Periprosthetic joint infection (PJI) is an uncommon but severe complication of total knee arthroplasty (TKA). The existing literature indicates that approximately 1% to 2% of patients undergoing primary arthroplasty experience PJI; moreover, PJI is the leading cause of revision arthroplasty.1 2 Individuals with PJI may experience a substantial decrease in quality of life and must undergo complex and costly treatments to resolve this complication.3

A previous study at our institution, Queen Mary Hospital in Hong Kong, examined 2543 patients who underwent elective primary TKA between 1993 and 2013.4 During that period, the reported incidence of PJI was 1.34% and the most common causative organism was methicillin-sensitive Staphylococcus aureus (MSSA).4 The number of TKAs performed at our centre is rapidly increasing. In the past 8 years, clinicians at our institution have performed 85% of the total number of TKAs between 1993 and 2013, a span of 20 years. Considering the rapid population ageing in Hong Kong, we anticipate a continued increase in the number of TKAs. In recent years, various measures have been proposed to further reduce the incidence of PJI, including restrictions on blood transfusion rates, preoperative optimisation of modifiable risk factors, and the implementation of stringent culture techniques to improve microbial yield.5 6 7 However, the limited availability of local data makes it difficult to assess the effectiveness of these techniques in reducing PJI incidence. It is important to analyse the efficacy of these interventions as part of ongoing efforts to improve surgical outcomes at our centre. Furthermore, the increasing consumption of antibiotics over the past two decades has led to inevitable changes in the microbiological landscape of infectious organisms.8

This study had three objectives. First, it aimed to provide current local data on the incidence of PJI after elective primary TKA. Second, it sought to identify changes in the microbiological landscape of PJI; this information may guide future treatment and prevention strategies. Third, it aimed to showcase the efficacy of interventions to reduce PJI incidence and encourage their adoption beyond our institution.

Considering the measures introduced to reduce infection at our institution, we hypothesised that the incidence of PJI after primary TKA decreased over the past decade. We also hypothesised that the proportion of methicillin-resistant S aureus (MRSA)–related PJI increased during this period due to the increasing global consumption of antibiotics.

This retrospective cohort study compared the incidence and bacteriology of PJI after TKA at our institution. Participants were included if they underwent primary elective TKA at our institution between 2014 and 2021, and met the 2011 Musculoskeletal Infection Society (MSIS) criteria for PJI.9 Exclusion criteria were infection after revision arthroplasty, knee arthroplasty for malignant joint conditions, and active bacteraemia. The primary outcomes of interest were the incidence and bacteriological patterns of PJI. Secondary outcomes included preoperative patient demographics and time to onset of PJI.

The Hong Kong Hospital Authority’s Clinical Data Analysis and Reporting System and the Local Joint Replacement Registry were utilised to identify all TKAs performed at our institution between 2014 and 2021. Records were then searched using the keywords ‘orthopaedic aftercare’ and ‘periprosthetic joint infection’ to identify potential cases of PJI. Patients who did not meet the 2011 MSIS criteria for PJI were excluded; the remaining patients comprised the study cohort.10 Two senior authors of this study (PK Chan and KY Chiu) independently screened the patient database using the 2011 MSIS criteria to identify suitable patients for further data collection. Any uncertainties or disagreements were resolved through discussion.

Using a predefined data extraction form, the same two senior authors extracted the following data from the records of all included patients: intraoperative joint fluid culture results, age, sex, medical co-morbidities (eg, diabetes mellitus, rheumatoid arthritis, and immunosuppression), date of the index operation, surgical technique, operative time, date of re-operation, and postoperative antibiotic regimen.

A previous retrospective cohort study by Siu et al4 assessed the incidence and bacteriology of PJI among patients who underwent TKA at our institution between 1993 and 2013. From that study, we extracted data on the incidence and bacteriology of PJI, patient demographics, and time to onset of infection for comparison with our cohort. For both cohorts, we recorded the mean operative time, number of joint specialists involved, surgical technique, use of patient-specific instrumentation, and infection control protocols; these data were used to identify potential confounders that could influence the incidence of PJI.

Patients in our cohort were classified according to the time to infection onset as early, delayed, or late. Early-onset PJI was defined as infection occurring within 3 months of the index operation. These infections commonly arise from intraoperative contamination by highly virulent microorganisms and therefore constitute a key focus of intervention. Delayed-onset PJI was defined as infection occurring between 3 and 24 months after the index operation. These infections are also typically acquired during surgery but involve less virulent microorganisms. Late-onset PJI was defined as infection occurring over 24 months after surgery. These infections are often caused by haematogenous pathogens unrelated to the index operation.11

In accordance with our institution’s guidelines, patients were invited to attend follow-up at 2 weeks, 3 months, 6 months, and 12 months postoperatively. Patients without complications were subsequently scheduled for annual follow-up. Regarding infection control, the preoperative, perioperative, and postoperative protocols for elective primary TKA remained consistent throughout the study period in both cohorts. Intravenous antibiotic prophylaxis (1 g of cefazolin, or vancomycin for patients with a penicillin allergy) was administered 1 hour prior to skin incision. Intraoperatively, laminar airflow and body exhaust systems were utilised. Antibiotic-loaded cement was not routinely used, and a single postoperative wound management and rehabilitation programme was implemented throughout the study period. Postoperative antibiotics were not routinely administered.

Categorical variables were grouped for analysis; prevalence was calculated and group differences were tested with the Chi squared test. Continuous variables were compared using independent two-tailed t tests. A P value <0.05 was considered statistically significant. All statistical analyses were conducted using SPSS (Windows version 27.0; IBM Corp, Armonk [NY], United States).

In total, 2543 and 2171 primary TKAs were performed at our institution between 1993-20134 and 2014-2021, respectively. The incidence of PJI was 0.64% (n=14; 95% confidence interval=0.39-0.89) between 2014 and 2021, significantly lower than the 1.34% (n=34; 95% confidence interval=0.97-1.71) recorded between 1993 and 2013 (P=0.018).4

The mean age of the 14 patients with PJI in our cohort was 68.5 ± 7 years (range, 56-85). Of these patients, eight were men (57.1%) and six were women (42.9%). In terms of medical co-morbidities, seven patients had diabetes mellitus (50.0%), one had rheumatoid arthritis (7.1%), and one had end-stage renal disease requiring immunosuppression (7.1%). The mean follow-up period in our cohort was 4 years 9 months (interquartile range [IQR], 4 years 0 months to 6 years 11 months). There were no significant differences in age, sex distribution, or medical co-morbidities (diabetes mellitus and rheumatoid arthritis) between the two cohorts. The cohort demographics are compared in Table 1.4

We analysed other potential confounding factors (eg, mean operative time, number of joint specialists involved, and surgical approach) to minimise their effects on the primary and secondary outcomes.12 The indications for TKA did not change at our institution during the two time periods. Our institutional guidelines state that patients with Kellgren and Lawrence Grade 3 or 4 end-stage knee osteoarthritis and debilitating symptoms refractory to nonoperative treatment are candidates for TKA. The mean operative times for primary elective TKA were 1 hour 56 minutes during 1993-20134 and 1 hour 33 minutes during 2014-2021. The difference was not statistically significant (P=0.170). The number of joint specialists involved increased from four to six between the two periods. From 1993 to 2019, all TKAs were exclusively performed using the conventional approach.

After the computed tomography–based robotic arm–assisted system for total joint arthroplasty was introduced in 2019,13 surgeons at our institution could choose between robotic-assisted TKA and conventional TKA. Currently, there are no specific indications for either approach; the choice remains a matter of surgeon preference.14 To our knowledge, no studies have compared PJI incidence between robotic-assisted and conventional TKA; future research should explore the infection rate associated with each procedure. Patients undergoing robotic-assisted TKA at our institution follow the same postoperative protocol established for those undergoing conventional TKA (follow-up at 2 weeks, 3 months, 6 months, 12 months, and annually thereafter). Because robotic-assisted TKA was recently introduced at our institution, the mean follow-up duration for patients treated with this approach was short (14 months; IQR, 4.5-26).

Early-onset PJI occurred in 7.1% (n=1) of patients in the study cohort, arising 60 days after arthroplasty. In the 1993-2013 cohort, 29.4% (n=10) of patients experienced early-onset infection at a median of 17 days after arthroplasty (IQR, 9-32).4 However, the incidence of early-onset PJI did not differ significantly between the two cohorts (P=0.095) [Table 1]. Delayed-onset PJI occurred in 28.6% (n=4) of patients in the study cohort, occurring at a median of 6 months after arthroplasty (IQR, 5-7). Late-onset PJI occurred at a median of 3 years after arthroplasty (IQR, 2 years 1 month to 3 years 7 months). A larger proportion of patients experienced infection during the first year after surgery in the 1993-2013 cohort4 compared with the 2014-2021 cohort (59% vs 36%).

Methicillin-sensitive S aureus remained the most common causative organism in cases of PJI between 2014 and 2021, affecting 57.1% (n=8) of patients. The proportion of PJI cases caused by MSSA was significantly greater in the 2014-2021 cohort than in the 1993-2013 cohort, in which 26.5% (n=9) of patients were infected with MSSA (P=0.043) [Table 2].4

Methicillin-resistant S aureus was the second most common causative organism in cases of PJI between 1993 and 2013 (17.6%, n=6)4; Streptococcus spp. (14.3%, n=2) was the second most common causative organism between 2014 and 2021. The two cases of streptococcal infection in the 2014-2021 cohort comprised one with Streptococcus dysgalactiae and one with Streptococcus agalactiae. Other causative organisms in PJI cases within the 2014-2021 cohort included MRSA (7.1%, n=1), methicillin-sensitive coagulase-negative staphylococci (7.1%, n=1), and Escherichia coli (7.1%, n=1). Methicillin-resistant strains accounted for 40% of all staphylococcal infections between 1993 and 20134; this proportion was 11.1% between 2014 and 2021. Table 2 compares the microbiological patterns of PJI between the two cohorts.

There was a non-significant decrease in the proportion of patients with culture-negative PJI between the two cohorts, from 23.5% (n=8) between 1993 and 20134 to 7.1% (n=1) between 2014 and 2021 (P=0.186) [Table 2].

This study showed that the incidence of PJI after primary TKA at our institution significantly decreased. Worldwide, the reported incidence of PJI after primary elective TKA ranges from 1% to 2%.1 2 Over the years, our institution has implemented various measures to reduce the incidence of PJI after TKA, including a preoperative patient optimisation programme and a restrictive blood management programme. These measures are summarised in Table 3.5 6

The association between blood transfusion and increased perioperative morbidity in patients undergoing TKA is well documented.15 16 The American College of Surgeons National Surgical Quality Improvement Program reported that patients receiving transfusions experienced up to a tenfold increase in the risk of adverse postoperative outcomes.17 Based on these findings, a more restrictive transfusion approach has been implemented in the past several years after the 2015 study17 to improve postoperative outcomes. A retrospective study of 12 590 patients demonstrated significant decreases in complications, 30-day readmissions, and hospital length of stay following implementation of a patient blood management programme for patients undergoing prosthetic joint arthroplasty.18 The programme aimed to reduce transfusion requirements by optimising red cell mass, minimising blood loss, and defining appropriate indications for transfusion.18 A patient blood management programme was introduced at our institution in 2014; subsequently, the mean transfusion rate among patients undergoing TKA decreased from 31.3% in 2013 to 1.9% in 2018.6

The preoperative optimisation programme at our institution emphasises the optimisation of modifiable risk factors for PJI prior to TKA.5 Rheumatological diseases such as rheumatoid arthritis, juvenile inflammatory arthritis, ankylosing spondylitis, and psoriatic arthritis are known to increase the risk of PJI.19 20 A previous review of 2543 TKAs showed that the incidence of PJI was 3.1% in patients with rheumatoid arthritis, significantly higher than the 1.2% observed in patients without rheumatoid arthritis.4 At our institution, patients with rheumatoid arthritis who exhibit a persistently elevated erythrocyte sedimentation rate or C-reactive protein level are referred to a rheumatologist for further assessment and treatment prior to surgery. Diabetes mellitus is also strongly associated with an increased risk of PJI. All patients scheduled for elective TKA at our institution undergo universal glycated haemoglobin and fructosamine screening, with referral to an endocrinologist for optimisation if the glycated haemoglobin level exceeds 7.5%.21 Other modifiable risk factors monitored at our institution include weight control, vitamin D status, and nutritional status.22

From 2014 to 2021, MSSA was the most common causative organism in cases of PJI at our institution (57.1%, n=8). This finding is consistent with the existing literature, which indicates that S aureus is the most common causative organism in PJI after primary joint arthroplasty (19%-29% of cases worldwide).23 24 25 The unique virulence factors of S aureus enhance its ability to adhere to implants, facilitating aggressive biofilm formation and enabling replication and survival within this microenvironment.26 27

Owing to the increasing use of antibiotics in the community, we hypothesised that the number of antibiotic-resistant causative organisms would increase in our cohort of patients with PJI. The SENTRY Antimicrobial Surveillance Program evaluated 20-year trends in antimicrobial susceptibility among S aureus isolates across 427 centres in 45 countries.28 The authors reported that the prevalence of MRSA peaked at 44.2% in 2005-2008, then declined to 42.3% in 2009-2012 and 39.0% in 2013-2016.28 The incidence of MRSA among PJI cases in our study was consistent with findings from other regions. For example, a multicentre study in New Zealand showed that 9.1% of PJIs were attributable to MRSA.29

It is well established that early-onset infection typically occurs during surgery through intraoperative contamination, whereas late-onset PJI commonly arises from haematogenous spread.10 We observed a decrease in the proportion of early-onset infection between the two cohorts; however, this difference was not statistically significant (P=0.095). Additional measures should be implemented to further reduce the incidence of PJI after primary TKA at our institution.

In recent years, the incidence of culture-negative PJI has increased among patients undergoing total joint arthroplasty.30 This increase has been hypothesised to result from a higher prevalence of low-virulence organisms, premature antibiotic treatment, and failure to use enriched culture media.31 32 An inability to identify causative organisms in cases of PJI represents a serious problem for surgeons and infection control teams because of the uncertainties associated with antimicrobial selection. To reduce the incidence of culture-negative PJI, our institution implemented recommendations published by Tan et al,7 including extending the incubation period, using blood culture bottles and flasks, and collecting an adequate number of separate intraoperative tissue samples from patients with suspected PJI.7 The incidence of culture-negative PJI at our institution declined; however, this decline was not statistically significant (P=0.186).

This study had several limitations. First, it included patients treated at a single academic centre in Hong Kong; therefore, the findings may not accurately reflect changing trends in PJI incidence and bacteriology across the region. Further multicentre studies are warranted to better understand these trends. Second, the duration of patient recruitment differed between the present and former studies (7 years vs 20 years). Third, the number of patients varied between the two cohorts (2543 vs 2171). Despite this difference, we proceeded with the 7-year recruitment period considering the clinical value of evaluating recent PJI incidence and bacteriology in Hong Kong. Considerable effort was made to standardise patient characteristics and baseline co-morbidities to ensure comparability between cohorts. Fourth, given the relatively short mean follow-up duration (4 years 9 months), some cases of late-onset PJI may have occurred after completion of follow-up. Fifth, inconsistencies in record-keeping over the past decade prevented analysis of all documented risk factors for PJI; this limitation was unavoidable because of the retrospective study design. Sixth, the limited number of PJI cases hindered further subgroup analyses (ie, assessment and comparison of PJI incidence between conventional and robotic-assisted approaches). A similar study with a larger cohort may therefore be beneficial. Despite these limitations, the present study represents the largest series of PJI cases in Hong Kong to compare bacteriological patterns across two time periods. We believe that the findings have important clinical implications for PJI management in local hospitals.

This is the first study in Hong Kong to assess changes in the incidence and microbiological patterns of PJI after TKA across two time periods. Our findings have substantial clinical implications, as they demonstrate the effectiveness of interventional measures implemented at our institution in reducing the incidence of PJI, the rate of culture-negative PJI, and the number of early-onset cases. Prevention of PJI improves patient outcomes and reduces the economic burden on the healthcare system. Larger, multicentre, prospective studies are required to further elucidate bacteriological trends in PJI in Hong Kong.

Concept and design: All authors.
Acquisition of data: JR Khoo.
Analysis or interpretation of data: JR Khoo, PK Chan.
Drafting of the manuscript: JR Khoo.
Critical revision the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This manuscript was presented as an oral presentation at the 42nd Annual Congress of the Hong Kong Orthopaedic Association (5-6 November 2022, Hong Kong).

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

This research was approved by the Institutional Review Board of The University of Hong Kong/Hospital Authority Hong Kong West Cluster, Hong Kong (Ref No.: UW 25-585). The requirement for informed patient consent was waived by the Committee due to the retrospective nature of the research.

1. Pulido L, Ghanem E, Joshi A, Purtill JJ, Parvizi J. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res 2008;466:1710-5. Crossref

2. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med 2004;351:1645-54. Crossref

3. Premkumar A, Kolin DA, Farley KX, et al. Projected economic burden of periprosthetic joint infection of the hip and knee in the United States. J Arthroplasty 2021;36:148-49.e3. Crossref

4. Siu KT, Ng FY, Chan PK, Fu H, Yan CH, Chiu KY. Bacteriology and risk factors associated with periprosthetic joint infection after primary total knee arthroplasty: retrospective study of 2543 cases. Hong Kong Med J 2018;24:152-7. Crossref

5. Chan VW, Chan PK, Fu H, et al. Preoperative optimization to prevent periprosthetic joint infection in at-risk patients. J Orthop Surg (Hong Kong) 2020;28:2309499020947207. Crossref

6. Chan PK, Hwang YY, Cheung A, et al. Blood transfusions in total knee arthroplasty: a retrospective analysis of a multimodal patient blood management programme. Hong Kong Med J 2020;26:201-7. Crossref

7. Tan TL, Kheir MM, Shohat N, et al. Culture-negative periprosthetic joint infection: an update on what to expect. JB JS Open Access 2018;3:e0060. Crossref

8. Roberts SC, Zembower TR. Global increases in antibiotic consumption: a concerning trend for WHO targets. Lancet Infect Dis 2021;21:10-1. Crossref

9. Workgroup Convened by the Musculoskeletal Infection Society. New definition for periprosthetic joint infection. J Arthroplasty 2011;26:1136-8. Crossref

10. Parvizi J, Zmistowski B, Berbari EF, et al. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res 2011;469:2992-4. Crossref

11. Tande AJ, Patel R. Prosthetic joint infection. Clin Microbiol Rev 2014;27:302-45. Crossref

12. Wang Q, Goswami K, Shohat N, Aalirezaie A, Manrique J, Parvizi J. Longer operative time results in a higher rate of subsequent periprosthetic joint infection in patients undergoing primary joint arthroplasty. J Arthroplasty 2019;34:947-53. Crossref

13. The University of Hong Kong. HKUMed introduces the latest robotic arm assisted joint replacement technology for enhancing surgical precision [press release]. 2020 May 28. Available from: https://www.hku.hk/press/press-releases/detail/21111.html#:~:text=Robotic%20Arm%20Assisted%20Joint%20Replacement%20Surgery,enable%20accurate%20patient%2Dspecific%20planning. Accessed 1 Apr 2026.

14. Deckey DG, Rosenow CS, Verhey JT, et al. Robotic-assisted total knee arthroplasty improves accuracy and precision compared to conventional techniques. Bone Joint J 2021;103-B(6 Suppl A):74-80. Crossref

15. Everhart JS, Sojka JH, Mayerson JL, Glassman AH, Scharschmidt TJ. Perioperative allogeneic red blood-cell transfusion associated with surgical site infection after total hip and knee arthroplasty. J Bone Joint Surg Am 2018;100:288-94. Crossref

16. Lu Q, Peng H, Zhou GJ, Yin D. Perioperative blood management strategies for total knee arthroplasty. Orthop Surg 2018;10:8-16. Crossref

17. Ferraris VA, Hochstetler M, Martin JT, Mahan A, Saha SP. Blood transfusion and adverse surgical outcomes: the good and the bad. Surgery 2015;158:608-17. Crossref

18. Loftus TJ, Spratling L, Stone BA, Xiao L, Jacofsky DJ. A patient blood management program in prosthetic joint arthroplasty decreases blood use and improves outcomes. J Arthroplasty 2016;31:11-4. Crossref

19. Kunutsor SK, Whitehouse MR, Blom AW, Beswick AD; INFORM Team. Patient-related risk factors for periprosthetic joint infection after total joint arthroplasty: a systematic review and meta-analysis. PLoS One 2016;11:e0150866. Crossref

20. Morrison TA, Figgie M, Miller AO, Goodman SM. Periprosthetic joint infection in patients with inflammatory joint disease: a review of risk factors and current approaches to diagnosis and management. HSS J 2013;9:183-94. Crossref

21. Duensing I, Anderson MB, Meeks HD, Curtin K, Gililland JM. Patients with type-1 diabetes are at greater risk of periprosthetic joint infection: a population-based, retrospective, cohort study. J Bone Joint Surg Am 2019;101:1860-7. Crossref

22. Kong L, Cao J, Zhang Y, Ding W, Shen Y. Risk factors for periprosthetic joint infection following primary total hip or knee arthroplasty: a meta-analysis. Int Wound J 2017;14:529-36. Crossref

23. Triffault-Fillit C, Ferry T, Laurent F, et al. Microbiologic epidemiology depending on time to occurrence of prosthetic joint infection: a prospective cohort study. Clin Microbiol Infect 2019;25:353-8. Crossref

24. Zeller V, Kerroumi Y, Meyssonnier V, et al. Analysis of postoperative and hematogenous prosthetic joint-infection microbiological patterns in a large cohort. J Infect 2018;76:328-34. Crossref

25. Benito N, Franco M, Ribera A, et al. Time trends in the aetiology of prosthetic joint infections: a multicentre cohort study. Clin Microbiol Infect 2016;22:732.e1-8. Crossref

26. Kherabi Y, Zeller V, Kerroumi Y, et al. Streptococcal and Staphylococcus aureus prosthetic joint infections: are they really different? BMC Infect Dis 2022;22:555. Crossref

27. Sendi P, Rohrbach M, Graber P, Frei R, Ochsner PE, Zimmerli W. Staphylococcus aureus small colony variants in prosthetic joint infection. Clin Infect Dis 2006;43:961-7. Crossref

28. Diekema DJ, Pfaller MA, Shortridge D, Zervos M, Jones RN. Twenty-year trends in antimicrobial susceptibilities among Staphylococcus aureus from the SENTRY Antimicrobial Surveillance Program. Open Forum Infect Dis 2019;6(Suppl 1):S47-53. Crossref

29. Ravi S, Zhu M, Luey C, Young SW. Antibiotic resistance in early periprosthetic joint infection. ANZ J Surg 2016;86:1014-8. Crossref

30. Bejon P, Berendt A, Atkins BL, et al. Two-stage revision for prosthetic joint infection: predictors of outcome and the role of reimplantation microbiology. J Antimicrob Chemother 2010;65:569-75. Crossref

31. Berbari EF, Marculescu C, Sia I, et al. Culture-negative prosthetic joint infection. Clin Infect Dis 2007;45:1113-9. Crossref

32. Malekzadeh D, Osmon DR, Lahr BD, Hanssen AD, Berbari EF. Prior use of antimicrobial therapy is a risk factor for culture-negative prosthetic joint infection. Clin Orthop Relat Res 2010;468:2039-45. Crossref