Numerous protocols are used to treat an infected total joint arthroplasty. We evaluated survival outcomes of patients with infected total joint arthroplasties over 35 years to determine if one treatment is more effective than others.
Of 17,561 total joint arthroplasties performed at our facility, 116 (0.66%) were diagnosed with deep infection, 22 of which (18.97%) were hematogenous. Success rates among resection, 1-stage, and 2-stage reimplantation arthroplasty were compared in each joint, as were organism types. After excluding 8 patients (6.9%) who died within 1 year of final treatment and 8 lost to follow-up, 30 hips and 70 knees were studied. Seventeen hips were treated with 2-stage arthroplasty; 70.6% succeeded after the first attempt, and 82.4% succeeded after the second. Fifty-one 2-stage knees had 54.9% success after the first attempt and 66.7% success after the second. Five of 5 hips treated with 1-stage arthroplasty succeeded after the first attempt, while 1 of 3 one-stage knees succeeded after the first attempt; none of these 8 joints had a hematogenous infection. No joints requiring a third treatment were successful after the third attempt. There were no differences in survival between different bacteria.
Treatment of deep prosthetic joint infection is difficult once the first attempt has failed and becomes more so after failure of the second attempt. Two-stage arthroplasty provides good results in most cases (82% in hips, 67% in knees), but a third attempt does not increase these percentages. The organism type does not significantly affect success in infected joint treatment.
Deep prosthetic joint infection is a devastating complication of total joint replacement. Documented infection rates in the literature range from 0.5% for the normal-risk patient to upwards of 1.6% for those patients considered to be at high risk.1-8 The infection rate reflects the presence of risk factors in patients,9,10 as well as various environmental controls used pre-, peri-, and postoperatively, such as systemic antibiotics,11,12 laminar air flow,11,12 antibiotic-loaded bone cement,13 and ultraviolet light.14,15
Various treatment methods have been used for cases of deep joint infection. Choices of treatment include irrigation and debridement,16 intravenous antibiotics,17 revision of 1 or several components,18 single-stage arthroplasty,16,19 2-stage excision and reimplantation arthroplasty,16,20,21 arthrodesis,22 and above-knee amputation.22 With each treatment, the goal is resolution, or at least mitigation, of infection and its associated pain and wound drainage. While most treatments for deep infection prove successful after the primary attempt,16-22 in many cases treatment must be continued if a new organism infects the joint or if the primary treatment does not eradicate the infection.
The purpose of this study was to evaluate the treatment of infected prosthetic joints over a 35-year period to determine the success of various treatment protocols. Our question was whether one method of treatment performed significantly better than others in treating an infected total joint arthroplasty. Focus is given to success after 2-stage revision arthroplasty, which has shown the highest success rates in recent years,20 as well as to successful treatment of Gram-positive and Gram-negative infections and in-center improvement in treatment over the 35-year period.
Materials and Methods
From November 1969 to June 2004, 17,561 primary total joint arthroplasties were performed at our institution by 6 orthopedic surgeons (R.A.M.). They included 6556 total hip arthroplasties (THAs; 37.3%) and 11,005 total knee arthroplasties (TKAs; 62.7%). Average patient age was 68.1±10.6 years (range, 17-95 years), average body mass index (BMI) was 29.7±5.7 (range, 10.9-64.3), and average follow-up was 8.1±4.7 years (range, 2.0-33.4 years). Of these 17,561 procedures, 116 joints (0.66%) in 111 patients developed an infection postoperatively, 22 of which (0.13%) had a hematogenous etiology. Infection was diagnosed following an observation of either active drainage or a positive wound culture. The cultures of 2 hips and 3 knees showed no organisms, but because they showed drainage, they were diagnosed as infections.
Thirty-three patients had an infected THA (35 hips; 0.53% of 6556). Seventeen men and 16 women had an average age of 66.2±15.1 years (range, 35-89 years), and average follow-up was 6.9±5.8 years (range, 0.1-23.3 years). Five of these hips (14.3%) had a hematogenous infection.
Seventy-eight patients had an infected TKA (81 knees; 0.74% of 11,005). Thirty men and 48 women had an average age of 65.0±10.9 years (range, 29-84 years), and average follow-up was 5.3±4.0 years (range, 0.2-18.0 years). Seventeen of these knees (21.0%) had a hematogenous infection (Table 1). Only deep prosthetic joint infections, be they primary or hematogenous, were included in this study.
Of the 116 infections, 8 joints (3 hips and 5 knees; 6.90%) in 8 patients were lost within 1 year of final treatment due to patient death resulting directly or indirectly from treatment of the infected joint arthroplasty. The causes of death included congestive heart failure, cardiogenic shock, right heart catheterization, aspiration, stroke, pneumonia, and cardiopulmonary failure. Eight other joints (2 hips and 6 knees; 6.90%) were lost to follow-up within 6 months of final treatment and thus no information regarding the end result of treatment was available in the hospital database. This final treatment occurred an average 127.5±151.0 days (range, 0-412 days) after first diagnosis of infection. Exclusion of these 16 cases left 100 joints (30 hips and 70 knees), implanted in 94 patients, available for further study.
Once diagnosed with deep prosthetic joint infection, patients were treated with irrigation and debridement; intravenous antibiotics specific to the infecting organism; and either permanent resection arthroplasty, 1-stage revision arthroplasty, or delayed 2-stage reimplantation arthroplasty. In cases that called for excision of the prosthesis and reimplantation, the choice between 1-stage and 2-stage arthroplasty was made by the operating surgeon. In earlier years of the study, we based most of these decisions on the findings of Buchholz et al,23 who recommended 1-stage arthroplasty. For 2-stage arthroplasties, the patient was treated with intravenous antibiotics for at least 6 weeks before reimplantation. Multiple cultures were obtained at debridement, excision, and reimplantation to document and identify infecting organisms. Five joints grew no organisms. These 2 hips and 3 knees were included in the analysis of success but not in any analysis of the influence of the organism. If revision arthroplasty failed to eradicate the infection, as evidenced by continued pain and drainage, treatment of the joint continued and additional procedures such as further antibiotic treatment, an additional debridement, another arthroplasty, arthrodesis, or above-knee amputation were performed as necessary.
Successful treatment of infection was defined as no clinical or cultural evidence of infection after treatment, including no pain, fever, or drainage. Arthrodesis (fusion), above-knee amputation, and suppression of infection without eradication constituted failure of treatment. Statistical tests were performed using SAS 9.1.3 software (SAS Institute Inc, Cary, North Carolina). These tests determined significant differences between the success rates of Gram-positive and Gram-negative infections, as well as differences between the various treatment protocols. We used the chi-square test for comparisons between data points >5 and Fisher’s exact test for comparisons between data points <5, and set significance at P<.05.
Of 30 infected THAs (including 4 hips with a hematogenous etiology), 5 (16.7%) were treated primarily with 1-stage exchange arthroplasty. All 5 were treated successfully after a single 1-stage procedure.
Seventeen of 30 hips (56.7%) were treated with 2-stage delayed reimplantation arthroplasty. Success after 1 procedure was 70.6% (12 of 17); the success rate increased to 82.4% (14 of 17) after 2 treatments, but did not increase after a third treatment.
Eight of 30 hips (26.7%) were treated primarily with resection arthroplasty. One of these (12.5%) was treated successfully after 1 treatment. After 2 treatments, the success rate was 37.5% (3 of 8).
Eighteen hips (60%) were successfully treated after 1 treatment. After a second treatment, the success rate increased to 73.3% (22 of 30). No hips succeeded after a third treatment. Seven hips failed after 1 treatment (suppressed infections), and 1 failed with a suppressed infection following a secondary resection.
Both 1-stage (P=.0047) and 2-stage (P=.0122) arthroplasty were more successful than resection arthroplasty in first-time treatment. First-time success was 100% for 1-stage arthroplasty (5 of 5), 70.6% for 2-stage arthroplasty (12 of 17), and 12.5% for resection arthroplasty (1 of 8). By the final treatment, 2-stage arthroplasty continued to fare significantly better than resection (P=.0249). One-stage arthroplasties also were more successful at final treatment, but this difference was not statistically significant (P=.0754). There were no other statistically significant relationships, including between 1-stage and 2-stage arthroplasty (Tables 2, 3).
Organisms did not significantly affect the chances of success. There was a slight difference in the results of final treatment for infected hips; hips infected with >1 Gram-positive organisms fared better than Gram-negative hips (70.8% [17 of 24] vs 50% [3 of 6]), but the difference was not significant (P=.3717). No other P value of any comparison was <.6244 (Tables 4, 5). No detected organisms were resistant to methicillin.
Of 70 infected TKAs (including 15 knees with a hematogenous etiology), 5 knees (7.1%) were cleared of infection after primary deep irrigation and debridement, which was performed an average of 89 days after the index surgery. All debridements were performed immediately after the diagnosis of infection. All knees were cleared of infection post-debridement.
Three of 70 knees (4.3%) were treated after the infection with 1-stage exchange arthroplasty. One of 3 knees was treated successfully with the 1-stage procedure, although resulting in a poor clinical score. The remaining 2 knees ended in failure, with 1 patient dying of septicemia and the other reinfected.
Fifty-one of 70 knees (72.9%) were treated with 2-stage delayed reimplantation arthroplasty after the diagnosis of infection. Twenty-eight of these knees (54.9%) were cleared of infection after 1 two-stage procedure, 8 of the 51 (15.7%) failed after the primary 2-stage arthroplasty with no further treatment, and the remaining 15 (29.4%) underwent continued treatment. Of these 15, six were treated successfully using various treatments (2 debridements, 3 two-stage arthroplasties, 1 resection), which increased the success rate to 66.7% (34 of 51). Two others failed without continued treatment, and 7 required a third treatment that was also unsuccessful.
Eight of 70 knees (11.4%) were treated for their infection with resection arthroplasty. Successful eradication of infection was achieved in 2 knees. Of the 6 failures, 5 were subsequently revised to 4 arthrodeses and 1 above-knee amputation; the sixth was revised to an above-knee amputation after infection of the secondary arthrodesis.
Three of 70 knees (4.3%) were treated primarily with either arthrodesis or amputation. Two were revised to an arthrodesis 7 and 32 months, respectively, after primary arthroplasty; irrigation and debridement were performed but failed to eradicate the infection. One knee underwent an above-knee amputation 5.5 years after primary arthroplasty; the other knee had been amputated after 2 failed 2-stage arthroplasties, and the first suffered from persistent draining.
Eight knees (11.4%) required 3 successive treatment programs postinfection, all of which resulted in failure of the tertiary treatment. Seven of these 8 knees underwent a 2-stage arthroplasty as the primary treatment for infection. Of these 7, two were subsequently revised with a second 2-stage procedure, followed by above-knee amputation after failure of secondary delayed reimplantation. The remaining 5 two-stage knees were treated with secondary resection, followed by 3 arthrodeses and 2 above-knee amputations. The final knee of this set underwent primary resection arthroplasty, which was subsequently revised to an arthrodesis; infection was found in the arthrodesis, and above-knee amputation was performed.
Two-stage arthroplasty was more successful (P=.0465) than resection in treating infected knees by the final treatment (66.7% [34 of 51] vs 25% [2 of 8]). Two-stage arthroplasty was more successful than resection after the first treatment also (54.9% [28 of 51] vs 25% [2 of 8]), but the difference was not significant (P=.1455). No other difference between treatment types was statistically significant (Tables 2, 3).
Organism type did not significantly affect the chances of success. Success of both the first and the final treatment for knee infections correlated with organism type even less than did hip infections, as the P value of each comparison did not fall below .8476 (Tables 4, 5). Four knees were infected with methicillin-resistant Staphylococcus aureus. Two were detected in the 1990s (1994 and 1997); these knees were found with an infection 3.9 and 5.3 years, respectively, after the primary arthroplasty. The other 2 cases were detected in 2003 and were cultured 19 and 20 days, respectively, after the primary arthroplasty. No cases of methicillin resistance were found before 1994.
After exclusion of 1 knee for death within 1 year of final treatment and 2 joints (1 hip and 1 knee) lost to follow-up within 6 months, 19 hematogenous infections were available for study. This included 4 hips and 15 knees. No joint with a hematogenous etiology received a 1-stage arthroplasty at the primary infection treatment.
Thirteen 2-stage arthroplasties (2 hips and 11 knees) were performed at primary treatment. Seven (2 hips and 5 knees) succeeded after 1 treatment (53.8%), and 1 knee failed after 1 treatment and did not undergo further treatment. Nine (2 hips and 7 knees) were successful by the second treatment (69.2%). None of the 3 knees treated 3 times ended successfully.
Three joints (1 knee and 2 hips) underwent primary resection arthroplasty. All 3 were treated 1 time; the knee failed due to continued infection, while 1 hip failed for continued infection and 1 hip had its infection eradicated after primary resection.
One hematogenous knee required only primary irrigation and debridement before its infection was treated. Two knees failed immediately postinfection and did not undergo treatment for eradication; 1 knee required primary fusion, and 1 knee required above-knee amputation.
Success Over the Study Period
We examined the success of all 100 joints and of the largest data set (knees treated with 2-stage arthroplasty) to determine changes in success over the 35-year period of the study. The overall success of treatment rate increased, but not significantly (P=.1205), over the course of the study. One of 4 joints from the 1970s were successful by the final treatment, followed by 11 of 32 joints (34%) in the 1980s, 25 of 43 joints (58%) in the 1990s, and 9 of 15 joints (60%) in the 2000s. For the 51 infected knees treated primarily with 2-stage arthroplasty, 1 of 1 was successful in the 1980s, 20 of 30 (66.7%) were successful in the 1990s, and 13 of 20 (65%) were successful in the 2000s. There was no significant trend in this set (P=.7691).
In our 35-year experience with deep prosthetic joint infection, 2-stage delayed reimplantation arthroplasty proved significantly better at treating infection the first time in both hips and knees than did resection arthroplasty. This advantage continued into the final treatment in infected hips. One-stage arthroplasty fared significantly better than resection in hips only, and only after the first treatment. We discovered a 6.9% rate of death from treatment in our group, and we also found that no third treatment resulted in eradication of infection.
Few studies have analyzed the success of treatment for infection in a large group that was not satisfactorily treated after first revision. One of these studies examined 12 patients with infected knees treated with 2-stage arthroplasty that were subsequently reinfected and treated with 9 secondary 2-stage arthroplasties and 3 arthrodeses.24 In our study, 18 of 68 joints treated primarily with 2-stage arthroplasties required further treatment. Eight of these 18 continuing cases (44.4%) were successfully treated after a second treatment protocol, with 2 end failures. The remaining 8 joints, all knees, failed after the third treatment; none went on to a successful outcome after the second treatment failed. The final endpoint of arthrodesis, above-knee amputation, and persistent drainage is far from an acceptable outcome, and successful treatment is more difficult to achieve after failure of the primary treatment and nearly impossible after failure of the secondary treatment.
Two-stage revision arthroplasty was expected to prove more successful than a 1-stage protocol. This was apparent in knees included in this study, although few 1-stage arthroplasties were performed. The superiority of the 2-stage revision TKA in our study verifies results found in other studies.1,16,17,21 Surprisingly, in hips, 1-stage arthroplasty was more successful than a 1-stage procedure, although there was no possibility of determining statistical significance because of the small sample sizes. Larger studies involving multicenter databases are required to more accurately determine a possible significant difference between the success of 1-stage and 2-stage revision THA.
A review of 43 infected hips determined that the type of infecting organism involved in prosthetic joint infection did not affect the end result.18 This has also been found more specifically in 2-stage reimplantation.25 Our finding that the success rates of Gram-positive and Gram-negative organisms did not affect success rates verifies these past data.
Our study has some limitations. Many of the inherent flaws applicable to retrospective analysis, including nonstandard technique and philosophical changes in protocol, apply to the findings presented here. Success of infection treatment improved (although not significantly) over the study period, largely because of our changes in treatment philosophy over time.23 The greatest strength found in our study is a large cohort size, significantly larger than most studies evaluating treatment for prosthetic joint infection.16-22 However, some subgroups in the study are too small to provide powerful predictions for future treatment, such as the group of 3 knees treated with 1-stage arthroplasty. The size of these groups prevents us from making a definitive recommendation of which primary treatment is most effective, but in our limited experience with deep prosthetic joint infection, we have found that 2-stage arthroplasty is the strongest treatment. Our strongest conclusion regarding the treatment path is our finding that no joints requiring 3 treatments were deemed successful.
- Gill GS, Mills DM. Long-term follow-up evaluation of 1000 consecutive cemented total knee arthroplasties. Clin Orthop Relat Res. 1991; (273):66-76.
- Mauerhan DR, Nelson CL, Smith DL, et al. Prophylaxis against infection in total joint arthroplasty. One day of cefuroxime compared with three days of cefazolin. J Bone Joint Surg Am. 1994; 76(1):39-45.
- Murray WR. Use of antibiotic-containing bone cement. Clin Orthop Relat Res. 1984; (190):89-95.
- Lecuire F, Gontier D, Carrere J, Giordano N, Rubini J, Basso M. Ten-year surveillance of nosocomial surgical site infections in an orthopedic surgery department [in French]. Rev Chir Orthop Reparatrice Appar Mot. 2003; 89(6):479-486.
- Scott WN, Rubinstein M, Scuderi G. Results after knee replacement with a posterior cruciate-substituting prosthesis. J Bone Joint Surg Am. 1988; 70(8):1163-1173.
- Steffen R, von Bremen-Kühne R, Eppe T. Complications after total knee arthroplasty [in German]. Zentralbl Chir. 2003; 128(1):74-77.
- Tang WM, Chiu KY, Ng TP, Yau WP, Ching PT, Seto WH. Efficacy of a single dose of cefazolin as a prophylactic antibiotic in primary arthroplasty. J Arthroplasty. 2003; 18(6):714-718.
- Chiu FY, Chen CM, Lin CF, Lo WH. Cefuroxime-impregnated cement in primary total knee arthroplasty: a prospective, randomized study of three hundred and forty knees. J Bone Joint Surg Am. 2002; 84(5):759-762.
- Peersman G, Laskin R, Davis J, Peterson M. Infection in total knee replacement: a retrospective review of 6489 total knee replacements. Clin Orthop Relat Res. 2001; (392):15-23.
- Ridgeway S, Wilson J, Charlet A, Kafatos G, Pearson A, Coello R. Infection of the surgical site after arthroplasty of the hip. J Bone Joint Surg Br. 2005; 87(6):844-850.
- Nelson JP. The operating room environment and its influence on deep wound infection. In: Murray WR, ed. The Hip: Proceedings of the Fifth Open Scientific Meeting of the Hip Society. St. Louis, MO: CV Mosby; 1977:129-146.
- Lidwell OM, Lowbury EJ, Whyte W, Blowers R, Stanley SJ, Lowe D. Infection and sepsis after operation for total hip or knee-joint replacement: influence of ultraclean air, prophylactic antibiotics and other factors. J Hyg (Lond). 1984; 93(3):505-529.
- Josefsson G, Lindberg L, Wiklander B. Systemic antibiotics and gentamicin-containing bone cement in the prophylaxis of postoperative infections in total hip arthroplasty. Clin Orthop Relat Res. 1981; (159):194-200.
- Moggio M, Goldner JL, McCollum DE, Beissinger SF. Wound infections in patients undergoing total hip arthroplasty. Ultraviolet light for the control of airborne bacteria. Arch Surg. 1979; 114(7):815-823.
- Ritter MA, Olberding EM, Malinzak RA. Ultraviolet lighting during orthopaedic surgery and the rate of infection. J Bone Joint Surg Am. 2007; 89(9):1935-1940.
- Borden LS, Gearen PF. Infected total knee arthroplasty. A protocol for management. J Arthroplasty. 1987; 2(1):27-36.
- Trebse R, Pisot V, Trampuz A. Treatment of infected retained implants. J Bone Joint Surg Br. 2005; 87(2):249-256.
- Balderston RA, Hiller WD, Iannotti JP, et al. Treatment of the septic hip with total hip arthroplasty. Clin Orthop Relat Res. 1987; (221):231-237.
- Izquierdo RJ, Northmore-Ball MD. Long-term results of revision hip arthroplasty. Survival analysis with special reference to the femoral component. J Bone Joint Surg Br. 1994; 76(1):34-39.
- Lentino JR. Prosthetic joint infections: bane of orthopedists, challenge for infectious disease specialists. Clin Infect Dis. 2003; 36(9):1157-1161.
- Wilson MG, Kelley K, Thornhill TS. Infection as a complication of total knee-replacement arthroplasty. Risk factors and treatment in sixty-seven cases. J Bone Joint Surg Am. 1990; 72(6):878-883.
- Johnson DP, Bannister GC. The outcome of infected arthroplasty of the knee. J Bone Joint Surg Br. 1986; 68(2):289-291.
- Buchholz HW, Elson RA, Engelbrecht E, Lodenkämper H, Röttger J, Siegel A. Management of deep infection of total hip replacement. J Bone Joint Surg Br. 1981; 63(3):342-353.
- Backe HA Jr, Wolff DA, Windsor RE. Total knee replacement infection after 2-stage reimplantation: results of subsequent 2-stage reimplantation. Clin Orthop Relat Res. 1996; (331):125-131.
- Goldman RT, Scuderi GR, Insall JN. 2-stage reimplantation for infected total knee replacement. Clin Orthop Relat Res. 1996; (331):118-124.
Dr Ritter and Mr Farris are from the Center for Hip and Knee Surgery, St Francis Hospital, Mooresville, Indiana.
Dr Ritter and Mr Farris have no relevant financial relationships to disclose.
Correspondence should be addressed to: Merrill A. Ritter, MD, 1199 Hadley Rd, Mooresville, IN 46158 (email@example.com).