point that practice underestimates. In the French multicentre study by Charret and colleagues, the largest contemporary series with 272 patients managed for septic bursitis of the olecranon or the patella, 43 of the 223 bursal samples remained sterile, that is 19.3 %¹⁷. An organism was identified in only 184 patients out of 272, that is 67.6 %.

Close to one septic bursitis in five has a negative bursal culture. A sterile culture does not cancel out a clear-cut clinical picture.

The reasons are known: antibiotics already started, slow-growing organisms, a poor sample. The case of Cutibacterium acnes reported by Skedros and colleagues is instructive on that last point: the culture only turned positive on the twenty-first day of incubation, whereas a first aspiration had been reported negative after 14 days²². A standard culture stopped at five days would have concluded that there was no infection.

And when the bursitis is frankly mechanical: do not aspirate

The opposite side is just as well documented. Bustamante and colleagues prospectively sampled 28 clinically aseptic olecranon bursitis cases and systematically sent the fluid to the laboratory. The result: zero positive cultures out of 28, zero positive Gram stains out of 28, a mean cellularity of 1,289 per cubic millimetre, and only two samples carrying calcium pyrophosphate crystals, without that changing what was done¹⁸. The authors conclude that routine analysis has limited clinical usefulness in this situation.

In other words: when the picture is frankly mechanical, analysis brings nothing back. And the procedure does cost something. Khodaee is explicit in his review: aspirating a microtraumatic bursitis is generally not recommended because of the risk of iatrogenic septic bursitis¹³.

The needle track, and why it matters

When aspiration is performed, it is classically done through a posterolateral approach, shifting the skin before the puncture to create a zigzag track. The logic is to avoid a short perpendicular track, which closes poorly and exposes the patient to a fistula or a chronic sinus track above the bursa. A medial approach is to be avoided because of the ulnar nerve.

This technical point explains a frequent clinical observation: an aspirated bursa that fills up again and ends up weeping through the puncture hole. It is also one of the reasons why repeated procedures are unfavourable, and why compression after aspiration is not a detail.

What are the risk factors and the natural course?

A 2024 study on nearly 55,000 subjects has renewed this chapter, and its most interesting result may be what it fails to find.

What the largest available study establishes

Shemesh and colleagues compared 10,301 patients with olecranon bursitis against 44,608 controls from a population database over the period 2005 to 2020²¹. It is by far the largest published sample on the subject.

Four associations stand out, all of small size: male sex (OR 1.406; 95 % CI 1.347 to 1.468), hyperlipidaemia (OR 1.243; 1.170 to 1.322), statin use (OR 1.130; 1.048 to 1.218) and smoking (OR 1.068; 1.020 to 1.118)²¹.

And diabetes is not associated (OR 0.979; p = 0.75), any more than cardiovascular disease or a history of stroke. That is a result worth pointing out, because diabetes appears in virtually every list of bursitis risk factors to be found online.

The factor that dominates everything, and that the study does not measure

None of these odds ratios comes close to the importance of the mechanical factor. The commonest cause remains chronic microtrauma from pressure¹³, and in the Calgary septic cohort the mean age was 44 years with 88 % men⁸, a profile that overlaps with manual trades loading the elbow. It is that exposure, unmeasurable in an administrative health database, that carries most of the risk.

The natural course

For the non-septic forms, the soundest landmark comes from the randomised trial of Kim and colleagues: 25 patients out of 30 (83 %) treated with simple compression and anti-inflammatories had resolved at four weeks¹². The natural tendency is therefore favourable.

One prognostic factor emerges from that same trial, and it is actionable: the duration of symptoms before treatment began was the only factor associated with failure at four weeks (6 weeks in case of failure against 4 weeks in case of success; p = 0.008)¹². Starting management early improves the odds, whichever treatment is chosen.

For the septic forms, the prognosis under appropriate treatment is good: in Charret's series of 272 patients, the overall failure rate was 5.9 %¹⁷. Recovery can, however, be long: Stell was already noting that healing a septic olecranon bursitis can take months⁷.

What management for a non-septic olecranon bursitis?

This is the physiotherapist's ground, and the good news is that the simplest treatment is also the one the data support best. The bad news is that the overall level of evidence remains low: a single contemporary randomised trial, two old trials, and a literature dominated by retrospective series.

The only contemporary randomised trial, and what it really says

Kim and colleagues randomised 90 patients with non-septic olecranon bursitis into three groups of 30: compression bandaging with non-steroidal anti-inflammatory drugs, aspiration alone, or aspiration followed by a corticosteroid injection. Treatment was repeated each week if the effusion recurred, and assessment ran over four weeks¹².

Three lessons, and a caveat the authors set out themselves.

First lesson: nothing is lost by starting with the simplest option. Compression bandaging with anti-inflammatories does as well as the invasive procedures at four weeks. It even does numerically better than aspiration alone, without the difference being significant.

Second lesson: the corticosteroid buys speed, not outcome. Resolution came at 2.3 weeks in the aspiration plus corticosteroid group, against 3.1 and 3.2 weeks in the other two (p = 0.015). One week gained on a condition that heals in more than eight cases out of ten.

Third lesson: the trial is small. The authors write explicitly that they were only powered to detect a 30 % difference, and offer their data as pilot data for sizing future trials¹². A difference not demonstrated is not the demonstration of no difference.

Their conclusion deserves quoting in its movement: because olecranon bursitis recurs, and because aspiration and injection can cause complications, they suggest that compression bandaging combined with a short course of anti-inflammatories offers the best balance between safety and effectiveness, as long as future trials have not shown a clear advantage to the procedures¹².

Intrabursal corticosteroids: a real disagreement in the literature

This is the point where the sources diverge most plainly, and it would be dishonest to smooth the disagreement over.

On one side, a double-blind randomised trial from 1989 on 42 patients split into four arms concludes that intrabursal injection of 20 mg of methylprednisolone is the most effective regimen for non-septic olecranon bursitis: the fastest reduction in swelling at one week, sustained improvement at six weeks, and fewer repeat aspirations at six months than in the naproxen or placebo groups⁶.

On the other, a long-term follow-up published five years earlier on 47 patients reviewed at a mean of 31 months gives the price of that effectiveness. The 25 patients treated with an injection of 20 mg of triamcinolone hexacetonide recovered quickly, generally within a week, but went on to show 3 infections, 5 skin atrophies and 7 chronic local pains. The 22 patients treated with simple aspiration recovered more slowly and with no complication at all. The authors conclude that a conservative approach is preferable since spontaneous resolution can be expected⁵.

Across 25 injected patients followed for a mean of 31 months: 3 infections, 5 skin atrophies, 7 chronic pains. Across 22 patients simply aspirated: no complication.

The systematic review comes down on the side of caution. Sayegh and Strauch, across 29 studies and 1,278 patients, find that corticosteroid injection in aseptic bursitis is associated with an increase in overall complications (p = 0.0458) and in skin atrophy (p = 0.0261), and conclude that it carries significant risks without improving the outcome¹⁰. The 2023 systematic review reaches the same practical position: reserve injection for refractory forms because of the higher complication rate¹⁹.

What rehabilitation actually does

None of these studies assesses a physiotherapy intervention on its own: that is a gap in the literature which has to be named. What is assessed is a package of “compression and anti-inflammatories”, whose compression component falls directly to the physiotherapist. Clinical reasoning is therefore built from what is supported, while saying so.

Removing the pressure is the causal measure. As long as the elbow keeps bearing on a hard surface, the bursa keeps being irritated. That means concrete work analysing situations: workstation and armrests, sleeping position, resting on the steering wheel, occupational movements on the floor, pressure during transfers in a patient who uses their elbows. Padded elbow protection has an obvious logic here, even if its own benefit has not been isolated in a trial.

Compression is the documented active component of the best-performing arm in Kim's trial. A compression bandage worn continuously for the first few days, watching vascular and skin tolerance, is justified.

Movement should be maintained, not forced. The elbow does not stiffen spontaneously in an uncomplicated bursitis, since movement is preserved from the outset. The aim is to maintain it, in particular the terminal flexion that the mass mechanically obstructs, without trying to gain range by forcing on an inflamed bursa.

What to avoid. Firm massage over the bursa, local friction techniques and local heat have no rationale on a distended, inflamed bursa, and are strictly to be avoided as long as infection has not been ruled out. If infection is in doubt, no local technique should be applied before the medical opinion.

Modalities and level of evidence

How is a septic olecranon bursitis treated?

This chapter is not the physiotherapist's direct concern, but knowing it changes how the patient is supported, how the timescales are explained, and how a treatment that goes off course is spotted.

Antibiotics first, surgery afterwards

The contemporary position is clear: medical treatment is enough in the majority of cases. In the French multicentre study on 272 patients, the success rate was equivalent between the medically managed group and the operated group, and only 26 % of patients were treated surgically¹⁷. Baumbach and colleagues had already concluded that the available data did not support the central European concept of immediate bursectomy for septic bursitis, and recommended a conservative regimen⁹.

Antibiotic treatment targets S. aureus first line, since it accounts for about 80 % of the organisms¹⁵. An initial intravenous route was preferred in the case of fever (p = 0.003) or extensive cellulitis (p = 0.002) in the French series, and concerned 41 % of patients¹⁷.

Treating early shortens the treatment. This is an old but robust observation: by culturing the fluid of 25 patients serially, Ho and Su showed that the time needed to sterilise the bursa correlated with the duration of symptoms before diagnosis (r = 0.68; p < 0.001). In patients treated within two weeks of symptom onset, sterility was obtained in under a week; beyond that, the response was delayed⁴. A delayed diagnosis therefore does not merely put healing back, it lengthens the treatment needed.

Duration: a second disagreement worth knowing

Two pieces of good-quality data point in different directions, and here again the key is that they do not describe the same situation.

On medical treatment, Charret and colleagues observe that failures were more frequent when antibiotic treatment lasted under 14 days (p = 0.02), in the surgical group as in the medical one, and conclude that a duration under 14 days deserves particular attention¹⁷.

After bursectomy, the randomised trial of Uçkay and colleagues on 164 inpatients shows that 7 days of oral antibiotics after excision of the bursa are enough, with 10 % failures in the single-stage arm¹⁴. The difference is explained: when the infected focus is removed, the bacterial load to treat is no longer the same.

That trial brings a second useful result, specific to the elbow: single-stage bursectomy with primary closure gave fewer wound dehiscences than the two-stage technique for elbow bursitis, 1 out of 66 against 9 out of 64 (p = 0.03), with a shorter hospital stay and a lower cost¹⁴.

Why surgery is not a trivial matter

The series of Pohl and colleagues, covering 61 open bursectomies, gives figures that invite restraint: 29.5 % postoperative complications, including 10 fluid collections, 4 ulnar nerve paraesthesias and 3 wound dehiscences, with a mean healing time of 17.4 days²⁰.

Above all, the functional result differs by form: aseptic patients significantly improved their physical score (41.5 to 46.8; p < 0.001), septic patients did not (43.6 to 40.5; p = 0.277), a septic diagnosis being associated with a poorer result on multivariate analysis (estimate −8.40; 95 % CI −14.42 to −2.38; p = 0.008)²⁰.

What the physiotherapist follows up after a septic bursitis

Three concrete points. First, recovery is long: Stell already pointed this out, it can take months⁷. Announcing that timescale saves the patient from concluding that it has failed.

Next, the posterior elbow scar is a fragile area, over which the skin tightens with every flexion. With 17.4 days of mean healing time and a documented risk of dehiscence²⁰, progression into flexion must be conducted while watching the scar, and elbow pressure prohibited for a long time.

Lastly, ulnar nerve paraesthesia is a known postoperative complication, found in 4 patients out of 61²⁰. Tingling along the ulnar border of the hand after bursectomy is not a trivial matter: it must be reported.

What to do about a posterior swelling of the elbow that is not a bursitis?

Not every posterior mass of the elbow is a hygroma, and two of the costliest errors are errors of over-confidence in the obvious diagnosis.

Triceps rupture, the error that costs a repair

This is the differential diagnosis to know first, because it is functionally serious and because it is regularly taken for a bursitis. The mechanism is the same in both cases: an elbow injury followed by posterior swelling.

The sign that separates them is simple and often left out: elbow extension against resistance. Shivdasani and colleagues report the case of a 53-year-old man, a weight trainer, who felt a tearing sensation with immediate pain and swelling of the posterior aspect of the right elbow, and showed weakness on extension against resistance. The initial diagnosis was olecranon bursitis, and three aspirations were performed over several weeks before an orthopaedic opinion corrected the diagnosis, five weeks after the injury²³.

The authors point precisely to the cause of the error: the extension-against-gravity test, specific to the function of the triceps tendon, is frequently left out because of the pain and the swelling²³. In other words: it is precisely because the elbow is swollen and painful that the test is not done, and it is precisely there that it must be.

This is no curiosity. In the surgical series of Pohl and colleagues, 20 patients out of 61 operated on for a bursectomy, that is 32.8 %, had a triceps repair at the same time, of which 8 were complete ruptures²⁰. An associated triceps lesion is therefore frequent in patients who end up in theatre.

For any posterior swelling of the elbow after trauma: test active extension against gravity, then against resistance. Weakness does not belong to the picture of bursitis.

Cubital tunnel syndrome

Compression of the ulnar nerve at the elbow can produce a swelling or a thickening of the posteromedial region, and above all it shares with bursitis both the site and the context of prolonged elbow pressure. The difference is neurological: paraesthesias along the ulnar border of the hand, of the fourth and fifth fingers, worsened by prolonged elbow flexion and at night, with an intrinsic motor deficit at an advanced stage.

Two situations demand that it be considered explicitly. The first is a posterior swelling accompanied by distal signs: bursitis gives no symptoms in the hand. The second is the appearance of paraesthesias after a bursectomy, ulnar paraesthesia being a documented postoperative complication, found in 4 patients out of 61 in Pohl's series²⁰.

The subject is covered in detail in our article on cubital tunnel syndrome (ulnar compression at the elbow).

Crystal-induced and inflammatory causes

Gout and chondrocalcinosis produce olecranon bursitis, sometimes as the presenting feature, and rheumatoid arthritis classically comes with nodules and bursitis in this region. Ho and Tice were already noting it: rheumatoid arthritis and gout can be accompanied by a non-septic bursitis².

The trap is a major one, and the next chapter devotes a case to it. Finding crystals does not close the discussion, because a bursa carrying crystals can also be infected. Bustamante and colleagues did in fact find calcium pyrophosphate crystals in 2 samples out of 28 from clinically aseptic bursitis without that changing management¹⁸: crystals are frequent and rule nothing out.

The other causes, rarer but worth knowing

What do concrete clinical cases teach us ?

Three published cases, chosen because each illustrates a different mechanism of error: the examination not done, the diagnosis that stops at the first result, and the culture stopped too soon.

Case 1: the bursitis that was a triceps rupture

A 53-year-old man, a recreational weight trainer. A tearing sensation with immediate pain and swelling of the posterior aspect of the right elbow. The general practitioner settled on olecranon bursitis and performed three aspirations over several weeks. The patient finally consulted an orthopaedic surgeon: complete rupture of the triceps tendon, repaired surgically five weeks after the injury. At one year, strength and range equal to the opposite side, a Mayo score of 100, weight training resumed without restriction at 16 weeks²³.

What the case teaches. The final result is excellent, which might make the delay seem less important. That would be a misreading: repairing triceps ruptures is all the simpler for being early, and five weeks is a delay that exposes the tendon to retraction. Above all, the error was avoidable by a ten-second test. The swelling that makes testing difficult is exactly the situation where the test counts most.

Case 2: the crystals that masked a mycobacterium

A 65-year-old immunocompetent man. Swelling of the right elbow recurring two years after a first bursectomy, with no trauma, no pain, no redness, no systemic sign. Aspiration found monosodium urate crystals: a diagnosis of gout was made, and the patient received indometacin then allopurinol. Despite treatment, the fluid re-formed.

The initial tests were reassuring: a search for acid-fast bacilli negative on direct microscopy, bacteriological and mycological cultures negative. It was the mycobacterial cultures, positive at ten days, that corrected the diagnosis, confirmed on a second aspiration: Mycobacterium kansasii. Treatment with azithromycin, ethambutol and rifampicin, then bursectomy at two months, with eight months of antibiotics in total. Ethambutol had to be stopped at six months for an optic neuropathy, with persisting visual sequelae²⁴.

What the case teaches. Three things, and all of them transfer.

• A positive result is not a complete diagnosis. Finding urate crystals stopped the reasoning. The bursa was indeed gouty, and it was also infected.
• The warning sign was not inflammatory, it was in the course. No fever, no redness, no pain. What was wrong was that the fluid kept coming back despite treatment appropriate to the diagnosis made. A bursitis that does not respond to treatment should make you reopen the diagnosis, not raise the dose.
• Infections with slow-growing organisms are seen in an elbow already operated on or already aspirated. The literature on atypical mycobacteria of the olecranon bursa describes diagnostic delays of several months, with sparse pictures often wrongly treated as aseptic bursitis, including with repeated injections that make the infection worse.

Case 3: the culture that was not long enough

A 53-year-old right-handed man, with pain and swelling at the back of the right elbow, no diabetes, no drug use, and no history whatsoever of injection into the bursa. A history, four years earlier, of septic bursitis with Staphylococcus aureus sensitive to meticillin after an abrasion, cured by debridement and a week of antibiotics.

The current episode resisted. The aspiration of 1^st October 2017 stayed sterile after 14 days. A sample from 6 October eventually grew: Cutibacterium acnes, on the twenty-first day of incubation. It took five debridements in three weeks, plus curettage in the clinic, intravenous antibiotics then three months of oral amoxicillin, to eradicate an infection with probable osteitis of the olecranon. No recurrence at four and a half years²².

What the case teaches. C. acnes is usually set aside as a contaminant, and it grows slowly: a standard culture reported at five days would have missed it, and a culture reported at 14 days did miss it. When a bursitis fails to heal while the cultures are negative, the question is not only “is it infected?” but “have we looked long enough?”. This case is also a reminder that recurrence at the same site, four years after a first episode, is a risk factor in itself.

How is this applied in practice?

What this article changes, session by session, for a physiotherapist who sees a swollen elbow come in.

The five minutes that count, at first contact

1. Look at the skin before touching. A graze, a scab, a wound, an ulcer, redness spreading past the outline of the swelling. Peri-bursal cellulitis extending beyond the bursa is the most useful clinical sign in community practice.
2. Compare the warmth of the two elbows with the back of the hand, staying aware that this does not have the performance of Smith's probe: clearly asymmetrical warmth counts, doubtful warmth is not reassuring.
3. Move the elbow passively. Mid-range free and painless: it really is the bursa. Pain throughout range: think joint.
4. Test active extension against gravity, then against resistance. Ten seconds, and it is what stops you missing a triceps rupture²³.
5. Ask about the hand. Paraesthesias of the last two fingers take you outside the picture.

What to tell the patient, and how to say it

The most important instruction is not about an exercise but about a timescale. A patient leaving with a bursitis judged to be mechanical must know what should bring them back without delay: redness that spreads, pain that increases, fever or chills, and generally any worsening within 48 hours.

Two received ideas also need defusing. The first is that the size of the lump measures the severity: that is false, a large painless bursa is often less worrying than a small very painful one. The second is that the bursa has to be “emptied”: Kim's trial shows that aspiration alone does numerically worse than simple compression¹², and the procedure is not neutral.

The management plan for a mechanical bursitis

Three errors not to make

Treating an elbow locally when infection has not been ruled out. Massage, frictions, ultrasound, heat: on a potentially infected bursa, these techniques have no rationale and may help to spread it. In case of doubt, you refer and you do not treat locally.

Being reassured by the absence of fever. This is the most predictable error, and the figure is worth remembering: only 20 % of confirmed septic bursitis cases had a temperature of 37.8 °C or above⁸.

Treating a recurrent bursitis as a mechanical inevitability. Recurrence after a procedure, especially in an elbow already operated on or aspirated, should raise the idea of a slow-growing infection before being put down to pressure²⁴.

Frequently asked questions

Bibliography

Each reference has been checked individually: existence of the PMID and reading of the abstract to verify that the source really establishes what is attributed to it. 24 sources. Click a superscript note marker in the text: the bibliography opens and highlights the source.

1. Smith DL, McAfee JH, Lucas LM, Kumar KL, Romney DM (1989). Septic and nonseptic olecranon bursitis. Utility of the surface temperature probe in the early differentiation of septic and nonseptic cases. Archives of Internal Medicine. PMID 2742432. doi:10.1001/archinte.149.7.1581.
2. Ho G, Tice AD (1979). Comparison of nonseptic and septic bursitis. Further observations on the treatment of septic bursitis. Archives of Internal Medicine. PMID 508024. doi:10.1001/archinte.1979.03630480051017.
3. Ho G, Tice AD, Kaplan SR (1978). Septic bursitis in the prepatellar and olecranon bursae: an analysis of 25 cases. Annals of Internal Medicine. PMID 666181. doi:10.7326/0003-4819-89-1-21.
4. Ho G, Su EY (1981). Antibiotic therapy of septic bursitis. Its implication in the treatment of septic arthritis. Arthritis and Rheumatism. PMID 7259803. doi:10.1002/art.1780240707.
5. Weinstein PS, Canoso JJ, Wohlgethan JR (1984). Long-term follow-up of corticosteroid injection for traumatic olecranon bursitis. Annals of the Rheumatic Diseases. PMID 6696516. doi:10.1136/ard.43.1.44.
6. Smith DL, McAfee JH, Lucas LM, Kumar KL, Romney DM (1989). Treatment of nonseptic olecranon bursitis. A controlled, blinded prospective trial. Archives of Internal Medicine. PMID 2684077. doi:10.1001/archinte.1989.00390110093021.
7. Stell IM (1996). Septic and non-septic olecranon bursitis in the accident and emergency department: an approach to management. Journal of Accident & Emergency Medicine. PMID 8894865. doi:10.1136/emj.13.5.351.
8. Laupland KB, Davies HD (2001). Olecranon septic bursitis managed in an ambulatory setting. The Calgary Home Parenteral Therapy Program Study Group. Clinical and Investigative Medicine. PMID 11558851.
9. Baumbach SF, Lobo CM, Badyine I, Mutschler W, Kanz KG (2014). Prepatellar and olecranon bursitis: literature review and development of a treatment algorithm. Archives of Orthopaedic and Trauma Surgery. PMID 24305696. doi:10.1007/s00402-013-1882-7.
10. Sayegh ET, Strauch RJ (2014). Treatment of olecranon bursitis: a systematic review. Archives of Orthopaedic and Trauma Surgery. PMID 25234151. doi:10.1007/s00402-014-2088-3.
11. Blackwell JR, Hay BA, Bolt AM, Hay SM (2014). Olecranon bursitis: a systematic overview. Shoulder & Elbow. PMID 27582935. doi:10.1177/1758573214532787.
12. Kim JY, Chung SW, Kim JH, Jung JH, Sung GY, Oh KS, Lee JS (2016). A randomized trial among compression plus nonsteroidal antiinflammatory drugs, aspiration, and aspiration with steroid injection for nonseptic olecranon bursitis. Clinical Orthopaedics and Related Research. PMID 26463567. doi:10.1007/s11999-015-4579-0.
13. Khodaee M (2017). Common superficial bursitis. American Family Physician. PMID 28290630.
14. Uçkay I, von Dach E, Perez C, Agostinho A, Garnerin P, Lipsky BA, Hoffmeyer P, Pittet D (2017). One- vs 2-stage bursectomy for septic olecranon and prepatellar bursitis: a prospective randomized trial. Mayo Clinic Proceedings. PMID 28602435. doi:10.1016/j.mayocp.2017.03.011.
15. Lormeau C, Cormier G, Sigaux J, Arvieux C, Semerano L (2019). Management of septic bursitis. Joint Bone Spine. PMID 31615686. doi:10.1016/j.jbspin.2018.10.006.
16. Nchinda NN, Wolf JM (2021). Clinical management of olecranon bursitis: a review. The Journal of Hand Surgery. PMID 33840568. doi:10.1016/j.jhsa.2021.02.006.
17. Charret L, Bart G, Hoppe E, Dernis E, Cormier G, Boutoille D, Le Goff B, Darrieutort-Laffite C (2021). Clinical characteristics and management of olecranon and prepatellar septic bursitis in a multicentre study. The Journal of Antimicrobial Chemotherapy. PMID 34293150. doi:10.1093/jac/dkab265.
18. Bustamante S, Boin M, Dankert J, Adekanye D, Virk MS (2022). The utility of routine cultures, cell count, and crystal evaluation of aspirate from aseptic olecranon bursitis. JSES International. PMID 35813143. doi:10.1016/j.jseint.2022.02.012.
19. Kaur IP, Mughal MS, Aslam F, Schram J, Bansal P (2023). Non-surgical treatment of aseptic olecranon bursitis: a systematic review. Reumatologia Clinica. PMID 37945181. doi:10.1016/j.reumae.2023.05.004.
20. Pohl NB, Brush PL, Toci GR, et al. (2023). Clinical outcomes following open olecranon bursa excision for septic and aseptic olecranon bursitis: an observational study. Cureus. PMID 37724223. doi:10.7759/cureus.43696.
21. Shemesh S, Itzikovitch R, Atzmon R, Kadar A (2024). Risk factors for the development of olecranon bursitis: a large-scale population-based study. Journal of Clinical Medicine. PMID 39768728. doi:10.3390/jcm13247801.
22. Skedros JG, Finlinson ED, Luczak MG, Cronin JT (2023). Septic olecranon bursitis with osteomyelitis attributed to Cutibacterium acnes: a case report. Cureus. PMID 36879721. doi:10.7759/cureus.34563.
23. Shivdasani K, Scheidt M, Anderson J, Okafor L, Garbis N, Salazar D (2024). Full-thickness triceps tears misdiagnosed as olecranon bursitis. JSES Reviews, Reports, and Techniques. PMID 38706666. doi:10.1016/j.xrrt.2024.02.002.
24. Guru S, Saroya J, Guru N, Bai P (2025). Mycobacterium kansasii chronic olecranon bursitis: a rare case report and literature review. IDCases. PMID 40726639. doi:10.1016/j.idcr.2025.e02326.