Abstract
Background
Septic arthritis is a debilitating diagnosis, and many of these patients elect to undergo total knee arthroplasty (TKA) and total hip arthroplasty (THA). The purpose of this systematic review and meta-analysis is to assess the risk of prosthetic joint infection (PJI) in those who undergo primary TKA or THA with a history of septic joint arthritis. Secondarily, we will evaluate patient-specific or surgical factors that could increase the risk of primary arthroplasty failure.
Methods
A comprehensive search of PubMed, CINAHL Plus, Embase, and Scopus was performed. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed for study selection, data extraction, and analysis. From a total of 404 articles, 18 studies were selected for inclusion.
Results
Our final cohort consisted of 1,758 arthroplasties with Staphylococcus species (n = 375 cases, 40%) being the most common causative septic arthritis agent. Male sex (70.5% vs. 51.1%, p = 0.0003) and Gram-negative bacteria cause (11.6% vs. 6.0%, p = 0.02) were associated with an increased risk of PJI following arthroplasty. There was no association found between arthroplasty failure and increasing age or body mass index. There was no difference in PJI rates between 1- and 2-stage arthroplasty (10.9% vs. 10.8%; p = 0.47).
Conclusions
Male sex and resistant organism as causes of septic arthritis were associated with an increased risk of PJI in patients with a history of native septic arthritis. Single- and 2-stage procedures had similar success rates. Future studies should be larger and focus on long-term outcomes in these patients.
Keywords: Hip arthroplasty, Knee arthroplasty, Prosthetic joint infections, Septic arthritis, Risk factors
Septic arthritis is a severe diagnosis that can lead to substantial pain and disability.1,2,3) Without appropriate treatment, it may lead to the irreversible destruction of cartilage and subsequent end-stage arthritis.4) It more commonly affects monoarticular joints such as the hip or knee.4,5) Many of these patients may eventually elect to undergo total hip arthroplasty (THA) and total knee arthroplasty (TKA) for improved pain relief and quality of life.2,6) However, those with a previous history of native septic joint arthritis may have an increased risk of multiple complications, including prosthetic joint infection (PJI), when compared to those with primary end-stage arthritis.6)
The risk of PJI in those who have a history of septic arthritis has been reported to range from 1% to 15% in several small studies.6,7,8,9) Despite this higher risk, these can be mitigated by various techniques such as the use of antibiotic spacers, antibiotic beads, or by waiting a specific time interval prior to proceeding with arthroplasty.5,7,10) However, there have been limited guidelines for when to perform a primary THA and TKA in someone with a history of septic joint arthritis.7,11) Although there have been some small studies evaluating the risks associated with PJI following primary arthroplasty in those with a history of septic arthritis, there is a need for a concise meta-analysis regarding safety and efficacy.
Therefore, the purpose of this systematic review and meta-analysis is to assess the risk of PJI in those who undergo primary TKA or THA with a history of septic joint arthritis. Secondarily, we will evaluate patient-specific or surgical factors that could increase the risk of primary arthroplasty failure.
METHODS
This study was exempt from Institutional Review Board approval and informed consent as it utilized previously published data with no identifiable participant information. This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A comprehensive search of 4 different electronic databases (PubMed, CINAHL Plus, Embase, and Scopus) was performed for all articles pertaining to this topic from January, 2014 to December, 2024. The following search strings were used: arthroplasty*[title] AND septic[title] AND arthritis[title]; total[title] AND knee[title] OR hip[title] AND septic[title] AND arthritis[title].
Studies that evaluated the outcomes of TKA and THA in those with a previous history of native septic joint arthritis were considered for inclusion in this study. The independence of various data subsets was confirmed by evaluating the time interval and institutions at which the studies took place. The following studies were excluded from this meta-analysis: (1) animal and lab studies; (2) non-English manuscripts; (3) current reviews or expert opinions; (4) isolated case reports; and (5) inadequate raw data. Each study was evaluated by 2 authors (SKS and MT) with all disputes being resolved by a third author (TPP).
Our study resulted in 404 studies. The following studies were excluded for the following reasons: (1) 22 non-human studies; (2) 3 non-English manuscripts; (3) 28 current reviews; and (4) 27 isolated case reports; and (5) 9 studies with inadequate raw data. Of the remaining 315 studies reviewed, 18 were selected for inclusion (Fig. 1, Table 1).5,6,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23) The quality of all included studies was evaluated using the Assessment of Quality in Lower Limb Arthroplasty (AQUILA) tool.
Fig. 1. Literature search flow diagram showing how a total of 18 studies were included.
Table 1. Studies Included.
| Study | Total (n) | PJI | Mean age (yr) | Male | Female | BMI (kg/m2) | Mean follow-up (yr) | Time from septic arthritis to TJA (yr) | MRSA | MSSA | Other Staphylococcus species | Strep species | Other resistant organisms | Other Gram positive | Anaerobes | Gram negatives | Polymicrobial | Fungus | Mycobacterium | Culture negative | Unspecified |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ohlmeier et al. (2020)12) | 68 | 2 | 62.8 ± 11.1 | 43 | 25 | 29.9 ± 5.3 | 5 ± 2.5 | 9.6 ± 14.2 | 3 | 24 | 19 | 4 | 3 | 1 | 8 | 22 | - | - | 6 | - | |
| Ni et al. (2020)10) | 24 | 0 | 61.6 ± 7.7 | 7 | 17 | - | 2.25 | - | - | 3 | 3 | - | - | 4 | - | - | 1 | 4 | - | 8 | - |
| Russo et al. (2021)15) | 47 | 2 | 55.9 ± 14.3 | 25 | 22 | 27.2 ± 4.1 | 7.1 ± 1.3 | - | 20 | 19 | 15 | 3 | - | - | - | 11 | 16 | 3 | 23 | - | |
| Wei et al. (2022)13) | 105 | 17 | 57.8 ± 13.2 | 63 | 42 | 25.2 ± 3.9 | 10.3 | 3.9 ± 7.78 | 14 | 18 | 5 | - | 15 | 1 | - | 12 | 4 | 3 | 7 | 43 | - |
| Portier et al. (2022)23) | 47 | 5 | - | 29 | 18 | - | - | 2 | 21 | 6 | 14 | - | - | 6 | 6 | - | - | - | 9 | ||
| Xu et al. (2019)14) | 74 | 9 | 49.4 ± 16.5 | 47 | 27 | 24.7 ± 4.4 | 4.7 | - | - | 6 | 15 | - | 6 | - | - | 7 | 6 | - | - | 24 | 12 |
| Zeng et al. (2019)18) | 45 | 1 | 45.9 ± 12.9 | 23 | 22 | 22.8 ± 3.1 | 6.4 | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Luo et al. (2019)22) | 101 | 0 | 52.3 | 51 | 50 | 23.3 | 6.1 | 24 | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Wang et al. (2019)19) | 56 | 1 | 47.8 | 23 | 33 | 22.5 | 10.7 | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Yang et al. (2019)20) | 49 | 0 | 44.3 ± 6.6 | 23 | 26 | 22.4 ± 3.6 | 8.7 | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Kunze et al. (2020)24) | 42 | 2 | 58.3 ± 15.1 | 26 | 14 | 30.9 ± 5.8 | 3.3 ± 1.7 | - | 4 | 6 | 10 | 3 | 1 | - | - | 3 | - | - | 11 | 5 | |
| Seo et al. (2014)11) | 62 | 6 | 67 ± 9.5 | 15 | 47 | - | 6.1 | - | 17 | 14 | 4 | 2 | - | 8 | 4 | - | - | 13 | - | ||
| Shaikh et al. (2014)25) | 13 | 0 | 65.46 ± 11.34 | 5 | 8 | - | 4 | - | 2 | 1 | - | - | - | - | - | 1 | 2 | 7 | - | ||
| Russo et al. (2024)5) | 114 | 9 | 67.7 ± 9.1 | 60 | 54 | 26.7 | 6 ± 1.9 | - | 4 | 13 | 3 | 3 | - | - | - | 4 | 4 | - | 4 | 11 | - |
| Tan et al. (2021)8) | 207 | 25 | 55.5 ± 14.4 | 103 | 104 | 27 ± 5.9 | 4.8 ± 3.1 | - | - | 38 | 18 | 12 | 13 | - | - | 17 | 15 | - | - | 51 | 17 |
| Tan et al. (2021)16) | 233 | 29 | 55.5 ± 14.4 | 118 | 115 | 27.1 ± 5.9 | 5.2 | - | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Bettencourt et al. (2021)6) | 215 | 16 | 62.9 ± 11.5 | 135 | 80 | 30.5 ± 5.4 | 9 | 4 | 11 | 25 | 65 | 20 | - | 5 | 7 | 14 | 2 | 1 | 10 | 55 | |
| Bettencourt et al. (2022)17) | 256 | 26 | 57.8 ± 11.4 | 156 | 100 | 30.6 ± 6.6 | 11 | - | 7 | 15 | 47 | 19 | - | 2 | 4 | 12 | 13 | 2 | 8 | 14 | 116 |
| Total | 1,758 | 150 | - | 952 | 804 | - | - | - | 84 | 189 | 220 | 82 | 38 | 15 | 4 | 96 | 105 | 13 | 23 | 221 | 214 |
Values are presented as mean ± standard deviation.
n: number of cases, PJI: prosthetic joint infection, BMI: body mass index, TJA: total joint arthroplasty, MRSA: methicillin-resistant Staphylococcus aureus , MSSA: methicillin-sensitive Staphylococcus aureus.
All data were inputted into a Microsoft Excel spreadsheet (Microsoft Corp.). Meta-analysis was conducted to estimate pooled outcomes regarding revision rates, etiologies of native infection, and patient characteristics. For the purposes of this study, 1-stage revision was considered the use of primary joint arthroplasty components, as opposed to the use of an antibiotic spacer followed by revision arthroplasty in the 2-stage revision cohort. To account for heterogeneity with continuous variables, we used the I2 statistic. Publication bias was assessed using Egger’s test. The Z test for populations was used to determine possible differences between proportions. All statistical calculations were performed using MedCalc (MedCalc Software Ltd.). A p-value of less than 0.05 was considered statistically significant.
RESULTS
Our final cohort consisted of 1,758 arthroplasties, including 952 men and 804 women. There were 150 diagnosed PJIs (8.5%). Mean follow-up ranged from 2 to 10 years among the studies. Mean ages ranged from 49 to 68 years (Table 1). When evaluating the etiologies of native septic joint, the most common finding was negative cultures (n = 221, 12.6%). When cultures were positive, the most common source of infection was unspecified Staphylococcus species (n = 220, 12.5%). The next most common etiologies, in descending order, were methicillin-sensitive Staphylococcus aureus (MSSA) (n = 189, 10.8%), polymicrobial (n = 105, 6.5%), Gram-negative bacteria (n = 96, 5.5%), methicillin-resistant Staphylococcus aureus (MRSA) (n = 84, 4.8%), Streptococcus species (n = 82, 4.7%), unspecified resistant organisms (n = 38, 2.2%), Mycobacterium species (n = 23, 1.3%), other Gram-positive bacteria (n = 15, 0.9%), and fungus (n = 13, 0.7%) (Table 2).
Table 2. Etiology of Septic Arthritis and Risk of PJI Following TJA.
| Etiology | % PJI | % control | p-value |
|---|---|---|---|
| Polymicrobial | 8.4 | 6.1 | 0.40 |
| Unspecified resistant organism | 8.4 | 4.0 | 0.05 |
| MRSA | 6.3 | 4.3 | 0.37 |
| MSSA | 15.8 | 13.1 | 0.48 |
| Other Staphylococcus | 9.5 | 10.0 | 0.87 |
| Streptococcus | 3.2 | 3.4 | 0.90 |
| Other Gram-positive bacteria | 2.1 | 0.4 | 0.05 |
| Gram-negative bacteria | 11.6 | 6.0 | 0.02 |
| Mycobacterium | 0 | 1.0 | 0.33 |
| Fungus | 1.1 | 0.3 | 0.25 |
| Culture negative | 11.6 | 14.6 | 0.44 |
PJI: prosthetic joint infection, TJA: total joint arthroplasty, MRSA: methicillin-resistant Staphylococcus aureus, MSSA: methicillin-sensitive Staphylococcus aureus.
There were 7 studies (n = 796) that could be evaluated for variables that increased the risk of failure of primary arthroplasty. When evaluating the role of Gram-negative bacteria etiology for native septic joint and risk of PJI following arthroplasty, the fixed and random effects rate of PJI in those with original Gram-negative septic joints was 18.1% (95% CI, 10%–28%) (Fig. 2A). This cohort showed no heterogeneity (p = 0.57; I2 = 0%) or publication bias (p = 0.3) (Fig. 2B). Those who did not have a PJI following arthroplasty with a previous native septic joint due to Gram-negative bacteria had a fixed and random effects rate of 8.8% (95% CI, 6.5%–11.4%) (Fig. 3A). There was no heterogeneity (p = 0.62, I2 = 0%) or publication bias (p = 0.35) among the cohort (Fig. 3B). Comparatively, etiology of Gram-negative bacteria (relative risk [RR], 1.93; 95% CI, 1.03–3.62; p = 0.02) conferred a higher risk of PJI (Table 3). No other etiology of native septic joint showed association with failure of primary arthroplasty (Table 2).
Fig. 2. (A) Fixed- and random-effects rates of Gram-negative prosthetic joint infection (PJI). (B) Bias assessment for Gram-negative PJI.
Fig. 3. (A) Fixed- and random-effects rates of Gram-negative controls. (B) Bias assessment for Gram-negative controls.
Table 3. Gram-Negative Etiology of Septic Joint Association with PJI.
| Cohort | Fixed effects rate (%) | 95% CI (%) | Random effects rate (%) | 95% CI (%) | Heterogeneity p-value (I2, %) | Publication bias (p-value) | Relative risk | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| Gram-negative controls | 8.8 | 6.5–11.4 | 8.8 | 6.5–11.4 | 0.62 (0) | 0.35 | Reference | Reference | Reference |
| Gram-negative PJI | 18.1 | 10–28 | 18.1 | 10–28 | 0.57 (0) | 0.3 | 1.93 | 1.03–3.62 | 0.02 |
PJI: prosthetic joint infection.
Further analysis showed that being a male and suffering from a PJI had a fixed effects prevalence of 69.9% (95% CI, 60.5%–78.5%) and a random effects rate of 69% (95% CI, 58.4%–78.7%) (Fig. 4A). There was no heterogeneity (p = 0.32, I2 = 14%) and significant publication bias (p = 0.02) within the cohort (Fig. 4B). Prevalence of males who did not have a PJI was a fixed effects rate of 51.1% (95% CI, 47.4%–54.8%) compared to the random effects rate of 53.1% (95% CI, 43%–63%) (Fig. 5A). This cohort had substantial heterogeneity (p = 0.0001, I2 = 85%) without publication bias (p = 0.45) (Fig. 5B). Comparatively, male sex was associated with a higher risk of PJI (RR, 1.39; 95% CI, 1.19 to 1.6; p < 0.0001) (Table 4).
Fig. 4. (A) Fixed- and random-effects rates of male sex prosthetic joint infection (PJI). (B) Bias asssessment for male sex PJI.
Fig. 5. (A) Fixed- and random-effects rates of male sex controls. (B) Bias assessment for male sex controls.
Table 4. Male Sex and Association with PJI.
| Cohort | Fixed effects rate (%) | 95% CI (%) | Random effects rate (%) | 95% CI (%) | Heterogeneity p-value (I2, %) | Publication bias (p-value) | Relative risk | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| Male control | 51.1 | 47.4–54.8 | 53.1 | 43–63 | 0.0001 (85) | 0.45 | Reference | Reference | Reference |
| Male PJI | 69.9 | 60.5–78.5 | 69.0 | 58.4–78.7 | 0.32 (14) | 0.02 | 1.39 | 1.19–1.6 | 0.0001 |
PJI: prosthetic joint infection.
There was no association found between age using the fixed effects (standardized mean difference [SMD], –0.09 years; 95% CI, –0.31 to 0.124 years; p = 0.43) and random effects models (SMD, 0.18 years; 95% CI, –0.51 to 0.15 years; p = 0.28) (Fig. 6A). Further analysis showed no heterogeneity (p = 0.07, I2 = 49%) and some publication bias (p = 0.04) (Fig. 6B). Additionally, there was no association with body mass index (BMI) and PJI risk using both fixed and random effects models (SMD, 0.22 kg/m2; 95% CI, –0.01 to 0.45 kg/m2; p = 0.07) (Fig. 7A). However, testing showed no heterogeneity (p = 0.54, I2 = 0%) or publication bias (p = 0.06) (Fig. 7B).
Fig. 6. (A) Fixed- and random-effects rates of forest plot for age and prosthetic joint infection (PJI). (B) Bias assessment for age and PJI.
Fig. 7. (A) Fixed- and random-effects rates of forest plot for body mass index (BMI) and prosthetic joint infection (PJI). (B) Bias assessment for BMI and PJI.
When evaluating the role of patient comorbidities and PJI risk, we were able to assess for a history of diabetes and smoking. The fixed effects prevalence of diabetes in those who were diagnosed with a PJI was 32.4% (95% CI, 23.5%–41.9%) as opposed to the random effects rate of 29.6% (95% CI, 18.1%–42.5%) (Fig. 8A). The cohort showed no heterogeneity (p = 0.13, I2 = 39%) or publication bias (p = 0.32) (Fig. 8B). When further evaluating the prevalence of diabetes in those who did have a PJI, the fixed effects rate was 13.3% (95% CI, 10.9%–15.9%) compared to a random effects rate of 12.7% (95% CI, 6.3%–21%) (Fig. 9A). However, further cohort analysis showed both heterogeneity (p = 0.0001, I2 = 88%) and publication bias (p = 0.04) (Fig. 9B). Comparatively, a diagnosis of diabetes was associated with an increased risk of PJI (RR, 2.29; 95% CI, 1.63–3.22; p < 0.0001) (Table 5).
Fig. 8. (A) Fixed- and random-effects rates of forest plot for diabetes mellitus (DM) and prosthetic joint infection (PJI). (B) Bias assessment for DM PJI.
Fig. 9. (A) Fixed- and random-effects rates of forest plot for diabetes mellitus (DM) and controls. (B) Bias assessment for DM controls.
Table 5. DM and Risk of PJI.
| Cohort | Fixed effects rate (%) | 95% CI (%) | Random effects rate (%) | 95% CI (%) | Heterogeneity p-value (I2, %) | Publicationbias (p-value) | Relative risk | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| DM control | 13.3 | 10.9–15.9 | 12.7 | 6.3–21.0 | 0.0001 (88) | 0.04 | Reference | Reference | Reference |
| DM PJI | 32.4 | 23.5–41.9 | 29.6 | 18.1–42.5 | 0.13 (39) | 0.32 | 2.29 | 1.63–3.22 | 0.0001 |
DM: diabetes mellitus, PJI: prosthetic joint infection.
When assessing the role of smoking history, the prevalence of smoking history in those who were diagnosed with a PJI was a fixed and random effects rate of 20% (95% CI, 12.8%–28.5%) (Fig. 10A). This cohort showed no heterogeneity (p = 0.71, I2 = 0%) or publication bias (p = 0.46) (Fig. 10B). The prevalence of those with smoking history without a PJI was a fixed effects rate of 13.9% (95% CI, 11.5%–16.6%) as opposed to a random effects rate of 12.4% (95% CI, 7.2%–18.7%) (Fig. 11A). Further statistical analysis revealed substantial heterogeneity (p = 0.0001, I2 = 80%) without publication bias (p = 0.16) (Fig. 11B). Comparatively, there was no association between smoking and PJI risk (RR, 1.32; 95% CI, 0.84–2.07; p = 0.12) (Table 6).
Fig. 10. (A) Fixed- and random-effects rates of forest plot for smoking and prosthetic joint infection (PJI). (B) Bias assessment for smoking and PJI.
Fig. 11. (A) Fixed- and random-effects rates of forest plot for smoking controls. (B) Bias assessment for smoking controls.
Table 6. Tobacco Use and Risk of PJI.
| Cohort | Fixed effects rate (%) | 95% CI (%) | Random effects rate (%) | 95% CI (%) | Heterogeneity p-value (I2, %) | Publication bias (p-value) | Relative risk | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| Smoking control | 13.9 | 11.5–16.6 | 12.4 | 7.2–18.7 | 0.0001 (80) | 0.16 | Reference | Reference | Reference |
| Smoking PJI | 20.0 | 12.8–28.5 | 20.0 | 12.8–28.5 | 0.71 (0) | 0.46 | 1.32 | 0.84–2.07 | 0.12 |
PJI: prosthetic joint infection.
When evaluating 1- vs. 2-stage arthroplasties, we were able to utilize 11 studies (992 arthroplasties). Within the 1-stage cohort, the fixed and random effects PJI rate was 10.9% (95% CI, 8.0%–14.2%) (Fig. 12A). The cohort showed no heterogeneity (p = 0.46, I2 = 0%) or publication bias (p = 0.19) (Fig. 12B). Comparatively, the 2-stage cohort had a fixed effects PJI prevalence of 10.8% (95% CI, 8.3%–13.5%) and a random effects PJI prevalence of 10.3% (95% CI, 6.8%–14.4%) (Fig. 13A). Further analysis showed no significant heterogeneity (p = 0.06, I2 = 51%) or publication bias (p = 0.49) (Fig. 13B). There was no difference in PJI occurrence among those treated with 1-stage or 2-stage arthroplasty (RR, 1.01; 95% CI, 0.74–1.39; p = 0.47) (Table 7).
Fig. 12. (A) Fixed- and random-effects rates of forest plots for 1-stage prosthetic joint infection (PJI). (B) Bias assessment for 1-stage PJI.
Fig. 13. (A) Fixed- and random-effects rates of forest plots for 2-stage prosthetic joint infection (PJI). (B) Bias assessment for 2-stage PJI.
Table 7. Studies Showing 1- versus 2-Stage Arthroplasty Failures.
| Cohort | Fixed effects failure rate (%) | 95% CI (%) | Random effects failure rate (%) | 95% CI (%) | Heterogeneity p-value (I2, %) | Publication bias (p-value) | Relative risk | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| One stage | 10.9 | 8.0–14.2 | 10.9 | 8.0–14.2 | 0.46 (0) | 0.19 | Reference | Reference | Reference |
| Two stage | 10.8 | 8.3–13.5 | 10.3 | 6.8–14.4 | 0.06 (51) | 0.49 | 1.01 | 0.74–1.39 | 0.47 |
DISCUSSION
Primary THA and TKA in those with a history of septic arthritis remains a challenging dilemma for practitioners. Currently, there are no clear sets of guidelines regarding when it is appropriate to perform a primary arthroplasty in this group of patients. Our systematic review and meta-analysis aimed to determine the true incidence of PJI in those with a history of septic arthritis and attempt to elucidate risk factors for said PJI. In addition to a PJI prevalence of 7.5%, we found that male sex and an etiology of unspecified resistant organisms in their original native joint at the time of septic arthritis diagnosis conferred an increased risk.
Our study has several limitations. The studies included were small and often included single institutions, which could lead to selection bias. Although there were 18 studies included in this systematic review, not all of them could be used for comparison for failure and success. This review cannot definitively account for all the potential variations in surgical protocol among the studies. Furthermore, although we cannot account for all potential confounders with heterogeneity and bias, the variables this was found in were within our control group that did not suffer from a PJI. Additionally, this study did not have enough cases to be able to perform various subgroup analyses based on various patient and surgery specific factors. Despite these limitations, we believe this systematic review provides insights into the risk of PJI following arthroplasty in those with a previous history of septic arthritis.
When exploring the source of septic arthritis, many cases had negative cultures. However, when cultures were positive, the most common etiology was multiple variations of the Staphylococcus species. Bettencourt et al.6) evaluated the outcomes of 215 primary TKAs with a history of previous septic arthritis. Among their cohort, Staphylococcus species accounted for 100 of the cases of septic arthritis (46.5%). Additionally, 65 cases (30%) had unknown etiology. Similarly, Russo et al.5) evaluated the outcomes of primary arthroplasty following a history of septic arthritis (n = 114 arthroplasties). Staphylococcus species were by far the most commonly isolated infection source (n = 54, 47.4%) with approximately 20% being culture negative (n = 23). Hence, although Staphylococcus remains the most commonly identified etiology for septic arthritis, there are a number of cases that will have negative cultures.
When exploring the risk factors for PJI in patients undergoing primary arthroplasty with a history of septic arthritis, there was an association found with male sex. Association with male sex has been affirmed by several small studies. Recently, Tan et al.8) assessed the outcomes of THA and TKA following septic arthritis at multiple institutions (n = 207 arthroplasties). After a mean follow-up of 4.8 years (SD ± 3.1 years), there were 25 patients who obtained a PJI, 19 of whom were males, conferring male sex as a risk factor (76% vs. 46%, p = 0.025). Similarly, Wei et al.13) evaluated the success of primary arthroplasty in those with septic arthritis history (n = 105). After a mean follow-up of 10.3 years (range, 1.6–21.3 years), 14 of the 17 PJI cases were male. After using an adjusted model, this conferred a substantial association with male sex and PJI in those with a history of septic arthritis (hazard ratio [HR], 9.95; 95% CI, 2.08–47.53; p < 0.01). Hence, male sex in those with previously treated septic arthritis may be considered a risk factor for primary arthroplasty failure.
Conversely, other small studies show no association between sex and PJI following arthroplasty. Xu et al.14) evaluated 74 patients who underwent 2-stage arthroplasties for septic arthritis at a single institution. After a mean follow-up of 4.7 years, 5 of the 9 failures were male. However, male sex was not found to be a risk for PJI (HR, 0.68; 95% CI, 0.17–2.8; p = 0.598). Similarly, Seo et al.11) assessed the outcomes of primary TKA in those with a previous septic arthritis history (n = 62 knees). After a mean follow-up of 6 years (range, 2–10.4 years), there were 7 PJIs, of which 2 were male and was not associated with arthroplasty failure (29% vs. 25%, p = 0.52). However, although there are studies that disagree with our findings of male sex association with arthroplasty failure in septic arthritis patients, this should be viewed cautiously given their smaller size and shorter follow-up time.
This meta-analysis determined that Gram-negative source of native septic joint conferred an increased risk of PJI. However, several studies have been unable to show an association with Gram-negative septic arthritis. Tan et al.16) evaluated the risk factors for PJI following 233 primary arthroplasties in those with a history of septic arthritis. After a mean 5-year follow-up, there was no association between Gram-negative bacterial septic arthritis and PJI following arthroplasty (HR, 1.19; 95% CI, 0.35–4.04; p = 0.78). Similarly, the previously mentioned study carried out by Xu et al.14) following 2-stage arthroplasty showed no association with Gram-negative bacteria (HR, 1.17; 95% CI, 0.15–9.40; p = 0.88). However, there have been other sources of native septic arthritis that have been associated with PJI. In the previously cited study performed by Wei et al.,13) they found a strong association between polymicrobial septic arthritis and PJI following arthroplasty (HR, 10.02; 95% CI, 1.48–68.06; p = 0.02). The study evaluating 2-stage arthroplasty performed by Xu et al.14) found a strong association with resistant organisms (HR, 13.96; 95% CI, 3.29–19.20; p = 0.001). It is possible this review was unable to determine true association with septic arthritis source due to a lack of raw data in studies for comparison. Hence, the association between causative agent of septic arthritis and PJI following primary arthroplasty remains an area of continued investigation.
This study found no association between the risk of arthroplasty failure and age or BMI. In the study performed by Wei et al.13) on 105 arthroplasties, they found no association between increasing age per year (HR, 0.98; 95% CI, 0.95–1.02; p = 0.34) and increasing BMI per kg/m2 (HR, 1.08; 95% CI, 0.95–1.22; p = 0.23). Similarly, Tan et al.8) found no association between mean age (56.7 years vs. 55.3 years, p = 0.84) or mean BMI (27.7 vs. 26.9 kg/m2, p = 0.28) with arthroplasty failure. However, although several studies did not show an association with age or BMI and PJI following arthroplasty in a history of septic arthritis, these are 2 factors that will continue to be investigated.
Additionally, diabetes history was associated with an increased risk of primary arthroplasty failure while smoking history had no association. These findings have studies which support and challenge this finding. Tan et al.16) evaluated the role of comorbidities in treatment failure and found a strong association with diabetes (HR, 4.09; 95% CI, 1.64–10.19; p = 0.003). Additionally, similar to our study, they found no association between PJI and smoking history (HR, 1.61; 95% CI, 0.64–4.05; p = 0.32). Conversely, Xu et al.14) failed to find an association with diabetes mellitus (HR, 2.16; 95% CI, 0.45–10.39; p = 0.34) or smoking (HR, 1.77; 95% CI, 0.37–8.52; p = 0.48) in their study of 74 2-stage arthroplasties. Given this study is smaller in size, the results should be interpreted cautiously. Hence, diabetes may still be considered a risk factor for surgical treatment failure in this setting.
One of the more intriguing findings of this systematic review was the fact there was no difference between single-stage or 2-stage treatment in patients with a history of septic arthritis. There were 2 studies that offered a comparison. The choice certainly varies between surgeons as advocates of 1-stage treatment will say it can reduce the time a patient goes without a fully functioning knee.26) Conversely, supporters of using 2-stage revision cite it as the gold standard of treatment and superior efficacy to 1-stage alternatives.26) The study from Tan et al.8) was the only study within this review that directly compared the efficacy of 1-stage (n = 97) or 2-stage (n = 110) arthroplasty. There was no difference in PJI risk between 1- or 2-stage treatment (10.3% vs. 13.6%, p = 0.67). Additionally, another study carried out by Tan et al.16) among 233 primary arthroplasties comparing these 2 treatment modalities showed no difference in efficacy (HR, 1.72; 95% CI, 0.58–5.06; p = 0.33). Given the fact this result is confirmed by 2 small studies from the same author, this remains an area of further investigation.
In conclusion, primary THA and TKA in those with a history of septic arthritis can be both safe and efficacious. However, there are certain risk factors such as male sex, history of Gram-negative septic joint, and a diagnosis of diabetes that could lead to a worse outcome. Practitioners should be aware of these risks so that modifiable risk factors can be addressed accordingly. For unmodifiable risk factors, patients should be educated regarding their risk of poorer outcomes. Future studies should be larger and focus on long-term outcomes in these patients while finding ways to maximize the safety and efficacy of primary arthroplasty, particularly in patients who may have a higher risk of PJI following a previous diagnosis of native septic arthritis.
Footnotes
CONFLICT OF INTEREST: No potential conflict of interest relevant to this article was reported.
References
- 1.Matthews PC, Dean BJ, Medagoda K, et al. Native hip joint septic arthritis in 20 adults: delayed presentation beyond three weeks predicts need for excision arthroplasty. J Infect. 2008;57(3):185–190. doi: 10.1016/j.jinf.2008.07.001. [DOI] [PubMed] [Google Scholar]
- 2.Li W, Fang X, Zhang C, et al. Comparison of efficacy and complications between two types of staging arthroplasty in treating chronic septic hip arthritis: a retrospective clinical study. Exp Ther Med. 2019;17(5):4123–4131. doi: 10.3892/etm.2019.7430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Papanna MC, Chebbout R, Buckley S, Stockley I, Hamer A. Infection and failure rates following total hip arthroplasty for septic arthritis: a case-controlled study. Hip Int. 2018;28(1):63–67. doi: 10.5301/hipint.5000538. [DOI] [PubMed] [Google Scholar]
- 4.Elsissy JG, Liu JN, Wilton PJ, Nwachuku I, Gowd AK, Amin NH. Bacterial septic arthritis of the adult native knee joint: a review. JBJS Rev. 2020;8(1):e0059. doi: 10.2106/JBJS.RVW.19.00059. [DOI] [PubMed] [Google Scholar]
- 5.Russo A, Migliorini F, Giustra F, Bosco F, Massè A, Burastero G. Two-stage total joint replacement for hip or knee septic arthritis: post-traumatic etiology and difficult-to-treat infections predict poor outcomes. Arch Orthop Trauma Surg. 2024;144(12):5111–5119. doi: 10.1007/s00402-024-05249-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bettencourt JW, Wyles CC, Fruth KM, et al. Outcomes of primary total knee arthroplasty following septic arthritis of the native knee: a case-control study. J Bone Joint Surg Am. 2021;103(18):1685–1693. doi: 10.2106/JBJS.20.01678. [DOI] [PubMed] [Google Scholar]
- 7.Aboltins C, Daffy J, Choong P, Stanley P. Current concepts in the management of prosthetic joint infection. Intern Med J. 2014;44(9):834–840. doi: 10.1111/imj.12510. [DOI] [PubMed] [Google Scholar]
- 8.Tan TL, Xu C, Kuo FC, et al. When total joint arthroplasty after septic arthritis can be safely performed. JB JS Open Access. 2021;6(2):e20.00146. doi: 10.2106/JBJS.OA.20.00146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sultan AA, Mahmood B, Samuel LT, et al. Patients with a history of treated septic arthritis are at high risk of periprosthetic joint infection after total joint arthroplasty. Clin Orthop Relat Res. 2019;477(7):1605–1612. doi: 10.1097/CORR.0000000000000688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ni M, Fu J, Deng T, et al. Clinical effects of staged joint replacement in patients with septic arthritic knee. J Orthop Surg Res. 2020;15(1):525. doi: 10.1186/s13018-020-02062-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Seo JG, Moon YW, Park SH, Han KY, Kim SM. Primary total knee arthroplasty in infection sequelae about the native knee. J Arthroplasty. 2014;29(12):2271–2275. doi: 10.1016/j.arth.2014.01.013. [DOI] [PubMed] [Google Scholar]
- 12.Ohlmeier M, Delgado G, Calderon CA, Hartwig CH, Gehrke T, Citak M. Are patients with a history of septic arthritis undergoing total knee arthroplasty at higher risk for revision surgery? A single-center study. J Arthroplasty. 2020;35(7):1857–1861. doi: 10.1016/j.arth.2020.02.065. [DOI] [PubMed] [Google Scholar]
- 13.Wei JT, Kuo FC, Wang JW, Ko JY, Lee MS, Wu CT. Outcome and predictors of septic failure following total joint arthroplasty for prior septic arthritis of hip and knee joint. J Arthroplasty. 2022;37(7):1375–1382. doi: 10.1016/j.arth.2022.03.011. [DOI] [PubMed] [Google Scholar]
- 14.Xu C, Kuo FC, Kheir M, Li X, Chai W, Chen JY. Outcomes and predictors of treatment failure following two-stage total joint arthroplasty with articulating spacers for evolutive septic arthritis. BMC Musculoskelet Disord. 2019;20(1):272. doi: 10.1186/s12891-019-2652-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Russo A, Cavagnaro L, Chiarlone F, Clemente A, Romagnoli S, Burastero G. Clinical outcomes and survivorship of two-stage total hip or knee arthroplasty in septic arthritis: a retrospective analysis with a minimum five-year follow-up. Int Orthop. 2021;45(7):1683–1691. doi: 10.1007/s00264-021-05013-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tan T, Xu C, Kuo FC, Ghanem E, Higuera C, Parvizi J. Risk factors for failure and optimal treatment of total joint arthroplasty for septic arthritis. J Arthroplasty. 2021;36(3):892–896. doi: 10.1016/j.arth.2020.09.020. [DOI] [PubMed] [Google Scholar]
- 17.Bettencourt JW, Wyles CC, Osmon DR, Hanssen AD, Berry DJ, Abdel MP. Outcomes of primary total hip arthroplasty following septic arthritis of the hip : a case-control study. Bone Joint J. 2022;104(2):227–234. doi: 10.1302/0301-620X.104B2.BJJ-2021-1209.R1. [DOI] [PubMed] [Google Scholar]
- 18.Zeng WN, Liu JL, Jia XL, Zhou Q, Yang L, Zhang Y. Midterm results of total hip arthroplasty in patients with high hip dislocation after suppurative hip arthritis. J Arthroplasty. 2019;34(1):102–107. doi: 10.1016/j.arth.2018.09.081. [DOI] [PubMed] [Google Scholar]
- 19.Wang D, Zeng WN, Qin YZ, Pei FX, Wang HY, Zhou ZK. Long-term results of cementless total hip arthroplasty for patients with high hip dislocation after childhood pyogenic infection. J Arthroplasty. 2019;34(10):2420–2426. doi: 10.1016/j.arth.2019.05.040. [DOI] [PubMed] [Google Scholar]
- 20.Yang Y, Yu QP, Wang SL, et al. Outcomes after total hip arthroplasty using a cementless S-ROM modular stem for patients with high hip dislocation secondary to hip pyogenic arthritis. Orthop Surg. 2019;11(3):460–466. doi: 10.1111/os.12485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Park CW, Lim SJ, Cha YT, Park YS. Total hip arthroplasty with subtrochanteric shortening osteotomy in patients with high hip dislocation secondary to childhood septic arthritis: a matched comparative study with Crowe IV developmental dysplasia. J Arthroplasty. 2020;35(1):204–211. doi: 10.1016/j.arth.2019.08.034. [DOI] [PubMed] [Google Scholar]
- 22.Luo Y, Yang Z, Yeersheng R, Li D, Kang P. Clinical outcomes and quality of life after total hip arthroplasty in adult patients with a history of infection of the hip in childhood: a mid-term follow-up study. J Orthop Surg Res. 2019;14(1):38. doi: 10.1186/s13018-019-1074-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Portier E, Zeller V, Kerroumi Y, Heym B, Marmor S, Chazerain P. Arthroplasty after septic arthritis of the native hip and knee: retrospective analysis of 49 joints. J Bone Jt Infect. 2022;7(2):81–90. doi: 10.5194/jbji-7-81-2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kunze KN, Sadauskas AJ, Kernzer B, Levine BR. Two-stage primary arthroplasty of native hips and knees that had previously failed treatment for septic arthritis: a single-center experience. Arthroplast Today. 2020;6(3):431–436. doi: 10.1016/j.artd.2020.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Shaikh AA, Ha CW, Park YG, Park YB. Two-stage approach to primary TKA in infected arthritic knees using intraoperatively molded articulating cement spacers. Clin Orthop Relat Res. 2014;472(7):2201–2207. doi: 10.1007/s11999-014-3545-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Patel D, Shannon V, Sharma S, Liu J, Skie M. A meta-analysis of success rates of one-stage versus two-stage revisions in knee prosthetic joint infections. Cureus. 2024;16(4):e57533. doi: 10.7759/cureus.57533. [DOI] [PMC free article] [PubMed] [Google Scholar]