Chronic periprosthetic infection is one of the most feared and difficult to treat complications of total joint arthroplasty1-3. Although treatment is controversial, a two-stage revision with interim placement of an antibiotic-impregnated polymethylmethacrylate (PMMA) spacer is considered the so-called “gold standard4-6.” Vancomycin and tobramycin are two of the most common antibiotics admixed with PMMA. Spacers with this combination of antibiotics have been demonstrated to elute bactericidal concentrations of antibiotics locally7-10.
While systemic concentrations of antibiotic used in this technique are generally low8,9,11-13, several factors can lead to high serum concentrations, including the amount of antibiotic implanted14, combining vancomycin and an aminoglycoside10,15, and manual mixing, which causes variation in elution pharmacodynamics15. Both vancomycin and tobramycin have been associated with renal failure when used systemically in 13.5% (thirty-nine of 289) and 12.9% (275 of 2130) of cases, respectively16,17; there is a synergistic effect when both are used concurrently18,19. To the best of our knowledge, although previous series have not demonstrated nephrotoxicity when these antibiotics were used in PMMA spacers8,13-15,20, previous case reports have21-24.
In this report, we present two cases of nephrotoxicity and elevated serum tobramycin levels that developed after the implantation of a PMMA spacer with high doses of vancomycin and tobramycin. Improvement in renal function occurred following removal of the spacers. Both patients were informed that data concerning their cases would be submitted for publication, and they provided consent.
Case 1. A seventy-seven-year-old woman with a history of hypertension, gout, a seizure disorder, and a right total knee arthroplasty presented to our clinic with two months of knee pain. Initial evaluation revealed a serum creatinine level of 1.5 mg/dL (normal range, 0.75 to 1.2 mg/dL), and knee aspiration cultures grew Staphylococcus lugdunensis and gamma-hemolytic streptococcus. She underwent removal of the implants, and a PMMA spacer constructed of 160 g of Palacos R+G with gentamicin (Zimmer, Warsaw, Indiana), 12 g of vancomycin, and 8 g of tobramycin was placed. Postoperatively, she was treated with intravenous vancomycin (Fig. 1.)
On postoperative day eighteen, a routine serum test revealed a creatinine level of 2.4 mg/dL. Several vancomycin trough levels were drawn between postoperative day twenty-five and thirty-five, the highest of which was 21.6 μg/mL, with all others less than 20 μg/mL (which is considered the high end of acceptable trough serum levels by our infectious disease specialists and laboratory). The patient's serum creatinine level continued to increase, peaking at 5.5 mg/dL on postoperative day twenty-seven (Fig. 2). On postoperative day twenty-eight, the serum tobramycin level was 4.8 μg/mL, even though she had received no intravenous tobramycin at any point (our infectious disease specialists and laboratory generally consider tobramycin trough levels of <2 μg/mL to be necessary and <1 μg/mL ideal to avoid nephrotoxicity). The patient underwent hemodialysis on the same day. The next day, the patient's serum tobramycin level was 3.5 μg/mL. On postoperative day thirty-one, the serum tobramycin level continued to rise to 3.9 μg/mL, and the patient underwent spacer removal; a replacement spacer constructed of PMMA and 10 g of vancomycin was implanted (Fig. 3). Serum creatinine and tobramycin levels (2.7 μg/mL on postoperative day thirty-two and 1.6 μg/mL on postoperative day thirty-three) steadily declined after removal of the initial spacer. At the last laboratory follow-up on postoperative day eighty-six, the serum creatinine level was 1.0 mg/dL.
Case 2. A seventy-three-year-old man with a history of hypothyroidism, hyperlipidemia, hypertension, cataracts, obesity, non-insulin-dependent diabetes mellitus type II, and a left revision total knee arthroplasty with an extensor mechanism allograft presented to us with left knee pain and purulent drainage. Aspiration of synovial fluid revealed infection with methicillin-resistant Staphylococcus aureus. After a year of management with chronic suppressive oral trimethoprim-sulfamethoxazole therapy, as requested by the patient, there was continued intermittent drainage and knee pain. The patient agreed to surgery for removal of the total knee implants as well as placement of an antibiotic-impregnated spacer. At the time of surgery, the patient's serum creatinine level was 1.3 mg/dL (Fig. 4).
A PMMA spacer constructed of 120 g of Palacos R+G with gentamicin, 9 g of vancomycin, and 6 g of tobramycin was placed. The patient was also treated with intravenous daptomycin. On postoperative day seventeen, the serum creatinine level was 2.5 mg/dL. Despite hydration and discontinuation of lisinopril, hydrochlorothiazide, allopurinol, and metformin therapy, the patient's renal function continued to worsen. Elevated tobramycin levels were demonstrated on multiple occasions (1.4 μg/mL on postoperative day seventeen, 1.3 μg/mL on postoperative day twenty, 1.5 μg/mL on postoperative day twenty-two, 2.2 μg/mL on postoperative day twenty-three, and 2 μg/mL on postoperative day twenty-four), even though he had received no intravenous tobramycin. Because of the elevated tobramycin levels, removal of the spacer was considered on postoperative day twenty-three. However, given the tenuous status of the periarticular soft-tissue envelope, spacer removal was delayed in light of the uncertainty about the contribution of tobramycin to the patient's renal failure.
The patient's renal function continued to worsen, and the serum creatinine levels peaked at 4.9 mg/dL on postoperative day forty-six. Three days later, the patient underwent spacer removal, and a replacement spacer constructed of PMMA and 10 g of vancomycin was implanted. Following this surgery, serum tobramycin levels declined to <0.5 μg/mL on postoperative day fifty-two. The patient's serum creatinine level steadily declined after spacer replacement. On postoperative day ninety-one, final laboratory follow-up revealed a serum creatinine level of 1.7 mg/dL.
Neither of the patients had any documented hypotension or signs of systemic infection. In both cases, a renal ultrasound was normal. Urine electrolyte levels revealed a fractional excretion of sodium of >5%, inconsistent with prerenal azotemia in both patients. Additionally, urine and serum eosinophil counts were negative on several occasions, demonstrating no signs of allergic interstitial nephritis in either patient. Even with the consultation of the nephrology team, no other causes of fulminant renal failure could be identified in either case.