Periprosthetic joint infection remains one of the most challenging complications following joint replacement and a leading cause of early implant failure.1 Although the overall infection rate is relatively low, being approximately 1% following hip and knee arthroplasties,2 it might dramatically increase in the presence of factors such as patients' immunosuppression, extensive soft tissue dissection, long operating times,3 and neoplastic diseases.4
A deep infection may be a devastating complication following megaprosthetic replacement after tumor resection, leading to challenges in treatment decisions5 and posing a high risk for repeated surgical procedures, prolonged hospital stay and rehabilitation, pain, poor functional outcome, and failed limb salvage.3,6,7 Overall, a recent systematic review8 reported a mean rate of perimegaprosthetic infection (PMI) of 10%; notably, a PMI rate as high as 43% may follow revision surgery.9 It has been suggested that long-term antibiotic prophylaxis decreases the risk of PMIs,8 but controversy remains about the role of other factors in the development of PMIs, thus supporting the need for a comprehensive literature review.
The purpose of this study was to systematically evaluate published clinical studies determining which variables predict the PMI rate after tumor resection.
Materials and Methods
The authors conducted a systematic literature search using the MEDLINE/PubMed and Cochrane Register of Controlled Trials databases. They wanted to identify and include in the search all English-language studies addressing septic complications of the use of megaprostheses in the treatment of bone tumors including an analysis of risk factors for PMIs. Level V studies were excluded from the search.
The MEDLINE/PubMed search was performed as follows: 1: “megaprosth* infect*” AND “English”[Language] NOT “case reports”[Publication Type] NOT “comment”[Publication Type] NOT “congresses”[Publication Type] NOT “editorial”[Publication Type] NOT “technical report”[Publication Type]; 2: “(tumor* OR tumour*) endoprosth* infect*” AND “English”[Language] NOT “case reports”[Publication Type] NOT “comment”[Publication Type] NOT “congresses”[Publication Type] NOT “editorial”[Publication Type] NOT “technical report”[Publication Type]; and 3: 1 OR 2. The search of the Cochrane Register of Controlled Trials database was performed using the terms “megaprosth*” OR “tumor* endoprosth*” AND “infect*”.
This search identified 249 pertinent records. These records were reviewed and cross-referenced to exclude repeated references, leaving 236 articles for review. Overall, 212 articles were excluded on the basis of titles or abstracts. Full-text articles were then retrieved and reviewed, in full, for the remaining 24 studies. Reference sections of all articles were searched for undetected studies, thus adding 9 studies to the review. Of the 33 studies, 25 were further excluded on the basis of the full-text review. The remaining 8 articles10–17 met the inclusion criteria and were analyzed in this systematic review (Figure).
Flowchart for the literature search.
All included studies (Table 1) retrospectively reviewed a series of patients having undergone a megaprosthetic replacement after tumor resection (level IV studies). These studies reported on a total of 2136 patients (range, 62–1161 per study); the mean patient age was 33.3 years (range, 5–90 years). In 2 studies,10,13 megaprostheses were used to treat lesions located in either the femur, tibia, humerus, or pelvis; 2 studies14,17 included only lesions of the proximal tibia; 2 studies11,12 included only lesions of the femur and tibia; 1 study15 included only lesions of the proximal femur; and 1 study16 included lesions located in either the distal femur or the proximal tibia. The mean PMI rate was 14.2% (range, 7.2%–25.8%), with PMIs having occurred from 9 days to 240 months postoperatively.
Details of Included Studies
Statistical analysis of risk factors for PMIs was performed through Kaplan-Meier survival curve comparisons (ie, using the Mantel-Cox log-rank test) in 5 studies,11,12,14–16 logistic regression analysis (ie, using the Hosmer-Lemeshow goodness-of-fit test) in 1 study,13 comparative statistics (ie, using the chi-square test) in 1 study,17 and preliminary comparative statistics (ie, using the chi-square test, Fisher's exact test, the Student's t test, or the Mann-Whitney U test, when appropriate) followed by stepwise logistic regression analysis in 1 study.10
All of the evaluated risk factors are summarized in Table 2. Patients' sex and age at surgery were not found to influence PMI rate.10,13,14,16 Increasing body mass index was associated with a significantly higher risk of PMI.13 Overall, the presence of comorbidities was also associated with PMIs,10 but neither a higher American Society of Anesthesiologists score10 nor the presence of diabetes mellitus13 specifically increased the risk. Perimegaprosthetic infections were also found to be more frequent with a coexisting skin necrosis or a superficial surgical site infection.16 Preoperative white blood cell count,14 febrile neutropenia,13,14 bacteremia,13 and the use of an indwelling port system14 were not associated with the development of a PMI.
Risk Factors for Perimegaprosthetic Infection
Six studies10,11,13–16 explored the role of lesion histology and tumor characteristics in determining PMI rate. Controversy remains regarding whether primary or metastatic lesions have a higher risk of PMI.11,15 Among primary tumors, no significant differences in PMI rates were found according to histological features10,11,14 and metastatic spread at diagnosis.10,13 No significant effects of chemotherapy10,11,13,14,16 and radiotherapy10,11,15 were found for the development of a PMI.
Five studies10–12,15,16 explored the role of localization of lesions in determining PMI rate. Proximal tibial and pelvic locations were identified as strong risk factors for PMIs10; in particular, pelvic extension of proximal femoral disease has led to a higher PMI rate than that following reconstruction of the proximal femur alone.15 Conversely, distal femoral location of the lesion appears to be a protective factor for PMI.10
A preoperative hospitalization of more than 48 hours has been found to predict a higher risk of PMI.10 Overall, features of perioperative antibiotic prophylaxis do not act as risk factors for PMIs; indeed, no significant differences in PMI rates have been found according to the choice of antibiotic used,13 the intraoperative dosing,13 the number of postoperative antibiotics used,16 and the length of postoperative prophylaxis.13,14
Specific operative features have been widely explored. No significant effects of the use of a clean air operating room were found for the development of a PMI.16 Two studies11,15 have shown that the use of a cemented megaprosthesis has led to a higher PMI rate than that following the use of an uncemented megaprosthesis; conversely, there were no statistical differences in PMI rates according to prosthesis type10,16 or hinge movement.11 Overall, width of resection margins,10 bone resection length,10,14,16 and extracapsular resection of knee tumors16 do not affect PMI rate; however, resection of greater than 37% of the proximal tibia was found to be related to a higher PMI rate.14 In distal femoral lesions, resection of 3 or 4 heads of the quadriceps muscle was associated with a significantly higher PMI rate than resection of 2 or fewer heads.16 The need for additional surgical procedures after the megaprosthesis implantation was associated with a higher risk of PMI.10 Two10,13 of 3 studies10,13,14 agreed that increasing surgical times statistically correlated with PMI rate; in particular, an operation lasting longer than 2.5 hours is a strong risk factor for the development of a PMI.10 One17 of 3 studies13,16,17 indicated that the routine use of a gastrocnemius flap for anterior reconstruction and megaprosthesis coverage following a proximal tibial resection led to a significant reduction of PMI rate. On the other hand, the PMI rate following megaprosthetic replacement of the proximal tibia was not found to increase with the use of synthetic material to fix the patellar tendon.14
An increased risk of PMIs was found for patients who needed a postoperative admission to the intensive care unit and for those having experienced a postoperative hematoma.13 Despite the fact that the estimated blood loss was not higher in patients with a PMI,10 two10,13 of 3 studies10,13,14 detected that an increased postoperative blood transfusion requirement posed a higher risk of PMI; in particular, transfusion of 2 or more units of allogeneic packed cells emerged as a predictor of PMI.10
Due to improved imaging modalities, understanding of tumor biology, and advances in adjuvant treatments, limb amputation is now used selectively in the treatment of bone tumors.18 Indeed, limb salvage surgery has become the standard of care for bone tumors, being applicable for up to 85% of patients18 and resulting in limb salvage for up to 90% of them.19 Because the use of megaprostheses is currently gaining momentum as the most common reconstructive technique after tumor resection,12,20 the goal of this systematic review was to identify factors affecting the risk of PMIs. Notably, the presence of several conflicting results together with the lack of comprehensive reviews to date exploring the exact causative factors for PMIs has often led to extending to megaprostheses previous information on risk factors for periprosthetic joint infections resulting from the analysis of primary arthroplasties.5
Both the presence of comorbidities and a higher body mass index at surgery have been reported as significant patient-related risk factors for PMIs. Of interest, the direct association between body mass index and infection rate is well established in conventional arthroplasties, as obesity increases both the morbidity from osteoarthritis and the risk of periprosthetic joint infections21; conversely, caution should be exercised in applying this risk factor to the tumor megaprosthesis population, as up to 18% of previous cohorts13 were underweight.
Localization of lesions was found to be an important risk factor for PMIs, with higher rates expected following megaprosthetic replacement of both pelvic and proximal tibial tumors. As previously argued,10 long operating times because of the complex anatomy, large bone defects after tumor resection, and difficulty in achieving wide surgical margins are all factors accounting for challenges in pelvic surgery. On the other hand, the high PMI rate following proximal tibial tumor resection may correlate with the difficulty in achieving a good soft tissue coverage, thus being confirmed by the marked reduction in PMI rate following the routine use of gastrocnemius flaps as part of the reconstruction.17 Notably, soft tissue condition has been shown to be a strong predictor of PMI and the most important risk factor for failed limb salvage.7
The most important finding of this systematic review was that the majority of factors accounting for a higher PMI rate are related to both surgical procedure and hospitalization. Of interest, prolonged preoperative hospital stay and operative time,22–24 the presence of postoperative hematoma,25 and the increasing number of allogeneic blood transfusions26 are all proven risk factors for periprosthetic joint infections, and their role has been confirmed following megaprosthetic replacement after tumor resection. The results of the current review indicate that the features of perioperative antibiotic prophylaxis do not act as risk factors for PMIs, thus contradicting a previously published systematic review.8 However, evidence and guidelines directing the prescription of prophylactic antibiotic regimens in musculoskeletal tumor surgery are lacking, so considerable variation in strategies exists.27
Of interest, this review found agreement that a higher PMI rate can be expected in cemented than in cementless megaprostheses,11–15 thus contradicting information regarding conventional arthroplasties.28 However, factors such as the antibiotic loading of acrylic cement,29 megaprosthesis alloy system,30 and surface treatments31 are all pertinent to consider.
This systematic review had limitations. First, the lack of randomized controlled trials (level I or II) on the role of specific factors in the development of PMIs impeded the performance of a strict meta-analysis to draw firm and univocal conclusions. In addition, there is wide variability in the reporting of risk factors for PMIs. Notably, none of the studies considered included all of the factors studied. Moreover, there is variability in the statistical measures used. These issues combined make it difficult to compare the results for all of the factors important to the surgeon and the patient.
This systematic review identified the role of several factors in the development of PMIs after tumor resection. Identified factors mostly relate to both surgical procedure and hospitalization. Therefore, there are several modifiable risk factors for PMIs. Physicians should consider these results when discussing the outcomes of limb salvage surgery with patients and trying to reduce the overall burden of PMIs.
- Willis-Owen CA, Konyves A, Martin DK. Factors affecting the incidence of infection in hip and knee replacement: an analysis of 5277 cases. J Bone Joint Surg Br. 2010; 92(8):1128–1133. doi:10.1302/0301-620X.92B8.24333 [CrossRef]
- Lamagni T. Epidemiology and burden of prosthetic joint infections. J Antimicrob Chemother. 2014; 69(suppl 1):i5–i10. doi:10.1093/jac/dku247 [CrossRef]
- Jeys LM, Kulkarni A, Grimer RJ, Carter SR, Tillman RM, Abudu A. Endoprosthetic reconstruction for the treatment of musculoskeletal tumors of the appendicular skeleton and pelvis. J Bone Joint Surg Am. 2008; 90(6):1265–1271. doi:10.2106/JBJS.F.01324 [CrossRef]
- Baek SH. Identification and preoperative optimization of risk factors to prevent periprosthetic joint infection. World J Orthop. 2014; 5(3):362–367. doi:10.5312/wjo.v5.i3.362 [CrossRef]
- Ercolano LB, Christensen T, McGough R, Weiss K. Treatment solutions are unclear for perimegaprosthetic infections. Clin Orthop Relat Res. 2013; 471(10):3204–3213. doi:10.1007/s11999-013-2852-7 [CrossRef]
- Pilge H, Gradl G, von Eisenhart-Rothe R, Gollwitzer H. Incidence and outcome after infection of megaprostheses. Hip Int. 2012; 22(suppl 8):S83–S90. doi:10.5301/HIP.2012.9576 [CrossRef]
- Hardes J, Gebert C, Schwappach A, et al. Characteristics and outcome of infections associated with tumor endoprostheses. Arch Orthop Trauma Surg. 2006; 126(5):289–296. doi:10.1007/s00402-005-0009-1 [CrossRef]
- Racano A, Pazionis T, Farrokhyar F, Deheshi B, Ghert M. High infection rate outcomes in long-bone tumor surgery with endoprosthetic reconstruction in adults: a systematic review. Clin Orthop Relat Res. 2013; 471(6):2017–2027. doi:10.1007/s11999-013-2842-9 [CrossRef]
- Capanna R, Morris HG, Campanacci D, Del Ben M, Campanacci M. Modular uncemented prosthetic reconstruction after resection of tumours of the distal femur. J Bone Joint Surg Br. 1994; 76(2):178–186.
- Dhanoa A, Ajit Singh V, Elbahri H. Deep infections after endoprosthetic replacement operations in orthopedic oncology patients. Surg Infect (Larchmt). 2015; 16(3):323–332. doi:10.1089/sur.2014.049 [CrossRef]
- Mavrogenis AF, Pala E, Angelini A, et al. Infected prostheses after lower-extremity bone tumor resection: clinical outcomes of 100 patients. Surg Infect (Larchmt). 2015; 16(3):267–275. doi:10.1089/sur.2014.085 [CrossRef]
- Capanna R, Scoccianti G, Frenos F, Vilardi A, Beltrami G, Campanacci DA. What was the survival of megaprostheses in lower limb reconstructions after tumor resections?Clin Orthop Relat Res. 2015; 473(3):820–830. doi:10.1007/s11999-014-3736-1 [CrossRef]
- Peel T, May D, Buising K, Thursky K, Slavin M, Choong P. Infective complications following tumour endoprosthesis surgery for bone and soft tissue tumours. Eur J Surg Oncol. 2014; 40(9):1087–1094. doi:10.1016/j.ejso.2014.02.241 [CrossRef]
- Cho WH, Song WS, Jeon DG, Kong CB, Kim JI, Lee SY. Cause of infection in proximal tibial endoprosthetic reconstructions. Arch Orthop Trauma Surg. 2012; 132(2):163–169. doi:10.1007/s00402-011-1405-3 [CrossRef]
- Funovics PT, Hipfl C, Hofstaetter JG, Puchner S, Kotz RI, Dominkus M. Management of septic complications following modular endoprosthetic reconstruction of the proximal femur. Int Orthop. 2011; 35(10):1437–1444. doi:10.1007/s00264-010-1054-0 [CrossRef]
- Morii T, Yabe H, Morioka H, et al. Postoperative deep infection in tumor endoprosthesis reconstruction around the knee. J Orthop Sci. 2010; 15(3):331–339. doi:10.1007/s00776-010-1467-z [CrossRef]
- Myers GJ, Abudu AT, Carter SR, Tillman RM, Grimer RJ. The long-term results of endoprosthetic replacement of the proximal tibia for bone tumours. J Bone Joint Surg Br. 2007; 89(12):1632–1637. doi:10.1302/0301-620X.89B12.19481 [CrossRef]
- DiCaprio MR, Friedlaender GE. Malignant bone tumors: limb sparing versus amputation. J Am Acad Orthop Surg. 2003; 11(1):25–37. doi:10.5435/00124635-200301000-00005 [CrossRef]
- Jeys L, Grimer R. The long-term risks of infection and amputation with limb salvage surgery using endoprostheses. Recent Results Cancer Res. 2009; 179:75–84. doi:10.1007/978-3-540-77960-5_7 [CrossRef]
- Pala E, Henderson ER, Calabrò T, et al. Survival of current production tumor endoprostheses: complications, functional results, and a comparative statistical analysis. J Surg Oncol. 2013; 108(6):403–408. doi:10.1002/jso.23414 [CrossRef]
- Zmistowski B, Alijanipour P. Risk factors for periprosthetic joint infection. In: Springer BD, Parvizi J, eds. Periprosthetic Joint Infection of the Hip and Knee. New York, NY: Springer-Verlag; 2014:15–40. doi:10.1007/978-1-4614-7928-4_2 [CrossRef]
- Muilwijk J, Walenkamp GH, Voss A, Wille JC, van den Hof S. Random effect modelling of patient-related risk factors in orthopaedic procedures: results from the Dutch nosocomial infection surveillance network ‘PREZIES’. J Hosp Infect. 2006; 62(3):319–326. doi:10.1016/j.jhin.2005.08.006 [CrossRef]
- 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. doi:10.1302/0301-620X.87B6.15121 [CrossRef]
- 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. doi:10.1097/00003086-200111000-00003 [CrossRef]
- Lima AL, Oliveira PR, Carvalho VC, Saconi ES, Cabrita HB, Rodrigues MB. Periprosthetic joint infections. Interdiscip Perspect Infect Dis. 2013; 2013:542796.
- Pulido L, Ghanem E, Joshi A, Purtill JJ, Parvizi J. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res. 2008; 466(7):1710–1715. doi:10.1007/s11999-008-0209-4 [CrossRef]
- Hasan K, Racano A, Deheshi B, et al. Prophylactic antibiotic regimens in tumor surgery (PARITY) survey. BMC Musculoskelet Disord. 2012; 13:91. doi:10.1186/1471-2474-13-91 [CrossRef]
- Corten K, Bourne RB, Charron KD, Au K, Rorabeck CH. What works best, a cemented or cementless primary total hip arthroplasty? Minimum 17-year follow up of a randomized controlled trial. Clin Orthop Relat Res. 2011; 469(1):209–217. doi:10.1007/s11999-010-1459-5 [CrossRef]
- Gasparini G, De Gori M, Calonego G, Della Bora T, Caroleo B, Galasso O. Drug elution from high-dose antibiotic-loaded acrylic cement: a comparative, in vitro study. Orthopedics. 2014; 37(11):e999–e1005. doi:10.3928/01477447-20141023-57 [CrossRef]
- Gosheger G, Goetze C, Hardes J, Joosten U, Winkelmann W, von Eiff C. The influence of the alloy of megaprostheses on infection rate. J Arthroplasty. 2008; 23(6):916–920. doi:10.1016/j.arth.2007.06.015 [CrossRef]
- Wafa H, Grimer RJ, Reddy K, et al. Retrospective evaluation of the incidence of early periprosthetic infection with silver-treated endoprostheses in high-risk patients: case-control study. Bone Joint J. 2015; 97-B(2):252–257. doi:10.1302/0301-620X.97B2.34554 [CrossRef]
Details of Included Studies
|Study||Year||Level||No. of Patients (Cases)||Mean Age (Range), y||Tumor Site||PMI Rate||PMI Onset||Risk Factor Analysis|
|Dhanoa et al10||2015||IV||150 (150)||25 (7–89)||Femur, tibia, humerus, pelvis||12.38%||9 d to 63 mo||Comparative statistics, logistic regression|
|Mavrogenis et al11||2015||IV||1161 (1161)||31.2 (7–80)||Femur, tibia||8.6%||3.7 y (0.5 mo to 19 y)||Kaplan-Meier method|
|Capanna et al12||2015||IV||199 (200)||43 (12–90)||Femur, tibia||9.5%||16 (2–51) mo||Kaplan-Meier method|
|Peel et al13||2014||IV||121 (121)||38 (14–86)||Femur, pelvis, tibia, humerus||14%||541 (41–952) d||Logistic regression|
|Cho et al14||2012||IV||62 (62)||26.5 (12–65)||Proximal tibia||25.8%||34 (2–240) mo||Kaplan-Meier method|
|Funovics et al15||2011||IV||166 (166)||49.8 (5.9–84.3)||Proximal femur||7.2%||39 (0–167) mo||Kaplan-Meier method|
|Morii et al16||2010||IV||82 (82)||31.1 (5–86)||Distal femur, proximal tibia||17%||10.9 (0–48) mo||Kaplan-Meier method|
|Myers et al17||2007||IV||194 (194)||21.5 (10–74)||Proximal tibia||19.5%||9 (2–22) mo||Comparative statistics|
Risk Factors for Perimegaprosthetic Infection
| ASA score||NS10|
| Comorbidities||OR, 5.13, P=.0210|
| Diabetes mellitus||NS13|
| Increasing BMI||OR, 1.12, P=.00913|
| Superficial SSI||70% vs 9.7%, P<.00116; NS13|
| Indwelling port system||NS14|
| WBC count||NS14|
| Skin necrosis||53.8% vs 10.1%, P<.00116|
| Primary tumor vs metastasis||10% vs 32% (at 10 years), P<.00111; 15.3% vs 1.1%, P<.0115; NS10,16|
| Metastatic spread||NS10,13|
| Malignant tumor||NS10|
| Ewing sarcoma||NS11|
| Preoperative chemotherapy||NS10,11,13,14,16|
| Preoperative radiotherapy||NS10,11,15|
| Localization (overall)||NS11,12,16|
| Proximal tibia||OR, 6.27, P=.03210|
| Distal tibia||NS10|
| Distal femur||OR, 0.216, P=.03810|
| Pelvis||OR, 90.32, P=.00210|
| Pelvic extension from proximal femur||42.8% vs 3.9%, P<.00115|
| Preoperative hospitalization >24 h||OR, 6.45, P=.04810|
| Vancomycin-containing antibiotic prophylaxis||NS13|
| Intraoperative antibiotic dosing||NS13|
| No. of postoperative antibiotics||NS16|
| Length of postoperative antibiotic prophylaxis||NS13,14|
| Clean air operating room||NS16|
| Cemented vs uncemented prosthesis||9.6% vs 8.5%, P=.04611; 25% vs 5.8%, P<.0515|
| Prosthesis type||NS10,16|
| Hinge movement||NS11,14|
| Resection margins||NS10|
| Extracapsular knee resection||NS16|
| Resection length (overall)||NS10,14,16|
| Tibial resection >37%||37.5% vs 10.3%, P=.01614|
| 3 or 4 vs 1 or 2 quadriceps heads resection||23.5% vs 5.7%, P=.0416|
| Additional procedures||OR, 11.67, P=.00310|
| Surgical time (overall)||OR, 1.01, P=.00313|
| Surgical time >2.5 h||OR, 8.96, P=.00410|
| Surgical time >3 h||NS14|
| Gastrocnemius flap coverage||14% vs 31%, P=.01417; NS13,16|
| Synthetic material to fix the patellar tendon||NS14|
| Postoperative ICU admission||OR, 3.82, P=.0213|
| Postoperative hematoma||OR, 7.21, P=.0213|
| Blood loss||NS10|
| Blood units transfused (overall)||OR, 1.15, P=.0213|
| >2 blood units transfused||OR, 5.74, P=.01610; NS14|