Antibiotic-loaded cement is a local depot, and no local depot stays purely local. Serum concentrations after implantation are measurable, persist for weeks, and track kidney injury in ways that grams per batch do not predict. Rifampicin, the oral backbone of staphylococcal prosthetic joint infection therapy, is the drug most often paired with that depot and least often measured next to it. The pharmacokinetic record is clear that vancomycin and aminoglycoside exposure continues long after the wound closes, and that baseline kidney function plus total implanted drug mass predict harm better than concentration does.
Why a local depot is never purely local
Two-stage exchange for periprosthetic joint infection pairs implant removal and debridement with systemic antimicrobials and a temporary antibiotic-loaded polymethylmethacrylate spacer. The spacer does two jobs at once: it holds the joint space open and it floods the local tissue with drug meant to suppress residual biofilm organisms.
Local delivery has been read as local containment, and that reading is unsafe. The antibiotic is physically embedded in the polymer rather than chemically bound to it. Release requires fluid to penetrate the matrix, dissolve accessible particles, and diffuse out through pores, channels, and cracks. What follows is an early burst from the surface and interconnected pores, then a lower and steadily declining phase. Local concentrations can stay biologically active for weeks, but only a fraction of the loaded drug ever leaves the cement.
Once drug reaches tissue and joint fluid, part of it reaches the circulation. Human studies have measured vancomycin, tobramycin, and gentamicin in serum after spacer implantation, with consequences running from silent detection to acute kidney injury, electrolyte disturbance, treatment change, ototoxicity, and in severe reports dialysis.
One distinction governs the rest of this article. Low-dose cement used for prophylaxis or fixation and high-dose temporary spacers used to treat established infection differ in load, geometry, patient population, and objective. Prophylaxis data cannot be carried over to the safety of a therapeutic spacer, and a good deal of the apparent disagreement in this literature comes from mixing the two.
How the pharmacokinetic evidence was built
This is a narrative mini review rather than a systematic one, and the search is described explicitly. A targeted bibliographic search ran to 30 September 2026 with backward reference screening, combining spacer terminology with vancomycin, tobramycin, or gentamicin, and with systemic absorption, pharmacokinetics, nephrotoxicity, or acute kidney injury. English-language human studies of temporary therapeutic hip or knee spacers were prioritized. Prophylactic cement work and bench studies supplied mechanism and material properties, not direct safety evidence.
Study design | Population | Spacer content and sampling | Key quantitative result |
Observational | 34 patients, 36 infected knees | Mean 3.4 cement batches, about 10.5 g vancomycin plus 12.5 g gentamicin; six weeks of systemic therapy; no serial levels | One transient creatinine increase, no reported clinical renal failure |
Prospective pharmacokinetic | 21 patients, 10 hips and 11 knees | Vancomycin 3.0 g per 40 g, tobramycin 2.4 g per 40 g, premixed gentamicin 0.5 g; weekly serum for 8 weeks | Gentamicin 0.275 to 0.364 mg/L, tobramycin 0.313 to 0.527 mg/L, vancomycin 5.46 to 15.34 mg/L, all detectable at 8 weeks |
Selected case series | 10 patients with renal dysfunction among 46 revisions | Mean tobramycin 8.2 g; no systemic aminoglycoside; levels drawn after creatinine had already risen | Mean 3.3 micrograms per mL, range 0.1 to 19.8; mean creatinine increase 260 percent |
Prospective, unselected | 20 hip and knee patients | Median total tobramycin 4.8 g; routine sampling on days 1 and 3 | Detectable in 13 of 20 on day 1 and 5 of 20 on day 3, most undetectable by 72 hours, no kidney injury through 14 days |
Retrospective, source-restricted | 62 gram-positive infections, 34 with renal insufficiency | Vancomycin-loaded spacers, systemic vancomycin excluded | Peak concentrations 1.23 to 5.43 mg/L; peak inversely associated with estimated filtration rate |
Rows two and four in that table are the ones to read together. Both are prospective. One sampled weekly for eight weeks and found every drug still detectable at the end. The other sampled on days 1 and 3 and found most concentrations gone by 72 hours. Neither is wrong. They covered different windows, different doses, and different sampling rules, and a reader who compares the raw numbers without noticing that will conclude the field contradicts itself.
Row three carries a bias that has to be named. Levels were drawn because kidney function had already declined, which is selection by outcome. It is strong evidence that the aminoglycoside came from the spacer, since no systemic aminoglycoside was given, and it says nothing about incidence.
What the measured concentrations actually mean
A serum concentration after antibiotic-loaded cement placement reflects two things at once, drug input and drug clearance. A high early value can come from burst release, from intravenous dosing of the same drug, from poor renal clearance, or from all three together. A single low value does not exclude meaningful cumulative exposure.
Vancomycin is where interpretation fails first. In the eight-week prospective study, systemic vancomycin use was independently associated with higher serum vancomycin, coefficient 6.24 with a 95 percent confidence interval of 2.72 to 9.75. When the same drug is running intravenously, no measured concentration can be assigned to the spacer.
Aminoglycosides allow cleaner attribution because they are usually not given systemically beside a spacer. Detect tobramycin in a patient receiving no systemic aminoglycoside and the spacer is the only credible source. That is why the aminoglycoside data from antibiotic-loaded cement carry the causal argument and the vancomycin data carry the confounding argument.
Acute kidney injury: incidence depends on who you count
Kidney injury rates after spacer placement span a range wide enough to be useless without the cohort attached. Patient mix, joint, formulation, monitoring intensity, and above all the definition drive the spread.
Study | Population | Kidney injury rate | Principal finding |
Systematic review of 24 studies | Hip, knee, and mixed cohorts | 4.2 percent in hip cohorts, 14 percent in knee cohorts, 27 percent in mixed cohorts | Range 0 to 35 percent across studies; risk factors included chronic kidney disease, low preoperative hemoglobin, transfusion |
Cohort plus meta-analysis | 340 first-stage exchanges, pooled 2,525 patients | 12.1 percent in cohort, 16.6 percent pooled | Acute dialysis in 1.4 percent; older age odds ratio 1.509 per decade; hypoalbuminemia odds ratio 3.593 |
Retrospective, 84 revision knees | Revision knee arthroplasty | 17 percent, 95 percent confidence interval 10 to 26 | Tobramycin above 4.8 g odds ratio 5.87; each additional gram odds ratio 1.24 |
Prospective, 50 revisions | Consecutive hip and knee revisions | 20 percent | Median onset postoperative day 2; day-1 tobramycin 1.9 versus 0.9 micrograms per mL, p equal to 0.01 |
Retrospective, 424 patients | 455 knee spacers | 14 percent without chronic kidney disease, 45 percent with it | Above 3.6 g per batch odds ratio 1.9 for vancomycin and 1.8 for aminoglycoside |
Randomized trial, 127 patients | High-dose spacer versus one-stage exchange | 22.7 percent versus 6.6 percent | Spacer odds ratio 7.48; chronic kidney disease odds ratio 3.84 |
Retrospective comparative, 277 patients | Stage-one revision versus debridement without local cement | 17.9 percent versus 14.7 percent | Odds ratio 1.43 with confidence interval 0.70 to 2.93, not significant |
The last row is the awkward one for anyone wanting a simple answer. It compares antibiotic cement against a procedure with no local cement and finds no statistically significant excess. The randomized trial in the sixth row finds a large one. Both are sound designs and they disagree, which is exactly why the review keeps evidence of absorption separate from evidence of harm.
Definition drift explains part of the range. The current standard defines injury as a creatinine increase of at least 0.3 mg per dL within 48 hours, an increase to at least 1.5 times baseline within seven days, or urine output below 0.5 mL per kg per hour for six hours. Older definitions used different windows, and one large cohort counted qualifying increases between weekly measurements even beyond the nominal window. An incidence figure without its definition cannot be compared with anything.
Mechanism differs by class, which matters because the two are usually co-loaded. Vancomycin injury runs through oxidative stress, mitochondrial dysfunction, tubular epithelial injury, inflammatory signaling, and intratubular casts. Aminoglycosides are filtered and taken up by proximal tubular cells, where lysosomal and mitochondrial accumulation produces tubular dysfunction and necrosis. Falling filtration then prolongs exposure, and that feedback loop turns one hit into a sustained one.
The counterintuitive finding: grams per batch is the wrong exposure metric
The number everyone quotes is grams of antibiotic per 40 g of cement. That number describes concentration in the mix. It does not describe exposure.
Total implanted mass depends on how many batches went in. One series averaged 3.4 batches per patient, and another averaged 8 g vancomycin plus 9 g aminoglycoside per construct. Two patients at identical concentration per batch can differ several-fold in total drug implanted, and therefore in how much was available to elute. Spacer geometry and surface area change release further, so a large articulating spacer and a small static block are different products at the same nominal concentration.
The review states the implication as a hypothesis rather than a finding: total implanted drug mass may predict systemic risk better than concentration per 40 g, and that proposition has not been tested directly. That honesty helps. The practical takeaway is to record both numbers rather than swap one unvalidated metric for another.
A second surprise concerns the peak. Day-one tobramycin separated patients who developed kidney injury from those who did not in one prospective cohort, at 1.9 versus 0.9 micrograms per mL. The absolute difference is small and the values sit far below therapeutic systemic dosing, yet the signal was statistically significant. Early exposure, not cumulative dose, was the discriminator. The same cohort reported an odds ratio of 8.2 for premixed gentamicin cement, which is a formulation effect rather than a dose effect.
Where rifampicin fits in antibiotic-loaded cement and oral therapy
Rifampicin is a legitimate edge case here, and the review's own framing shows why. It notes that less commonly used agents need evidence of heat stability, polymer compatibility, retained antimicrobial activity after curing, and acceptable mechanical effects. Rifampicin clears a higher bar than vancomycin on the thermal and mechanical side, and that is a materials question before it becomes a pharmacology question.
Its clinical role in prosthetic joint infection is different and better established. It is the oral component that penetrates biofilm, added to a systemic regimen once the implant is out and susceptibility is known, and one prospective cohort in this review recorded postoperative systemic therapy containing cephalosporins, vancomycin, or rifampin alongside a tobramycin and vancomycin spacer. That study is the clearest picture of an interaction the review never analyzes: a patient can be on rifampicin systemically while an aminoglycoside elutes from cement, and both act on the same kidney.
Two consequences follow. Rifampicin is a strong inducer of hepatic and intestinal enzymes and transporters, so any drug whose exposure is being managed around a spacer has to be assessed against the whole regimen rather than against the depot alone. And if a rifampicin-loaded or rifampicin-adjacent spacer is being developed, the elution and mechanical data have to be generated for that formulation, because release from polymethylmethacrylate cannot be inferred from vancomycin.
For an ingredient supplier the point is unglamorous. The specification that matters for a spacer program is not only assay and impurities but thermal stability across the curing exotherm, a particle size distribution consistent with the mixing and vacuum protocol, and a polymorphic form that survives processing. A rifampicin grade qualified for oral solid dosage starts that conversation rather than ending it.
Practical decisions for a risk-management framework
The review proposes a management pathway and labels it proposed rather than prospectively validated. Each element is still actionable:
- Assess renal risk against a stable baseline, and do not treat a creatinine-derived filtration estimate as stable clearance while injury is evolving.
- Document each antibiotic's grams per batch, the number and mass of batches, total grams implanted, and spacer geometry or estimated surface area.
- Coordinate local and systemic prescribing explicitly, because the same drug can arrive from two directions and a measured concentration cannot separate them.
- Monitor high-risk inpatients early rather than universally, using daily creatinine and urine output read over the standard injury windows.
- Measure serum levels selectively when exposure or toxicity is suspected, and treat the value as one input rather than a threshold.
- Escalate to multidisciplinary review, and consider spacer removal for progressive injury or persistent clinically important exposure, weighing operative risk against infection control.
One caveat belongs beside that list. If intravenous vancomycin is indicated for a serious resistant infection, exposure-guided dosing applies to the systemic regimen and must not be read as a validated safety threshold for spacer-derived exposure. The two exposures share a compartment, not a dosing rule.
Limitations
The review is narrative, performs no formal risk-of-bias assessment, and pools nothing. Apart from one randomized trial the underlying studies are observational, mostly small, with heterogeneous injury definitions, inconsistent dose reporting, and incomplete separation of spacer-derived from intravenous exposure. Several cohorts selected patients who already had renal dysfunction, which pushes incidence estimates up. Follow-up often stops at discharge, and ototoxicity is plausibly underreported because it is not systematically screened. The management pathway has not been validated prospectively. Rifampicin appears only as a component of systemic therapy in one cohort and is not analyzed as a spacer antibiotic.
Bottom line
Antibiotic-loaded cement delivers drug into the circulation for weeks, and that part is solid: gentamicin 0.275 to 0.364 mg/L, tobramycin 0.313 to 0.527 mg/L, and vancomycin 5.46 to 15.34 mg/L, all still measurable eight weeks after implantation. The harm signal is messier. Reported kidney injury runs from 4.2 percent in hip cohorts to 27 percent in mixed cohorts, one randomized trial reported an odds ratio of 7.48 for a high-dose spacer, and one comparative cohort found no significant excess. What holds across studies is that baseline kidney function and total implanted drug mass matter more than grams per batch, and that rifampicin's role sits on the systemic side of the regimen, where its induction effects have to be managed alongside whatever the depot releases.
FAQ
What does antibiotic-loaded cement actually release
It releases drug through fluid penetrating the polymer, dissolving accessible particles, and diffusing through pores, channels, and cracks. The profile is an early burst followed by a slower declining phase. Local concentrations can stay active for weeks, but only a fraction of the incorporated drug elutes, and a portion of what elutes reaches the bloodstream and stays measurable for weeks after surgery.
How high are serum levels after spacer implantation
In a prospective study of 21 patients sampled weekly for eight weeks, gentamicin averaged 0.275 to 0.364 mg/L, tobramycin 0.313 to 0.527 mg/L, and vancomycin 5.46 to 15.34 mg/L, with all three still detectable at week eight. A separate series with a mean tobramycin load of 8.2 g reported a mean concentration of 3.3 micrograms per mL, ranging from 0.1 to 19.8.
Why does dose per batch not predict systemic exposure
Grams per 40 g of cement describes formulation concentration, not the mass implanted. Total drug mass depends on how many cement batches were used and their mass, and one cohort averaged 3.4 batches while another averaged 8 g vancomycin plus 9 g aminoglycoside per construct. Spacer geometry and surface area change release further, so two spacers at identical concentration can differ several-fold in total elutable drug.
How common is acute kidney injury after a spacer
Reported rates depend heavily on the cohort and the definition. A systematic review of 24 studies reported means of 4.2 percent in hip cohorts, 14 percent in knee cohorts, and 27 percent in mixed cohorts, with ranges up to 35 percent. A large cohort reported 14 percent in patients without chronic kidney disease and 45 percent in those with it. A randomized comparison of a high-dose spacer against one-stage exchange reported 22.7 percent versus 6.6 percent.
Does the spacer cause kidney injury or just correlate with it
The evidence splits. Detectable aminoglycoside in patients receiving no systemic aminoglycoside establishes that absorption happens, not that the concentration caused injury. Dose and concentration associations are consistent with causation but remain open to confounding, and falling clearance can itself raise concentrations. One randomized trial strengthens causal inference; one comparative cohort found no significant excess risk. Both are legitimate and they disagree.
Which patients carry the highest renal risk
Patients with chronic kidney disease carry the highest risk by a wide margin, with one cohort reporting 45 percent injury against 14 percent without it. Other reported predictors include older age, low preoperative hemoglobin, hypoalbuminemia, perioperative transfusion, diuretic use, and concurrent systemic vancomycin. Premixed gentamicin cement was associated with an odds ratio of 8.2 in one prospective cohort, which suggests a formulation effect as well as a patient effect.
Can vancomycin levels be attributed to the spacer
Only when systemic vancomycin is excluded. When both sources are present, the measured concentration cannot be partitioned, and one prospective study found systemic use independently associated with higher serum vancomycin with a coefficient of 6.24. Aminoglycosides give cleaner attribution because they are usually not given systemically alongside a spacer, which is why the causal argument rests mainly on tobramycin and gentamicin data.
What is the correct way to report spacer dosing
Report grams of each antibiotic per batch, the number and mass of batches, total grams implanted, and spacer geometry or estimated surface area. Concentration per 40 g alone is insufficient for pharmacologic interpretation. The review proposes that total implanted mass may better predict systemic risk than concentration, and explicitly labels that as a hypothesis requiring direct comparative validation rather than an established finding.
Does rifampicin belong in the spacer or in the systemic regimen
Its established role is systemic, as the oral biofilm-active component added once the implant is removed and susceptibility is known, and one cohort in this review recorded rifampin among postoperative systemic agents alongside a vancomycin and tobramycin spacer. Adding it to the cement is a materials question first, because the review notes that less commonly used agents must demonstrate heat stability, polymer compatibility, retained activity after curing, and acceptable mechanical effects.
How should a high-risk patient be monitored
Assess renal risk against a stable baseline, record every local and systemic dose, monitor creatinine and urine output daily in high-risk inpatients over the standard injury windows, measure serum drug levels selectively when exposure or toxicity is suspected, and escalate to multidisciplinary review. Consider spacer removal for progressive injury or persistent clinically important exposure, balancing operative risk, mechanical needs, and infection control. The pathway is proposed, not prospectively validated.
What is still unknown about antibiotic-loaded cement
Whether total implanted mass outperforms concentration as a risk metric, how much of the reported kidney injury is attributable to the spacer rather than to perioperative physiology or concurrent drugs, what a validated exposure threshold would be, how often ototoxicity occurs, and how rifampicin behaves as a cement additive in terms of elution and mechanical strength. Standardized dose, sampling, exposure, and outcome reporting would answer several of these at once.