1 Introduction

Total knee arthroplasty (TKA) is one of the most common orthopedic procedures with an incidence of 8.7 per 1000 U S. Medicare enrollees.¹ The incidence and utilization of TKA has continued to increase, with primary TKA rates nearly tripling over a 12-year period in the U.S., a trend reflected worldwide.^1–3 The burgeoning popularity of TKA can be attributed to its ability to produce significant, rapid improvements in quality of life for patients who suffer from end-stage arthritis as well as various chronic knee pathologies. The most common indication for TKA is osteoarthritis, a degenerative condition that increases in prevalence with age and is the underlying cause in 95% of TKAs performed in the U.S.⁴ Unfortunately, many other comorbidities are associated with advanced age, which may contribute to increased complication rates after TKA. An example of a relatively common medical comorbidity is end stage renal disease (ESRD), which has more than doubled in prevalence in people over the age of 65 between 2000 and 2008.⁵ The main reason for the increase in ESRD can be attributed to chronic diabetes and hypertension.⁵ Dialysis, administered as either hemodialysis or peritoneal dialysis, is the treatment for ESRD.

The presence of kidney disease and therapeutic dialysis may increase the medical and surgical risk in patients undergoing TKA. Renal osteodystrophy, seen most commonly in dialysis patients due to its association with long-standing renal disease, is characterized by metabolic abnormalities, such as hypocalcemia and hyperphosphatemia. The hypocalcemia in turn can induce secondary hyperparathyroidism, resulting in osteomalacia and decreased bone quality.^6,7 Additionally, dialysis-related amyloidosis, stemming from deposition of ß2-microglobulin, presents with bony cysts that may further compromise underlying bone stock.^8–10

Local and systemic infections are also a major concern. Native septic arthritis is more commonly seen in dialysis patients secondary to their immunodeficient status from ESRD combined with the risk of recurrent bacteremia from dialysis treatments.^9–11 Furthermore, antibiotic selection and appropriate antibiotic dosing for the prevention and treatment of periprosthetic joint infection (PJI) is more challenging in patients undergoing dialysis. Commonly used antibiotics for preoperative prophylaxis include cefazolin, the first-generation cephalosporin, and vancomycin, which is used as adjunctive prophylaxis when there is concern for methicillin-resistant Staphylococcus aureus (MRSA) and skin colonization.^12,13 However, both are primarily renally eliminated, requiring adjusted doses in dialysis patients. If a PJI does occur and exchange arthroplasty is indicated, antibiotic-impregnated spacers containing aminoglycosides and occasionally vancomycin are commonly used.^14–16 Therefore, the nephrotoxicity and systemic toxicity of these antibiotics create limitations for their use in dialysis patients, which raises concern for PJI treatment outcomes.

Dialysis treatment thus presents several challenges to the TKA patient. A few systematic reviews have investigated ESRD patients receiving dialysis who underwent total hip arthroplasty (THA) and reported higher rates of revision and PJI compared to non-dialysis controls and nationally reported standards.^17,18 However, a systematic review investigating the impact of ESRD and dialysis on outcomes of patients after primary TKA, including perioperative and postoperative complications, has not been reported. The purpose of this systematic review is to evaluate ESRD patients on dialysis undergoing primary TKA and the effect on postoperative complications, including PJI, reoperation and revision, morbidity, and mortality, in order to better estimate the risk of TKA in this patient population.

2 Materials and methods

The data collection for this systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁹

3 Results

3.1

3.1 Study selection

125 distinct papers were initially generated from the search criteria (Fig. 1). These were subjected to initial title and abstract screening for inclusion. 14 papers passed this round and subsequently underwent full text screening for exclusion, which resulted in 9 final papers for this systematic review. The 5 studies that were excluded at this stage included 2 studies that did not distinguish patients who received a primary TKA from a primary THA, 1 study that did not distinguish patients who received dialysis from those who received renal transplants, 1 study that did not distinguish which ESRD patients underwent dialysis from those who did not, and 1 study that did not explicitly specify that dialysis was the treatment ESRD patients received. All 9 included studies were retrospective cohort studies.^23–31 Four reported on dialysis patients only,^28–31 while the other 5 were comparative studies that additionally reported outcomes for a control, non-dialysis cohort.^23–27 The total mean modified Coleman score of this review was 45.4 ± 3.4, which indicates moderate overall study methodology.

Close

Fig. 1 Flow chart diagram.

Export to PPT

3.3

3.3 Overall complications

Eight studies reported overall complication rates following primary TKA producing a cumulative rate of 25% (N = 925/3702) for patients who received dialysis.^{23–26,28–31} Of these, 4 were comparative studies, with a cumulative rate of 26.8% (N = 894/3334) for the dialysis group compared to 7.9% (N = 363,659/4,560,023) for non-dialysis controls.^23–26 This difference was statistically significant (OR 2.26, 95% CI: 1.78–2.88, I² = 80.1%, p < 0.001) (Fig. 2).

Close

Fig. 2 Forest plot comparing overall complications after primary Total Knee Arthroplasty (TKA) in dialysis patients to non-dialysis controls. OR, odds ratio; CI, confidence interval.

Export to PPT

3.4

3.4 Death

Eight studies reported on death following primary TKA, with a cumulative mortality rate of 2.5% (N = 92/3702) for patients who received dialysis.^{23–26,28–31} Of these, 4 studies were comparative, with a cumulative mortality rate for the dialysis group of 2.4% (N = 79/3334) in comparison to the controls’ cumulative rate of 0.10% (N = 4607/4,560,023).^23–26 This difference was statistically significant (OR 8.60, 95% CI: 6.03–12.26, I² = 0.0%, p < 0.001) (Fig. 3).

Close

Fig. 3 Forest plot comparing mortality after primary Total Knee Arthroplasty (TKA) in dialysis patients to non-dialysis controls. OR, odds ratio; CI, confidence interval.

Export to PPT

3.5

3.5 Infection

Eight studies reported on postoperative infection after TKA.^23–29,31 However, there was variation in the definition that the studies used. Two studies used International Classification of Diseases (ICD-9) codes to identify patients.^23,26 Notably, one of these was significantly more expansive, including codes 998.51, 998.59, 711.06, 711.46, 711.86, 711.96, 996.66, 996.67, and 996.69 under the umbrella category of PJI.²³ The other study included only ICD-9-CM codes, 996.6 and 998.5, for wound infections.²⁶ The remaining 6 studies provided descriptive categorizations. Four of these studies specifically included both deep (deep surgical site infection, PJI) and superficial (superficial surgical site infection, wound necrosis/infection) site infections in their infection definition.^24,25,28,31 The fifth study limited their definition to just PJIs,²⁹ while the sixth did not provide any details beyond simply “infections”.²⁷ None of these 6 studies detailed how the infection diagnosis was made, such as using the Musculoskeletal Infection Society (MSIS) criteria or positive synovial cultures.

The cumulative infection rate for dialyzed patients across the 8 studies was 2.5% (N = 90/3578).^23–29,31 All 5 comparative studies reported on infections, with a cumulative rate of 2.55% (N = 89/3496) for dialyzed patients, in comparison to the cumulative rate for the control group of 0.35% (N = 15,928/4,615,593).^23–27 This difference was statistically significant (OR 2.47, 95% CI: 1.83–3.35, I² = 23.4%, p < 0.001) (Fig. 4).

Close

Fig. 4 Forest plot comparing infections after primary Total Knee Arthroplasty (TKA) in dialysis patients to non-dialysis controls. OR, odds ratio; CI, confidence interval.

Export to PPT

3.6

3.6 Reoperation

Seven studies included reoperation in their postoperative complications.^{23–25,27–29,31} The general categories of the additional procedures consisted of revision surgery in 2 studies,^23,27 arthrotomy/debridement in 2 studies,^29,31 and a generic “return to the operating room” in 2 studies.^24,25 The final study reported zero reoperations. The cumulative reoperation rate for dialyzed patients across the 7 studies was 3.74% (N = 71/1895). Of these studies, 4 were comparative studies, with a cumulative rate of 3.81% (N = 69/1813) for dialyzed patients, in comparison to the controls’ cumulative rate of 1.52% (N = 6598/434,389).^23–25,27 This difference was statistically significant (OR 2.12, 95% CI: 1.40–3.20, I² = 51.9%, p < 0.001) (Fig. 5).

Close

Fig. 5 Forest plot comparing reoperations after primary Total Knee Arthroplasty (TKA) in dialysis patients to non-dialysis controls. OR, odds ratio; CI, confidence interval.

Export to PPT

3.7

3.7 Length of hospital stay

Three comparative studies reported statistics for length of hospital stay following primary TKR.^24–26 Dialysis patients who underwent TKA had a mean hospital stay of 5.8 ± 1.2 days, while the control groups had a mean hospital stay of 2.9 ± 0.9 days. This difference was also statistically significant (WMD: 2.71, 95% CI: 1.48–3.95, I² = 94.8%) (Fig. 6).

Close

Fig. 6 Forest plot comparing length of hospital stay after primary Total Knee Arthroplasty (TKA) in dialysis patients to non-dialysis controls. WMD, weighted mean difference; SD, standard deviation; CI, confidence interval.

Export to PPT

4 Discussion

End stage renal disease requiring dialysis is a chronic condition that is increasing in prevalence and which most commonly affects older patients.⁵ This is concerning to the orthopedic surgeon since this is also the age group that predominantly undergoes TKA as they also often have concomitant degenerative joint disease.^1,4 The combination of ESRD and dialysis treatment may place this population at a greater risk for postoperative complications including infection, reoperation, and mortality, but this risk has not been quantified in large scale studies.^6–11 To our knowledge, no prior study has analyzed the risk that dialysis patients encounter when undergoing TKA. The purpose of this systematic review and meta-analysis is to better understand the postoperative complication rates following TKA in dialysis patients compared to non-dialysis controls.

The results from the 9 reviewed studies, which includes a total of 3864 patients on dialysis for ESRD, clearly demonstrate that dialysis patients have higher rates of all complication categories, including infection, reoperation, and mortality, compared to controls. We also found a significantly greater length of hospital stay among dialysis patients, which in conjunction with increased complications means that performing a TKA in these patients is associated with increased cost and health care utilization.

The complication rates reported in the control TKA patients included in the comparative studies is similar to other reported complication rates in the literature, supporting the reliability of these comparisons. Following standard primary TKAs in the general population, 90-day mortality has been reported to be 0.19% in a recent meta-analysis.³² The incidence of PJI is estimated to be 0.8%–1.9%,^33,34 while reoperation rate has been reported at 1.18%³⁵ and length of hospital stay in the majority of patients is generally less than 3 days.³⁶

Of all the complications included in this review, mortality rate showed the most striking difference in dialysis patients following primary TKA. As suggested elsewhere in the literature, this difference likely reflects a dose-dependent relationship of kidney disease severity, in addition to possible effects of dialysis treatment itself.^27,37 For example, a study of chronic kidney disease (CKD) patients, encompassing primarily those with disease on the milder end of the spectrum in addition to a minority of patients with ESRD on dialysis, found no differences in 30-day and 90-day mortality compared to TKA patients without CKD.³⁷ There were also no differences in rates of deep surgical site infections, pulmonary embolisms, and deep vein thromboses. However, when mortality at any point during their median 2.1-year follow-up was considered, a significant difference was found between the 2 groups.³⁷ This reflects chronic changes in health that are more attributable to CKD than surgical sequelae, the latter of which would be expected to present more acutely.

In contrast, although the studies in our review tracked mortality to varying follow-up timepoints, the 2 comparative studies with the shortest follow-ups (both 30 days postoperative) still noted high mortality rates of 1.2%²⁴ and 0.88%,²⁵ in comparison to their control rates of 0.11%²⁴ and 0.10%,²⁵ respectively. When considered alongside the concomitant differences in postoperative complications detailed above, our mortality findings likely stem from a higher predisposition to surgical complications in these patients due to much greater kidney disease severity and exposure to the deleterious effects of dialysis. In particular, PJIs represent a significant concern in dialysis patients, as the requirement of adjusted antibiotic regimens may make it more difficult for patients to clear the infection, ultimately putting them at greater risk of bacteremia and life-threatening sepsis. Moreover, the presence of ESRD already places patients at a minimum of an ASA III classification preoperatively,³⁸ and is also strongly associated with cardiovascular disease and diabetes, two comorbidities that make patients worse surgical candidates in general.³⁹

Our study is the first systematic review to compare the outcomes of dialysis and non-dialysis patients undergoing TKA. Two prior systematic reviews evaluated the impact of dialysis on outcomes after THA.^17,18 Those studies both compared the dialysis patients to those who received renal transplants, but did report absolute rates for deep infections and revision surgeries in the dialysis-only cohort. Dialysis patients undergoing THA had high rates of deep infections (cumulatively 8.5%¹⁷ and 10.8% ¹⁸) and revision surgeries (cumulatively 16.3%¹⁷ and 15.7% ¹⁸). The complication rates after THA in dialysis patients are even higher than similar complications in dialysis patients after TKA. However, these reviews had significantly longer follow-up, which limits the ability to make direct comparisons between THA and TKA in dialysis patients, yet they highlight the increased risk of arthroplasty found in this patient cohort.

The impact of these higher complication rates on patient outcomes and overall healthcare costs are important considerations for dialysis patients and their providers to consider when deciding between undergoing a TKA versus more conservative treatment approaches. Counseling of ESRD patients on dialysis prior to TKA should include the standard risk assessment, including detailed discussion of their higher mortality risk. For these patients, however, greater consideration should be given to the prevention of PJI. A multidisciplinary approach, including input from infectious disease specialists and the patients’ nephrologists, could greatly aid in optimizing preoperative antibiotic prophylaxis, as well as put contingency plans in place for timely intervention. If a PJI does occur, an aggressive approach should be pursued, with exchange arthroplasty likely to yield superior results to debridement, antibiotics, and implant retention (DAIR) given the increased risk of re-infection seen in DAIR and challenges with long term antibiotic suppression.⁴⁰ Moreover, alternative PJI treatments that do not involve arthroplasty, such as permanent resection, arthrodesis, or amputation should be strongly considered, but the long-term outcomes of dialysis patients after reimplantation is still unknown.

There are several limitations to this review. First, only one study distinguished patients who underwent peritoneal dialysis from those on hemodialysis. As such, we were unable to compare postoperative complications for these two types of dialysis.²³ However, this study showed that peritoneal dialysis patients had lower rates of infection compared to hemodialysis patients after TKA with comparable results for other complications. This important finding calls for further research on peritoneal dialysis and complications after THA and TKA.²³ In addition, only 3 of the 5 comparative studies assessed for differences in baseline comorbidities between dialysis and non-dialysis control groups. Future studies should focus on better controlling for these confounding risk factors to isolate the influence of dialysis on postoperative complications. Additionally, the mean follow-up was fairly short, with over half the studies reporting complications up to one year postoperatively. Time of last follow-up has the greatest impact on reoperation rate, with Postler et al. finding that 73.7% of first revision surgeries following primary TKA took place after 2 years.⁴¹ Non-revision reoperations (NRR) tend to take place earlier in the postoperative period, yet 16% of NRRs still were found to take place following the first postoperative year.⁴² Thus, although the trend of higher reoperation rates in dialysis patients compared to controls should still hold true, the reported absolute rates are likely to be an underestimate.

5 Conclusion

Patients with ESRD who are on dialysis have a significantly higher complication rate of major complications following TKA compared to non-dialysis patients. Dialysis patients undergoing TKA have a significantly higher overall complication rate, mortality rate, periprosthetic infection rate, reoperation rate, and length of hospital stay. Dialysis patients undergoing TKA require additional counseling on risks and benefits and will likely require additional post-operative health care resources and associated cost. Further research should be dedicated to distinguishing between the outcomes associated with hemodialysis and peritoneal dialysis, especially for deep infection, given the relative dearth of these studies in the literature.

Funding

Not applicable.

Availability of data and material

Not applicable.

Code availability

Not applicable.