TRT and knee replacement: a 13,000-patient study just flagged three serious risks.

The study: what it was and how it was run

The paper at the center of this is a retrospective cohort study by Omurzakov and colleagues at the Hospital for Special Surgery (HSS) in New York, published in The Journal of Arthroplasty in late 2025 and featured at the American Academy of Orthopaedic Surgeons (AAOS) 2026 annual meeting [1]. It drew from a large national insurance database covering patients who underwent primary total knee arthroplasty (TKA) before February 2020, with complete 5-year follow-up data required for inclusion.

Patients were stratified by whether they had used TRT within one year before surgery. Those with a history of septic arthritis, osteonecrosis, or pathologic fractures were excluded — the intent was to isolate elective, primary joint replacement. Propensity score matching (1:1) balanced the two cohorts on confounders including age, comorbidity burden, and other clinical variables. After matching, 6,625 patients were in each arm, for a total matched sample of 13,250 patients [1].

That sample size matters. Much of the existing literature on TRT and arthroplasty involves single-center cohorts of 150 to 500 patients — useful for generating hypotheses, not for detecting risk at odds ratios of 1.3 to 1.8. The Omurzakov study had the statistical power to find effects that smaller studies miss.

The three risks flagged at 90 days weren't marginal signals. Pneumonia and sepsis both came in at OR 1.8. AKI at OR 1.5. Pulmonary embolism at OR 1.4. In absolute surgical risk terms, those are not numbers you wave away.

The three 90-day risks

The study measured medical complications at 90 days and one year, and periprosthetic complications through five years. The 90-day window is where the headline signals live.

The pulmonary embolism finding is the one that reads as the AAOS headline because PE is the feared early complication of any lower-extremity orthopedic surgery. TKA is already a pro-thrombotic procedure — the operative limb tourniquet, surgical trauma to venous structures, and immobilization in recovery all create clot risk. Standard of care includes anticoagulation prophylaxis for weeks post-op. The Omurzakov data suggests that TRT users carry an additional PE risk on top of that already-elevated baseline [1].

Pneumonia at OR 1.8 is the less intuitive finding and arguably the more informative one. Pneumonia is not a direct mechanical consequence of knee surgery. Its elevation in the TRT group points toward a systemic physiological state — specifically, changes in immune regulation — rather than a surgical site phenomenon. The same logic applies to sepsis at OR 1.8 [1].

Acute kidney injury (AKI) is the third primary flag. TKA involves anesthesia, blood pressure fluctuations, NSAIDs and nephrotoxic analgesics in the post-op protocol, and often meaningful blood loss. Kidneys are already stressed perioperatively. An OR of 1.5 in TRT users suggests the baseline renal reserve is differently compromised entering surgery — a mechanism discussed in the next section [1].

Why TRT changes surgical physiology

Three physiological mechanisms are the most plausible contributors to what the Omurzakov data found. None of these is speculative — they're established effects of testosterone therapy with independent literature behind them.

Erythrocytosis and blood viscosity. Testosterone drives erythropoietin (EPO) production in the kidneys, which stimulates red blood cell synthesis. Elevated hematocrit is one of the most consistent dose-related adverse effects of TRT — it's why monitoring is built into prescribing guidelines. When hematocrit rises above the physiologically normal range, blood viscosity increases and the balance of coagulation tilts toward thrombosis. In the context of surgical venous stasis, tourniquet use, and reduced mobility during recovery, that viscosity shift is directly relevant to pulmonary embolism and deep vein thrombosis (DVT) risk [2, 3]. A 2024 review from the American Society of Hematology noted that polycythemia is a common adverse effect of testosterone therapy requiring ongoing monitoring, and that it can drive VTE risk even at therapeutic doses [2].

Immune modulation. Testosterone is immunosuppressive at the cellular level — it shifts the immune axis toward reduced T-helper and natural killer cell activity and downregulates pro-inflammatory cytokine production. This is not uniformly bad (it's part of why men with autoimmune conditions sometimes see symptomatic relief), but in the perioperative window it means the immune system's early-response capacity to bacterial invasion and pulmonary insults is altered. The pneumonia and sepsis signals in the Omurzakov data are consistent with this immunosuppressive effect. Higher TRT doses compound this [1].

Renal hemodynamics. Testosterone influences renal tubular sodium reabsorption and can promote mild fluid retention and changes in intrarenal blood flow. In the surgical setting — where blood pressure is managed actively, NSAID analgesics are commonly used, and intravascular volume fluctuates — kidneys operating on an altered hemodynamic baseline are more vulnerable to acute injury. The AKI signal at OR 1.5 aligns with this mechanism [1].

The signal that doesn't stop at 90 days

The 90-day findings are the surgical-safety story. The 5-year findings are the implant-survival story. They're different, and the 5-year numbers may ultimately matter more for the TRT community, because the population using TRT is disproportionately middle-aged men — the same demographic now getting TKAs in their 50s and 60s for lifestyle-related osteoarthritis.

The periprosthetic joint infection (PJI) finding at OR 2.3 is the number that will drive orthopedic surgeon behavior. PJI is one of the most catastrophic complications in joint replacement — it typically requires removal of the implant, prolonged antibiotic therapy (often via intravenous lines for 6 weeks), repeat surgery, and in some cases permanent implant-free spacer placement if the infection cannot be cleared. The 2.3-fold elevation in TRT users compared to matched controls is a large signal. It aligns directionally with a separate Duke cohort study of TRT users undergoing total hip arthroplasty (THA), which found 90-day PJI rates of 3.4% in TRT users versus 0.3% in controls — an 11-fold raw difference, though the sample was far smaller [4].

The periprosthetic fracture finding at OR 2.7 is somewhat harder to explain through a single mechanism. Testosterone generally supports bone mineral density; hypogonadal men have lower bone density and are typically the ones prescribed TRT in the first place. One hypothesis is that the fracture elevation reflects the underlying hypogonadal baseline in TRT users — i.e., their bone architecture entering surgery is more compromised than the matched controls even after TRT normalizes serum testosterone levels, and the propensity matching could not fully account for this pre-existing deficit. Another possibility is a pharmacological effect on bone remodeling during high-dose or long-duration TRT. The data doesn't resolve this [1].

Revision surgery at OR 1.5 is the downstream consequence of elevated PJI and fracture — if you have more infections and more fractures, you have more revisions. This is the practical number that orthopedic practices will use when counseling TRT patients on lifetime implant survival probability.

A 2.3× elevation in periprosthetic joint infection at 5 years is not a marginal footnote. PJI is the complication that ends implant careers. For a 55-year-old on TRT getting his first TKA, that number changes the conversation.

The perioperative question: pause or continue?

The Omurzakov study defined TRT use as exposure "within one year of surgery" — it did not analyze whether patients who paused TRT before surgery had better outcomes than those who continued. This is the critical clinical gap the data creates without filling.

Current orthopedic and endocrine society guidelines do not offer a unified, evidence-graded recommendation on perioperative TRT management. Some anesthesia and surgical societies advise stopping medications that increase thrombotic risk (including estrogen-containing contraceptives) 4–6 weeks prior to major elective surgery; whether this framework should be applied to TRT is an active question that the Omurzakov findings will accelerate.

The practical clinical framework that's emerging from commentary on this data:

• Hematocrit check before surgery. A hematocrit above 52–54% substantially amplifies thrombosis risk. If TRT has driven erythrocytosis, therapeutic phlebotomy or a TRT dose reduction to normalize hematocrit before elective surgery is a reasonable precaution that most surgeons would support.
• Discussion of TRT pause. A 4–6 week wash-out period for injectable testosterone (which has the longest half-life of common formulations) would not fully normalize hematocrit in that window, but it would reduce the hormonal signal driving erythrocytosis going forward. Gels and creams clear faster. The trade-off is that hypogonadal men stopping TRT abruptly may experience symptoms (fatigue, mood changes, reduced strength) that complicate surgical recovery. This is an individualized risk-benefit discussion, not a blanket protocol.
• Enhanced anticoagulation surveillance. For TRT users who undergo TKA without pausing, the 1.4× PE signal argues for aggressive adherence to prophylactic anticoagulation protocols and extended-duration prophylaxis conversations with the surgical team.
• Longer-term infection vigilance. The 5-year PJI signal means that any joint pain, warmth, swelling, or fever in a TRT user with a knee implant should trigger a lower clinical threshold for PJI workup — joint aspiration, synovial fluid culture, serum CRP and ESR. Early detection dramatically improves PJI outcomes.

The counterargument worth reading

The full picture of TRT and orthopedic surgery is not one-directional. A 2024 systematic review in Arthroscopy by Flynn and colleagues evaluated five randomized controlled trials of perioperative testosterone supplementation in orthopedic patients — including hip fracture fixation and elective joint reconstruction — and found that four of the five showed functional improvements: better lean mass preservation, higher Harris Hip Scores, faster gait recovery, improved quality of life [5].

The apparent contradiction is resolvable. The Flynn analysis focused largely on patients with hip fracture — an older, frailer population where sarcopenia and hypogonadism compound each other, and where testosterone's muscle-preserving and bone-density effects plausibly improve outcomes. The Omurzakov analysis focused on elective primary TKA in patients who were already on TRT — a different population with a different hormonal baseline.

The distinction the data supports: testosterone supplementation initiated perioperatively to address sarcopenia and hypogonadism in fragile orthopedic patients may be beneficial. Continuation of pre-existing TRT at full therapeutic doses through elective knee replacement in a middle-aged man with already-normalized (or supraphysiologic) testosterone levels carries a measurable complication signal. These are not contradictory findings — they describe different interventions in different populations.

What this means if surgery is on your horizon

If you're on TRT and knee replacement is a near-term possibility, the Omurzakov data gives you three things to bring to the conversation with your surgical team and prescribing physician.

First: get your hematocrit tested before anything else. Elevated hematocrit is the most addressable risk factor in this data set. If your hematocrit is under 50%, the thrombotic risk from TRT-driven erythrocytosis is substantially lower than if it's sitting at 54–56%. This is a simple blood draw, and the result gives you and your surgeon real information rather than a theoretical conversation.

Second: the perioperative pause question needs an answer, not a shrug. Surgeons have historically been inconsistent on TRT management before elective ortho — some ask patients to stop, many don't address it at all. The Omurzakov data is going to change that. Go into your pre-surgical consult expecting the question. If your surgeon doesn't raise it, you should. The relevant facts are your current TRT formulation, dose, duration of use, and your most recent hematocrit and testosterone level.

Third: the 5-year data changes your monitoring posture after surgery. If you proceed with TKA while on TRT — whether you paused it or not — the 2.3× PJI elevation means your lifetime vigilance for joint infection should be higher than the average knee replacement patient. Any unexplained pain, swelling, heat, or drainage around the knee after the first 12 months warrants a call to your surgeon, not a wait-and-see approach. PJI caught at 6 months is treatable. PJI caught at 4 years after a failed course of oral antibiotics is frequently implant-ending.

Bottom line

A 13,250-patient propensity-matched analysis presented at AAOS 2026 found that men on TRT face meaningfully higher rates of pulmonary embolism (OR 1.4), pneumonia (OR 1.8), and acute kidney injury (OR 1.5) in the 90 days after total knee arthroplasty. Those risks persist: DVT and cardiac events remain elevated at 1 year, and periprosthetic joint infection (OR 2.3), periprosthetic fracture (OR 2.7), and revision surgery (OR 1.5) are all significantly higher at 5 years [1].

The mechanisms — erythrocytosis-driven viscosity, immunosuppression, altered renal hemodynamics — are physiologically coherent and independently supported in the TRT literature [2, 3]. The data is observational, not randomized, and the study did not isolate the effect of pausing TRT before surgery. Those are real limitations.

The practical floor: if you're on TRT and considering knee replacement, this study should be on the table in your pre-surgical consultation. Get your hematocrit checked. Have the pause conversation with your prescribing physician. And if you proceed, carry the 5-year infection signal as a reason to have a lower threshold for getting your knee checked if something feels wrong.

The data isn't a reason to avoid TRT or to avoid necessary surgery. It's a reason to go into elective orthopedics with your hormonal baseline actually factored in — which, until this study, most pre-surgical workups didn't do.

Testosterone, hCG, enclomiphene, and the GH-axis recovery peptides relevant to perioperative planning are profiled in The Peptide Manual.