Can a Shoulder Replacement Get Infected?

Yes, though it is uncommon. Between 1 and 4 out of every 100 shoulder replacements develop a deep infection around the implant. Most are caused by an ordinary skin bacterium (Cutibacterium acnes) that produces no fever, no redness, and no drainage, which is what makes this infection both difficult to prevent and difficult to recognize.

Three things patients should know:

• What is a periprosthetic joint infection? An infection around an orthopedic joint replacement. Bacteria colonize the surface of the implant itself, not the skin or the incision. Once they establish themselves on metal and plastic, antibiotics alone almost never clear them.
• How do we prevent it? Through a series of small measures that begin well before you reach the operating room. The two with the strongest scientific evidence are the antibiotic given before the incision and allowing three months to pass after a cortisone injection before surgery.
• How do we find it? Slowly, and by assembling several tests rather than relying on one. No single blood test, x-ray, or fluid sample settles the question, and the reason is worth understanding.

What the infection actually is

A shoulder replacement introduces metal and plastic into the joint. If bacteria reach those surfaces, they adhere and build a protective layer over themselves, a structure called a biofilm. Beneath that layer the bacteria are very difficult for either the immune system or antibiotics to reach.

This is why a deep implant infection nearly always means further surgery to remove or exchange components. Antibiotics alone do not penetrate the biofilm, and your body cannot clear the organism on its own.

Why this particular organism is so troublesome

The bacterium responsible in most cases is Cutibacterium acnes, the same organism involved in ordinary acne. It resides in the sebaceous glands of the skin, and those glands are unusually concentrated across the chest, shoulder, and upper back. This leads to three challenges.

1. It is already present. It lives in the skin of almost all patients. A review of 80 studies found it on shoulder skin before any washing roughly half the time, in people who were entirely healthy.
2. Surface cleaning solutions cannot reach all of it. Antiseptic solutions clean the surface of the skin. The bacteria sit in the glands beneath it and repopulate the surface within the hour.
3. It causes almost no immune response. No fever, no pus, frequently no abnormal blood work. What it produces instead is pain and stiffness that fail to resolve, which is easily mistaken for a mechanical problem with the implant.

How we prevent it

Before you reach the operating room

Timing after a cortisone injection. Corticosteroid injected into a joint suppresses the local immune response, and this can increase infection. Analyses of large insurance and Medicare databases have consistently found higher rates of infection and revision when shoulder replacement follows an injection by less than three months, with odds roughly doubled. Beyond three months, the association is much weaker.

Limiting the number of cortisone injections to one. Cortisone is not dangerous, and this is not an argument against using it. However, recent data shows that 2 or more injections in the year before a shoulder replacement increase the risk of a periprosthetic infection.

Washing your shoulder at home before surgery. Most surgeons send patients home with an antiseptic wash to use for several days beforehand. Which one you are handed matters more than it might appear.

Both chlorhexidine soap and benzoyl peroxide are often sent home with patients or purchased over the counter to wash the shoulder once a day for several days before surgery. Benzoyl peroxide is stronger and has better evidence to support it but can cause skin irritation.

Your general health. Nonmodifiable risk factors for infection include previous surgery and male gender. We can’t change those, but we can change some things. Modifiable risk factors include smoking, testosterone use, and poorly controlled diabetes. These are taken into consideration prior to surgery, with many surgeons working with patients to reduce risk by treating these preoperatively.

In the operating room

The antibiotic given before the incision. This is the single most important thing we do. All patients receive IV antibiotics within one hour of surgery. Cefazolin is the best antibiotic for this — a study of 7,713 shoulder replacements at one institution showed that patients who received cefazolin had roughly two-thirds less infection than those who received a substitute.

The patients who receive a substitute almost always have a penicillin allergy in their chart, which is often inaccurate. If your penicillin allergy dates to childhood and has never been confirmed, it is worth asking about. This is one of the few points in the process where a patient asking a single question can measurably reduce their own risk.

Skin preparation. The shoulder is painted with an antiseptic solution before draping. Chlorhexidine combined with alcohol removes more total bacteria from shoulder skin than iodine-based solutions, which is why it has become standard. Adding hydrogen peroxide appears to help further when applied to intact skin before the incision.

Opening the skin. Some surgeons make the incision with electrocautery rather than a scalpel, on the reasoning that heat sterilizes the cut edge. It does. Whether that advantage persists through the remainder of the operation is less certain.

Limiting exposure. Longer operations mean more time for contamination, and more handling of tissue means more opportunity for the organism to travel from the cut skin edge onto gloves, instruments, and eventually the implant. Efficiency is itself a prevention measure.

Antibiotic powder in the wound. Some surgeons place vancomycin powder in the wound before closing. The evidence here is genuinely divided, and the largest study, covering more than 28,000 replacements, found no benefit.

Other things. Washing the wound with iodine or chlorhexidine-based solutions, peroxide, frequently changing gloves, and minimizing implant contact with skin are all frequently used measures to help reduce rates of infection.

What this means

No single item on this list is decisive. Infection after shoulder replacement was uncommon before most of these measures existed and remains uncommon now.

Prevention in this setting is a matter of stacking small advantages rather than deploying one intervention that solves the problem.

Prevention, at a glance

Measure	What it does	Strength of evidence
Cefazolin before the incision	Kills bacteria before they reach the implant	Strong
Testing an unconfirmed penicillin allergy	Permits the more effective antibiotic to be used	Strong
Allowing 3 months after a cortisone injection	Avoids operating on a recently immunosuppressed joint	Strong
Smoking cessation, glycemic control	Improves healing and immune function	Strong
Benzoyl peroxide wash at home	Penetrates the sebaceous glands where the organism lives	Good, may cause skin irritation
Chlorhexidine wash at home	Cleans the skin surface	Good
Chlorhexidine-alcohol prep in the OR	Reduces overall bacterial load at the incision	Good
Hydrogen peroxide before incision	Extends reach into the skin	Moderate
Electrocautery incision	Sterilizes the cut dermal edge	Moderate
Vancomycin powder in the wound	Delivers antibiotic directly into the joint	Moderate

How we diagnose an infection after a shoulder replacement

This can sometimes be very difficult, and where there’s a difference between bacteria living in or around the shoulder and a true infection.

Growing a bacterium from a sample is not the same as having an infection. This organism lives on everyone. It is recovered from operating room air roughly 10% of the time and from deep tissue during routine first-time shoulder replacements about a third of the time, in patients who never develop any problem. A single positive bacterial culture is therefore a clue rather than a verdict, and the entire diagnostic process is built around separating a true infection from an organism that simply happened to be present.

What patients notice

In many cases, the infection isn’t subtle. Redness, swelling, wound drainage, fevers, and extreme pain can all point to an acute infection with C acnes or something different, like MRSA.

In other cases, pain and stiffness, and the pattern of the symptoms, matter more. Pain that never resolved after the original surgery, or that resolved and returned months later, is the characteristic. Stiffness that will not yield counts as well. The other findings people associate with infection (redness, warmth, swelling) are usually absent, which is why these patients are often worked up for implant loosening or instability first.

Blood tests

The standard markers, ESR and CRP, measure inflammation. In some cases, a shoulder infection generates so little of it that the tests frequently remain normal. In one series of revision shoulder surgeries, the better performing test of the two still missed roughly a third of proven infections, and other markers have performed worse, including one that identified only 1 infection in 7.

A normal result does not clear you. That is the single most important thing to understand about blood work in this setting.

X-rays

More useful than their simplicity suggests. The finding that carries the most weight is loosening of the humeral stem, the component seated in the arm bone. That component rarely loosens on its own in the absence of infection, so loosening on an X-ray moves infection substantially up the differential.

Aspiration: drawing fluid from the joint with a needle

This sounds as though it ought to settle the question. It does not. This is where fluid is drawn from the shoulder and sent off for various tests.

In a series of 106 shoulder aspirations performed before revision surgery, 24 shoulders proved to be infected, and the fluid tests identified only 10 of them. It therefore misses roughly two of every three infections. When it does return positive, however, it is correct about 98% of the time.

That asymmetry is the useful part. A positive aspiration carries considerable weight. A negative one carries very little and should never be treated as clearance.

Cultures taken during surgery

At the time of revision surgery, at least five tissue specimens are taken from separate locations and sent to the laboratory for bacterial culture growth, especially when the cause for revision is uncertain.

Two pieces of information separate a true infection from a specimen that merely picked up a passenger:

• How many specimens grew. One positive out of five describes a very different situation from five out of five.
• How quickly they grew. Specimens that grow quickly (within the first few days) are a sign of a true infection, while specimens that grow later may be contaminants.

How much of this is noise

In one carefully controlled study, surgeons took tissue specimens during 117 routine shoulder operations on patients with no evidence of infection. About 1 in 5 grew bacteria. The same team cultured 54 sterile sponges that never touched a patient, and about 1 in 8 of those grew bacteria as well.

That study was small, and it does not establish that a positive culture in a symptomatic patient is meaningless. What it does establish is that a single positive result, standing alone, cannot be trusted, which is precisely why the number of positive specimens and the speed of growth carry as much weight as the result itself.

In 2018, an international consensus group produced criteria specific to the shoulder. Three findings confirm infection outright: a sinus tract draining pus from the implant to the skin, visible pus within the joint, or two specimens growing the same virulent organism. Everything else earns points, and the total yields a verdict of probable, possible, or unlikely. It is a formal acknowledgment that no single finding decides the matter.

Tests, at a glance

Test	How often it identifies a true infection	Reliability when positive	Bottom line
Blood tests (CRP, ESR)	About 2 in 3	Frequently wrong	Normal results do not clear you
X-ray showing a loose humeral stem	Variable	Strongly suggestive	Inexpensive and underrated
Needle aspiration of the joint	About 1 in 3	Almost always correct	A positive means a great deal, a negative very little
Five tissue specimens at surgery	The gold standard	The gold standard	Read alongside how many grew and how quickly

---

Midhat Patel, MD is a fellowship-trained shoulder and elbow surgeon at Banner Health in Phoenix and Scottsdale, Arizona, and a Clinical Assistant Professor of Orthopaedic Surgery at the University of Arizona College of Medicine – Phoenix.

Sources

Figures quoted above, with the underlying study detail.

• Antibiotic choice. Marigi EM, et al. Antibiotic prophylaxis with cefazolin is associated with lower shoulder periprosthetic joint infection rates than non-cefazolin alternatives. J Bone Joint Surg Am. 2022;104(10):872–880. Single-institution registry, 7,713 primary shoulder arthroplasties, minimum 2-year follow-up. Infection rate 1.3%. Cefazolin associated with 69% lower all-cause infection risk and 78% lower C. acnes infection risk. Vancomycin hazard ratio 2.32 (95% CI, 1.22–4.40); clindamycin 5.07 (95% CI, 2.83–9.05). doi:10.2106/JBJS.21.00445
• Penicillin allergy testing. Wyles CC, et al. 2019 John Charnley Award. Bone Joint J. 2019;101-B(6 Supple B):9–15. Of 2,576 allergy-tested patients, 2,493 (97%) cleared for cephalosporins.
• Cefazolin despite a documented cephalosporin allergy. Jolissaint JE, et al. J Bone Joint Surg Am. 2026. 482 patients with documented cephalosporin allergy given cefazolin; 0.0% allergic reactions versus 0.51% (4 of 786) on alternative antibiotics.
• Cortisone timing, three-month interval. Werner BC, et al. The timing of elective shoulder surgery after shoulder injection affects postoperative infection risk in Medicare patients. J Shoulder Elbow Surg. 2016;25(3):390–397. Arthroplasty within 3 months of injection: infection 3.0% at 3 months (OR 2.0, p = .007) and 4.6% at 6 months (OR 2.0, p = .001) versus matched controls. No association when surgery occurred more than 3 months after injection. doi:10.1016/j.jse.2015.08.039
• Cortisone timing, revision endpoint. Stadecker M, et al. Bone Joint J. 2022;104-B(5):620–626. 4,252 patients; injection within 3 months carried odds ratio 2.61 for all-cause revision (95% CI, 1.77–3.28), with 71.4% of revisions attributed to infection. doi:10.1302/0301-620X.104B5.BJJ-2021-0024.R3
• Cortisone timing, narrower window. Baksh N, et al. Arch Orthop Trauma Surg. 2023. 25,422 total shoulder arthroplasties stratified into 0–4, 4–8, and 8–12 week intervals; significant increase confined to the 0–4 week group (OR 2.29 at 1 year; 95% CI, 1.19–3.99). Listed for completeness; the three-month interval above reflects the larger body of evidence and consensus guidance.
• Meta-analysis of injection timing. Schoell K, et al. J Shoulder Elbow Surg. 2024. 7 studies, 136,233 patients; overall odds ratio 1.13 (95% CI, 1.06–1.19), with rates rising as injections approached the date of surgery.
• Patient risk factors. Seok HG, et al. J Clin Med. 2022. Meta-analysis, 8 studies, 420 infected cases and 28,464 controls. · Richards J, et al. Clin Orthop Relat Res. 2014;472(9):2809–2815. 3,906 primary shoulder arthroplasties; male sex hazard ratio 2.59 (95% CI, 1.27–5.31). doi:10.1007/s11999-014-3696-5 · Kunutsor SK, et al. PLoS One. 2016. 66 studies, 512,508 patients; smoking relative risk 1.83 (95% CI, 1.24–2.70).
• Home skin washes. Sewpaul Y, et al. Am J Sports Med. 2024. Network meta-analysis, 17 randomized trials, 1,350 patients. Benzoyl peroxide the only agent significantly reducing positive C. acnes cultures versus placebo or soap and water (odds ratio 0.12; 95% CI, 0.04–0.36).
• Benzoyl peroxide versus chlorhexidine. Kolakowski L, et al. Neer Award 2018. J Shoulder Elbow Surg. 2018. 80 patients randomized, contralateral shoulder as control.
• Operating room skin preparation. Saltzman MD, et al. J Bone Joint Surg Am. 2009;91(8):1949–1953. 150 patients randomized. Positive cultures after preparation: 7% chlorhexidine-alcohol, 19% iodine-alcohol, 31% povidone-iodine. doi:10.2106/JBJS.H.00768
• Hydrogen peroxide, four trials. Chalmers PN, et al. JSES. 2019;28(8):1554–1561 (positive, pre-incision). Stull JD, et al. JSES. 2020;29(2):212–216 (positive, pre-incision, 17.1% vs 34.2%). Grewal G, et al. JSES. 2021;30(8):1827–1833 (negative, applied to dermis after incision). Crutcher WL, et al. JSES. 2024;33(9):1905–1908 (negative, repopulation at 60 minutes 78% in both arms).
• Electrocautery versus scalpel. Kim HM, et al. J Shoulder Elbow Surg. 2024;33(1):6–13. 64 patients randomized. Dermal edge cultures positive in 31% of the scalpel group and 0% of the cautery group; glove cultures identical later in the case. doi:10.1016/j.jse.2023.07.014
• Spread from the cut edge. Moor BK, et al. J Shoulder Elbow Surg. 2021;30(7):1537–1543. Randomized, 108 patients; subcutaneous povidone-iodine reduced contamination of gloves and retractors. doi:10.1016/j.jse.2020.11.018
• Vancomycin powder, positive study. Garofalo R, et al. J Shoulder Elbow Surg. 2023;32(8):1638–1644. Retrospective, 422 treated versus 405 controls, minimum 12 months; 0 versus 13 infections. doi:10.1016/j.jse.2023.02.129
• Vancomycin powder, larger negative study. Joshi T, et al. J Shoulder Elbow Surg. 2025. 28,098 shoulder arthroplasties, propensity-matched; no reduction at 90 days or 2 years.
• Prevalence of the organism. Razi A, et al. J Shoulder Elbow Surg. 2024. Review of 80 studies. C. acnes in operating room air mean 10%, on unprepared skin mean 47%, in deep tissue at primary arthroplasty mean 29%. Unexpected positive cultures not associated with worse outcomes.
• Sterile sponge control study. Mook WR, et al. J Bone Joint Surg Am. 2015;97(12):957–963. 117 open shoulder surgeries with no prior surgery or suspicion of infection. At least one positive tissue culture in 20.5%; positive growth in 13.0% of 54 sterile control sponges (p = 0.234). doi:10.2106/JBJS.N.00784
• False positive rates. McCarroll TR, et al. J Shoulder Elbow Surg. 2020. 95 patients; false-positive rate 17.0% open, 10.4% arthroscopic.
• Serum markers. Siegert P, et al. Arch Orthop Trauma Surg. 2021. 136 revision shoulder arthroplasties. Optimized CRP cutoff of 7.2 mg/L gave 69% sensitivity and 74% specificity.
• Interleukin-6. Villacis D, et al. J Bone Joint Surg Am. 2014;96(1):41–45. 34 patients; sensitivity 0.14, specificity 0.95. doi:10.2106/JBJS.L.01634
• Aspiration accuracy. Hecker A, et al. J Shoulder Elbow Surg. 2020;29(3):516–520. 106 aspirations; 24 shoulders infected, only 10 (42%) yielded organisms. Sensitivity 33%, specificity 98%. doi:10.1016/j.jse.2019.07.016
• Alpha-defensin. Ecker NU, et al. Clin Orthop Relat Res. 2019;477(7):1712–1718. 105 patients, 24 infected. Sensitivity 75% (95% CI, 53–90), specificity 96% (95% CI, 90–99). doi:10.1097/CORR.0000000000000762
• Radiographic findings. Pottinger P, et al. J Bone Joint Surg Am. 2012;94(22):2075–2083. 193 revision shoulder arthroplasties. Male sex, humeral osteolysis, and cloudy fluid each associated with more than a six-fold increase in positive culture. doi:10.2106/JBJS.K.00861
• Speed of culture growth. Frangiamore SJ, et al. J Bone Joint Surg Am. 2015;97(14):1149–1158. 46 revision cases. True positives grew at a median of 5 days (IQR 4–7), probable contaminants at 9 days (IQR 6–12), p = 0.002. None of the 37 true positives appeared after day 11; 44% of contaminants did. doi:10.2106/JBJS.N.00881
• The 2018 criteria. Garrigues GE, et al. J Shoulder Elbow Surg. 2019;28(6):S8–S12. doi:10.1016/j.jse.2019.04.034
• Criteria applied in practice. Crook BS, et al. J Shoulder Elbow Surg. 2025. 386 revision shoulder arthroplasties; 38 (9.8%) definite infection.
• Infection rate in primary replacement. Wright JO, et al. J Shoulder Elbow Surg. 2024;33(3):618–627. doi:10.1016/j.jse.2023.10.019
• Ricchetti ET, Patel M, Sogbein O, Namdari S. Preventative strategies and diagnostic evaluation of postoperative infection including prosthetic joint infection. In: Rockwood and Matsen’s The Shoulder, 7th ed. Elsevier; 2025:Ch 58.