Introduction

As total joint arthroplasty (TJA) becomes more prevalent in the USA, the burden of revision arthroplasty is expected to increase as well [1]. One of the most common reasons for revision arthroplasty is prosthetic joint infection (PJI) [1, 2], which is associated with significant morbidity and mortality [3, 4] An estimated ~ 1% of all total joint arthroplasties (TJA) result in a PJI, but it has been suggested as the driving factor behind the recent increase in revision burden [5]. Two-stage revision arthroplasty is considered by many the gold standard for treatment of a chronic PJI, particularly in the USA [6]. The first stage of a revision arthroplasty includes placement of an antibiotic-impregnated cement spacer (ACS), often containing nephrotoxic antibiotics such as vancomycin and tobramycin [7]. Two-stage revision arthroplasty has been shown to be highly effective in eradicating PJI in over 80% of cases [8]. Despite this success rate, it is associated with substantial morbidity and an alarmingly high mortality rate [3, 9]. This population tends to have a higher comorbidity burden which predisposes them to longer hospital stays, increase in overall health care dollars, and greater complication rates.

One such complication is acute kidney injury (AKI), which has been extensively studied in primary TJA but less so in the two-stage revision arthroplasty setting [10, 11]. This is clinically important because AKI carries a significant increase in patient morbidity and mortality [12]. It is important for both orthopedic surgeons and medical clinicians to be made aware of these risk factors and better help optimize patients and lead to safer outcomes for patients with chronic PJIs, particularly in a vulnerable population undergoing implantation of ACS that contain nephrotoxic agents.

The goals of this systematic review were to assess the existing literature to (1) establish an incidence of AKI following first-stage exchange arthroplasty, (2) identify perioperative risk factors for AKI risk, (3) determine the antibiotic concentration thresholds in ACS over which AKI risk is significantly increased.

Methods

Literature search strategy

This systematic review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [13]. An electronic search was conducted of the PubMed database of all studies published from inception to 1/21/2022 (see search terms in Appendix). Medical Subject Headings and Emtree terms were utilized to improve search results.

Eligibility criteria

Studies were included for review if they (1) involved subjects undergoing ACS placement in the first stage of treatment for PJI, (2) included AKI as an outcome, as defined by either the Kidney Disease Improving Global Outcomes (KDIGO) [14] or the Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease (RIFLE) [15] classifications, and (3) had N ≥ 10 (case series and case reports excluded).

Study selection

All screening and text review was conducted independently by two authors (YPC and ML). All disagreements were resolved through discussion among screening authors, with senior author involvement included as needed (RMA).

Data extraction and synthesis

All data extraction was performed by three authors (YPC, ML, and HH). Extracted variables included country in which the study was conducted, study inclusion and exclusion criteria, patient characteristics (age, BMI, sex, preexisting chronic kidney disease [CKD], baseline serum creatinine level), procedure type (knee vs. hip revision), cement spacer type, antibiotic impregnated in the spacer, quantity of antibiotic used, systemic antibiotics administered during the study period, AKI definition, and incidence of AKI. Risk factors for AKI based on univariate and multivariate analysis were also extracted, as well as the subsequent odds ratios (OR) with 95% confidence intervals (CI). Given the high heterogeneity in antibiotics and spacer types, a meta-analysis could not be performed.

Quality appraisal

Quality appraisal was performed using the Methodological Index for Non-Randomized Studies (MINORS) criteria for non-randomized studies, which is scored out of a maximum 16 points for non-comparative studies and 24 points for comparative studies [16]. This was performed independently by two authors (YPC, HH).

Results

Study characteristics

A total of 882 citations were identified (Fig. 1) and screened for inclusion. Fifty-one articles were identified for full text review, with 8 meeting the criteria for inclusion [17,18,19,20,21,22,23,24] (Table 1). Six studies were conducted in the USA [18,19,20,21,22,23]. The remaining two were conducted in Canada [17] and Germany [24]. This included 540 knee revisions and 943 hip revisions. Three studies included only knee revisions [18, 20, 23], one included only hip revisions [19], and the remaining four had both [17, 21, 22, 24]. ACS types, quantity of antibiotics added, and systemic antibiotic use are reported in Table 2. The incidence of AKI within this population was 21% (309 cases) (Table 3). The range of AKI incidence in the included studies was 10–33%. Two studies split up cases on the basis of preexisting CKD, finding higher rates of AKI in those with preexisting CKD (45% vs. 14% in 455 knee revisions [20], 28% vs. 7% in 256 hip revisions [19]).

Fig. 1
figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart

Full size image
Table 1 Descriptive characteristics of eight studies assessing rates of acute kidney injury in first-stage revision arthroplasty involving the use of antibiotic cement spacers
Full size table
Table 2 Descriptive characteristics of antibiotic cement spacers and systemic antibiotics in eight studies assessing rates of acute kidney injury in first-stage revision arthroplasty
Full size table
Table 3 Definitions, incidence, and risk factors for development of postoperative acute kidney injury
Full size table

Risk factors

Perfusion-related factors such as lower preoperative hemoglobin, higher median number of units of blood transfused or intraoperative transfusion, and postoperative hypovolemia were commonly found to have an association with increased risk for AKI. Other commonly reported risk factors included increasing age, perioperative use of nephrotoxic medications such as nonsteroidal anti-inflammatory drugs, and comorbidity burden (CKD, diabetes mellitus, hypertension, cardiovascular disease, etc.). Six studies performed multivariate analysis along with univariate analysis, while two studies only performed univariate analysis [17, 23]. One study found no predictors associated with AKI risk on either univariate or multivariate analysis [21], while another found several univariate predictors but none after multivariate analysis [19]

ACS antibiotic concentration

Two studies reported thresholds of antibiotic concentration in ACS associated with increased AKI risk. One found that > 3.6 g of vancomycin or aminoglycosides per batch of cement in first-stage knee revisions was associated with increased AKI risk [20]), while another demonstrated > 4 g of vancomycin or > 4.8 g of tobramycin in an ACS was associated with increased risk [23]. The remaining six studies found no associations with antibiotic dose in ACS and AKI [17,18,19, 21, 22, 24].

Quality appraisal

The results of quality appraisal for the included studies are presented in Table 4.

Table 4 Quality appraisal of eight studies assessing rates of acute kidney injury in first-stage revision arthroplasty involving the use of antibiotic cement spacers
Full size table

Discussion

Acute kidney injury (AKI) has been shown to increase morbidity and mortality following revision TJA. The incidence of AKI after the first stage of a two-stage revision arthroplasty in the current literature ranges between 10 and 33%. Although current evidence is limited and precludes meta-analysis, the most commonly associated risk factors reported include perfusion-related factors (low hemoglobin, perioperative transfusion, hypovolemia), older age, increased comorbidity burden, and perioperative nephrotoxic drug use. Only two studies found that certain antibiotic concentration thresholds in ACS were linked to increased AKI risk, but neither of these were the result of multivariate analysis accounting for other risk factors.

A similar systematic review performed by Luu et al. in 2013 on ten studies consisting of 544 patients reported an AKI incidence of approximately 5% [25], a substantially lower figure than the 21% reported in our study. One possible reason for this discrepancy is the difference in definition of AKI. The current study restricted inclusion criteria to studies using either KDIGO or RIFLE definitions for AKI, whereas the review performed by Luu et al. did not have such restrictions, including many studies that did not specify a threshold. Additionally, of the ten studies in their review all were performed in 2012 or earlier, while the studies included in the current review were performed between 2012 and 2021. Finally, the current review included 1483 total combined revision hip and knee cases, approximately 900 greater than the Luu et al. study. Other recent studies provide similar rates of AKI after ACS placement—a retrospective review conducted in 2018 by Yadav et al. of 3,218 revision TJA found a 29% incidence of AKI after ACS implantation, compared to 3% in aseptic revision cases [26]. AKI in the first stage of revision arthroplasty is an underreported aspect of PJI management in current literature and is a uniquely widespread issue for this patient population, especially when compared to primary or aseptic revision TJA.

In assessing the risk factors for AKI following ACS implantation, the most commonly reported were related to poor perfusion, such as low hemoglobin, and perioperative blood transfusions or hypovolemia. These are well-established risk factors in primary TJA literature, but it is important to recognize the increased rates of blood loss and transfusion with septic revision arthroplasty [27, 28]. This may be due to increased bony work required when explanting prostheses (particularly with the removal of cemented components), edematous and inflamed tissue in the setting of infection, or prolonged operative time. Other risk factors identified in the review of selected studies included use of nephrotoxic agents such as NSAIDs, increasing age, or an increased burden of medical comorbidities such as CKD, diabetes mellitus, or hypertension. These risk factors are prevalent in this population, and it is important for patients to be medically optimized prior to undergoing revision arthroplasty. A review of comorbidities in TKA patients at risk for postoperative mortality and PJI using the Medicare database by Bozic et al. demonstrated that 27% of PJI patients have heart disease, 22% have diabetes, and 70% have hypertension, a major risk factor for developing CKD [29]. Other potential avenues to reduce AKI risk perioperatively could involve the use of single-stage revision in select patients, as this would avoid the implantation of an ACS and involve decreased blood loss [30]. While two-stage revisions are widely referred to as the gold standard of treating PJI, several recent meta-analyses suggest that single-stage revisions may have similar success [8, 31,32,33]. However, further study is warranted to identify the risk of AKI solely attributed to an ACS. Additionally, while restrictive transfusion protocols are being used widely for primary TJA [34, 35], it may be worth weighing the benefits of a less restrictive policy in a population receiving several nephrotoxic agents to avoid a state of hypoperfusion.

Regarding the ideal concentration of antibiotics in an ACS, this review included two studies that demonstrated ACS antibiotic concentration thresholds associated with increased risk of AKI [19, 20]. Dagneaux et al. found that ACS concentrations > 3.6 g of vancomycin or aminoglycoside were associated with an increased risk of AKI in a retrospective review of 455 cases involving the first stage of revision knee arthroplasty [20]. Furthermore, they found that the nephrotoxic characteristics of these antibiotics worked synergistically as AKI rates were higher when using both when compared to one or the other. In a retrospective review of 84 cases of revision TKA, Menge et al. demonstrated that ACS containing > 4.0 g of Vancomycin or > 4.8 g of Tobramycin had a statistically significant association with the development of an AKI (100% of patients were found to have AKI post-operatively) [23]. Both studies only reported these associations of ACS concentrations with AKI on univariate analysis, however, and the remaining six studies in this review were not able to demonstrate any association with ACS concentration. Additionally, in the Dagneaux et al. paper, the association between spacer antibiotic dosage and AKI rate was not statistically significant after accounting for other factors in a multivariate analysis. In a retrospective review of 91 hip revisions and 156 knee first-stage revisions, Geller et al. demonstrated that prefabricated ACS were associated with increased risk of AKI but did not specify the antibiotic concentration in the spacers [22]. A difficult aspect of identifying AKI risk and the role of ACS concentrations in these cases as there are several perioperative factors such as nephrotoxic agents, preexisting comorbidities, differing degrees of blood loss, and systemic antibiotics that may all play a role. Additionally, the different concentrations of antibiotics used in various studies may also have affected the study outcomes. Even with higher concentrations of spacer antibiotic dosages, several studies like those performed by Aeng et al. and Theil et al. still were unable to demonstrate an association between concentration and AKI risk [17, 24]. Finally, there is a potentially enormous bias in the presence of systemic antibiotic use, as many of these carry substantial nephrotoxic side effects [36, 37]. This is well discussed in the literature, and while high doses of local antibiotics present in spacers may provide a contributing factor there is scarce literature on how this may contribute to kidney injury. The ideal dosage to add to a cement spacer is also controversial, as amount of antibiotic added can also contribute to mechanical weakening of the spacer as well [38]. Further study is warranted to determine safe thresholds, particularly in patients with poor preoperative kidney function. Patients with underlying CKD were found to have nearly a five-fold increase in AKI on CKD compared to those without renal issues preoperatively [23]. Orthopedic surgeons and medical clinicians alike need to be mindful of the type and amount of antibiotic used when trying to safely eradicate infection in these patients.

A significant limitation of the current review was the heterogeneity of the included studies with regards to ACS type, concentrations, concomitant systemic antibiotics, and study design. This variation unfortunately did not allow for meta-analysis. Additionally, AKI is a multifactorial diagnosis—it is difficult to state a cause–effect relationship regarding systemic antibiotic use in two-stage revision. Strengths of this review include the strict inclusion criteria when screening studies, particularly the requirement for a clear and widely recognized definition for AKI. Additionally, revision TJA for PJI is difficult to study in volume, as coding for these procedures is notoriously inaccurate making database studies unreliable [39]. Large databases may not have access to more granular data such as ACS use, antibiotics included in the ACS, or systemic antibiotics. Thus, analysis of these risk factors is limited to level 2 and 3 studies such as those included in this review.

Conclusion

The high incidence rates of AKI following the first stage of two-stage revision arthroplasty must be considered by providers in the treatment of chronic PJI. Identification of associated risks should be used to better guide preoperative optimization and treatment of these patients, such as reducing the risk of perioperative hypovolemia. Further study is warranted to identify the AKI risk attributed to ACS versus other potential causes of nephrotoxicity as well as the ideal dosage of antibiotics added to ACS constructs.