Periprosthetic joint infection following Staphylococcus aureus bacteremia

Journal of Infection (2011) 63, 17e22

www.elsevierhealth.com/journals/jinf

Periprosthetic joint infection following Staphylococcus aureus bacteremia Parham Sendi a,b,c,*, Florian Banderet a, Peter Graber a, Werner Zimmerli a a

Unit of Infectious Diseases, Basel University Medical Clinic Liestal, Switzerland University Clinic for Infectious Diseases, University Hospital Bern, Switzerland c Institute for Infectious Diseases, University of Bern, Switzerland b

Accepted 6 May 2011 Available online 14 May 2011

KEYWORDS S. aureus bacteremia; Prosthetic joint infections; S. aureus

Summary Objectives: The incidence of haematogenous periprosthetic joint infections (PJI) among patients with remote infections has been reported to be less than 1%. This incidence may be much higher in cases after documented Staphylococcus aureus bacteremia (SAB). We evaluated the incidence of haematogenous PJI following SAB in patients with previously uninfected arthroplasties. Methods: A retrospective analysis of our cohort including patients with SAB and prosthetic joints at the Basel University Medical Clinic Liestal from 1998 to 2008. Results: We identified 31 patients with 45 uninfected prosthetic joints in situ at the time of SAB. In 12 patients (39%) and 13 arthroplasties (29%), SAB caused PJI. In comparison to nosocomial SAB, infections occurred only in cases with community-acquired SAB (p Z 0.002). PJI was diagnosed within a median time of 2.5 days (IQR 1e3.5) after admission. The comparison between patients with and without PJI revealed no significant difference in gender, age, comorbidities and number of prostheses per patient and age of the prosthesis. Conclusions: The rate of PJI after SAB is high, ranging from 30% to 40%, and clearly higher than rates reported for bacteremia with other pathogens. PJIs were observed in community-onset bacteremia, in which there is a typically delay from symptoms to antimicrobial treatment. ª 2011 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

Introduction The number of primary total hip and knee arthroplasties has been significantly increasing in recent years, with nearly 800,000 prosthetic joints implanted in the United States

alone in 2006.1 Periprosthetic joint infection (PJI) is an uncommon but serious complication. It occurs exogenously (e.g. infection inoculated at implantation) or haematogenously through seeding from a remote origin to the implant. The overall incidence of haematogenous PJI among patients

€hlstrasse 51, CH-3010 Bern, Switzerland. Tel.: þ41 31 * Corresponding author. Institute for Infectious Diseases, University of Bern, Friedbu 632 32 99; fax: þ41 31 632 49 66. E-mail address: [email protected] (P. Sendi). 0163-4453/$36 ª 2011 The British Infection Association. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2011.05.005

18 with remote infections has been reported to be 0.1e0.6%.2 However, after documented Staphylococcus aureus bacteremia (SAB), the risk of bacterial seeding on a prosthetic joint has been reported much higher. To our best knowledge, only 2 studies have evaluated this risk, reporting an incidence of 29% and 34%, respectively.3,4 Considering that in parallel to the number of arthroplasties, the incidence of SAB has also increased over the last decades,5 we evaluated the incidence of haematogenous (i.e. secondary) PJI following SAB in patients with previously uninfected arthroplasties.

P. Sendi et al. for S. aureus9 infection and PJI1 were compared between cases with and without PJI. The variables included chronic heart disease, chronic lung disease, chronic liver failure, chronic renal failure, rheumatic disease, corticosteroid use, diabetes mellitus, peripheral vascular occlusive disease, cancer, metabolic syndrome, chronic skin disease and IVDU. Further variables for comparison included communityacquired bacteremia, hospital-acquired bacteremia, 30-day mortality and SAB-related mortality.

Statistical analysis

Patients and methods Study population A blood culture database (Basel University Medical Clinic, Liestal, Switzerland) was screened for S. aureus from 1998 through 2008. The patient charts of all episodes with S. aureus grown in blood cultures were reviewed for the presence of both prosthetic hip and/or knee joints and symptoms of sepsis.

Definitions, inclusion and exclusion criteria Symptoms of sepsis were defined according to the consensus statement of the American College of Chest Physicians and Society of Critical Care Medicine.6 SAB was defined as 1 positive blood culture plus sepsis syndrome. SAB was considered hospital-acquired when it appeared after 48 h of hospital admission. Otherwise, SAB was considered as community-acquired. Diagnosis of PJI in association with SAB included 1 positive synovial fluid or tissue culture and 1 of the following criteria: purulence surrounding the joint, or acute inflammation in histopathology.7 In cases of SAB-related death, only one of the latter criteria was required for PJI diagnosis. SAB-related death was defined as decease caused by the bacteremia.8 PJI was classified as haematogenous only if the infection occurred 2 years after implantation or if it occurred <2 years and an obvious distant focus was evident and the arthroplasty had no previous signs of infection. Inclusion criteria: Patients with documented SAB and no clinical or radiological signs for PJI in their previous history were included in the study. Exclusion criteria: We excluded referred patients with positive blood cultures results from external laboratories and patients with haematogenous PJI but without documented SAB (i.e. missing or negative blood culture results). Patients and prosthetic joints with definite or possible primary PJI (i.e. exogenous) were excluded from analyses.

Variables associated with PJI and Outcome For the outcome analyses of patients with SAB and previously uninfected arthroplasties, development of PJI was looked for. Patients without PJI were prospectively followed to evaluate whether infection presented later (i.e., after stopping antimicrobial treatment for SAB). Patient-specific variables, comorbidities, and risk factors

Univariate assessment by chi-square, Fisher’s exact test, and t test were performed where appropriate. Variables having a 2-tailed p value of <.05 were considered significant. The statistical software used was S-Plus 2000.

Results During the 11-year period, 33 patients with 50 prosthetic joints in situ at the time of SAB were identified. Two patients and 5 prosthetic joints were excluded from analyses (Fig. 1). In one patient, early exogenous PJI was evident. In a second patient with 2 total hip arthroplasties, double PJIs occurred 11 months after implantation. Although haematogenous infection was likely, exogenous PJI could not be ruled out, because no distant focus was identified and the time interval between implantation and infection was <2 years (see definition of haematogenous PJI in Patients and methods). In addition, 2 arthroplasties (but not the patients) in cases with multiple prosthetic joints were excluded because the joints were the source of SAB (i.e. early exogenous PJI). Therefore, 31 patients with 45 previously uninfected prosthetic joints and SAB were included in analyses (Fig. 1).

S. aureus bacteremia and mortality All S. aureus strains were methicillin susceptible. The sources of SAB are presented in Table 1. In 5 episodes, the time interval between implantation and SAB was <2 years, but an obvious primary distant focus was identified. Strain typing to establish whether the PJI and SAB isolates were identical was not performed. Primary empiric intravenous antimicrobial treatment exerted antistaphylococcal activity in all cases and included penicillin derivatives in 20 (65%) episodes and cephalosporins in 8 (26%). The SAB-related mortality rate was 23% (7 patients), and the overall 30-day mortality rate was 39% (12 patients).

Haematogenous seeding to arthroplasty with no previous signs of infection In 12 of 31 patients (39%) and 13 of 45 arthroplasties (29%), PJI resulted from haematogenous seeding (Fig. 1). Details of these cases are presented in Table 2. In all cases with PJI, S. aureus grew in blood cultures within 24 h after admission. Hence, PJI occurred only in cases with community-acquired SAB (p Z 0.002, Table 3). PJI was diagnosed within a median time of 2.5 days (IQR 1e3.5) after admission. In 3 of 12 patients, the presumed diagnosis was confirmed at necropsy,

S. aureus bacteremia and prosthetic joint infections S.aureus bacteremia and prosthetic joints 33 patients with 50 prosthetic joints

Definite or possible exogenous PJI 2 patients 5 prosthetic joints

19 prosthesis, and hence no puncture was performed. In the second case, joint puncture revealed only 1 cell/mm3 and no growth of microorganisms. In the third case (exitus day 25, not SAB related), clinical examination of the joint arthroplasties was without pathological findings, but ultrasound, joint puncture, and necropsy were not performed. In the other 13 cases without PJI, the median interval between SAB and last clinical follow-up was 11.8 months (IQR 7e34).

PJI versus no PJI S.aureus bacteremia and prosthetic joints without previous signs of infection 31 patients 45 prosthetic joints Haematogenous seeding to arthroplasty (patients)

PJI 12 patients (39%)

no PJI 19 patients (61%)

Haematogenous seeding to arthroplasty (prosthetic joints)

PJI 13 joints (29%) Figure 1

no PJI 32 joints (71%) Flowchart of the study.

although in one patient PJI was also evident by synovial culture. In 19 cases without PJI, 6 died within 30 days (Table 3). In 3 of these cases, PJI was excluded at necropsy. In the other 3 cases, prosthetic joints were examined during hospitalization, but showed no pathological findings. In the first case, ultrasound showed no synovial fluid around the

Table 1

Comparison of the most frequent characteristics between patients with and without PJI is presented in Table 3. There was no significant difference in gender, age, and number of prostheses per patient. The median age of the prosthesis was similar in both groups, 11 years (IQR 6e15) in arthroplasties with infection (n Z 13) versus 7 years (IQR 4e11) in those without infection (n Z 32). There was also no difference in the frequency of rheumatic disorder and corticosteroid use. However, patients with PJI had significantly more often diabetes (p Z 0.037). As indicated previously, communityacquired SAB was significantly more frequent in PJI cases than it was in cases without PJI (p Z 0.002, Table 3).

Discussion Patients with SAB are at risk for a wide range of complications, and metastatic infections can occur in almost any body site. The rate of haematogenous seeding to native joints has been estimated at 3e10%.10e12 However, the risk of haematogenous seeding to prosthetic joints is much higher, since the presence of an implant massively decreases the minimal infecting dose of staphylococci. In an experimental model, 107 colony-forming units (cfu) of S. aureus did not produce any abscesses in the absence of foreign material, whereas 102 cfu were enough to permanently infect subcutaneous implants.13 In a clinical study, the risk of prosthetic joints becoming infected by haematogenous seeding after SAB was 34% (15 of 44 patients).3 In their analysis, an infection was classified as haematogenous, if it occurred after 1 year of implantation and there was no evidence of direct inoculation. Lalani et al.4 reported a rate of 29% (16 of 55 patients). However, only one third of their patients had prosthetic joints, and two

Sources of Staphylococcus aureus bacteremia in 31 patients with prosthetic joints.

Infection source

No. (%)

Skin and soft tissue No source identified Catheter associateda Vertebral osteomyelitisb Pneumonia Contralateral periprosthetic jointc

9 8 6 4 2 2

a b

c

(29) (26) (19) (13) (6) (6)

Source included central line, peripheral line and Sheldon catheter. Vertebral osteomyelitis might be due to silent bacteremia that occurred prior to presentation at our center. All cases claimed of back pain. Three patients developed community-acquired, one patient hospital-acquired bacteremia. In one case, haematogenous seeding to prosthetic joint was noted (Table 2). Two arthroplasties in patients with multiple prosthetic joints were early exogenous PJI, and hence the source of SAB.

20

Table 2

Clinical characteristics of patients with Staphylococcus aureus bacteremia and periprosthetic joint infection. Infected Time of SABa prosthesis

Age of Mode of PJI prosthesis, y diagnosis

Time of PJI Death Death diagnosis, db (30 d)c related to SAB

13 7 15 7 21

Histology sf, histology sf sf, histology Histology

3, necropsy 3 4 1, necropsy 5, necropsy

Yes Yes Yes Yes Yes

Yes No Yes Yes Yes

Left hip 4 Both knees 3 Right hip 10

sf, biopsy sf sf, biopsy

0 9 1

No No No

e e e

Left hip Right hip

15 11

sf, biopsy sf, biopsy

1 1

No No

e e

Left hip Left hip

23 3

sf sf

2 5

Yes No

Yes e

No. Sex Age, No. of y prostheses

Total no. of Risk factor(s) comorbidities of PJI or SAB

Source of SAB

1 2 3 4 5

m f m m f

80 77 72 70 66

2 1 2 2 2

2 1 4 2 3

Unknown 0 Cellulitis 0 Pneumonia 0 Foot ulcer 1d Vertebral 1d osteomyelitis

Right hip Left hip Left hip Left hip Left knee

6 7 8

f m f

58 88 55

1 THA 2 TKAs 1 THA þ 1 TKA

Unknown Pneumonia Unknown

9 10

m f

63 41

1 THA 0 2 THAs þ 2 TKAs 2

11 12

m m

84 78

1 THA 1 THA

THAs THA THAs THAs TKAs

1 3 2

2 1

Corticosteroids None Diabetes Diabetes Diabetes; corticosteroids; obesity None Diabetes Rheumatoid arthritis; obesity None Rheumatoid arthritis; corticosteroids Diabetes None

0 0 0

Unknown Unknown

0 1d

Unknown Unknown

0 0

m Z male; f Z female; THA Z total hip arthroplasty; TKA Z total knee arthroplasty; PJI Z periprosthetic joint infection; SAB Z Staphylococcus aureus bacteremia; y Z years; d Z days; sf Z synovial fluid culture. a Time of SAB Z interval from admission to culture positivity; 0 Z bacteria grew in blood culture within 24 h. b Time of PJI diagnosis Z interval from admission to diagnosis of PJI. c Mortality within 30 days after admission.

P. Sendi et al.

S. aureus bacteremia and prosthetic joint infections Table 3

21

Comparison of patients’ characteristics with and without periprosthetic joint infections.

Characteristics

Median age, y (IQR) Male, no. (%) Median no. prostheses per patient Median no. comorbidities Diabetes, no. (%) Rheumatic disease, no. (%) Corticosteroid use, no. (%) CA bacteremia, no. (%) 30-day mortality, no. (%) Sepsis-related mortality, no. (%)

PJI Yes (n Z 12)

No (n Z 19)

71 7 2 2 5 2 3 12 6 5

78 12 1 2 1 3 3 11 6 2

(62e79) (58)

(42) (17) (25) (100) (50) (42)

(73e83) (63)

(5) (16) (16) (58) (32) (11)

p value (two-tailed) 0.14 0.79 0.27 0.43 0.037 0.95 0.56 0.002 0.33 0.08

PJI Z periprosthetic joint infection; IQR Z interquartile range, CA bacteremia Z community-acquired bacteremia.

thirds had various other orthopedic devices. They classified an infection as haematogenous, if there was no contiguous focus of infection associated with the site of prosthesis insertion. Our results were similar (29% of prosthetic joints and 39% of patients), even though our criteria for haematogenous PJI were stricter, requiring at least a 2-year interval between implantation and infection, or the presence of an obvious primary distant focus. We chose a cut-off of 2 or more years to include only ‘late’ PJIs which are almost always due to haematogenous seeding, and hence, reliably exclude exogenous infections. The classification of PJI in early (<3 months after surgery), delayed (3e24 months after surgery) and late (>24 months after surgery) is arbitrary, but helpful in proposing the pathogensis.14 Early exogenous infections are typically caused by virulent microorganisms, whereas delayed exogenous infections are caused by less virulent bacteria such as coagulase-negative staphylococci or Propionibacterium acnes.7 However, haematogenous seeding from a remote focus to the arthroplasty can occur at any time after implantation. Therefore, this classification is not optimal for distinguishing exogenous and haematogenous PJI, if virulent microorganisms, such as S. aureus are involved in an infection that occurred less than 2 years after surgery. Nevertheless, we reviewed all patient charts in detail and only one case in our study caused difficulties to distinguish between exogenous and haematogenous origin (i.e. SAB 11 months after implantation) and was therefore excluded. The rate of haematogenous PJI after SAB may be overestimated in our study because the true incidence of SAB in a large community is unknown (i.e., from various clinics and private laboratories). Also, SAB is not always detected (i.e., treatment without obtaining blood cultures). On the other hand, we included only patients in whom SAB was documented in our laboratory. We excluded referred patients with results from external laboratories and patients with haematogenous PJI but without documented SAB. Therefore, we consider a PJI rate of 30e40% after documented community-acquired SAB to be an accurate risk estimate. Moreover, although the numbers of patients in Murdoch et al.’s,3 Lalani et al.’s4 and our study are small, the results of these series are consistent. Likewise, similar high incidences of metastatic infections after SAB have been

reported in other studies,15,16 in particular when a foreign body material was involved.17,18 Community-acquired SAB was significantly more common among patients with PJI. This finding is in agreement with other studies showing that patients with community-acquired bacteremia have a higher risk for secondary foci than do those with nosocomial bacteremia.11 This observation may be related to the duration of bacteremia, which is commonly longer in communitye than in hospital-acquired cases.19 However, in our study, the onset, and hence, the duration of community-acquired SAB prior to hospital admission are unknown. Although patients with diabetes are more prone to SAB,9 we cannot explain why diabetes was more frequent in cases with than in those without PJI (Table 3). Considering the small numbers in each group, this result is difficult to interpret. The SAB-related mortality rate, however, was 23%, similar to that of other reports.20 Our single center study is retrospective and consists of a small number of patients. However, PJI is a rare event. Therefore, evaluating the incidence of haematogenous PJI following SAB in patients with previously uninfected arthroplasties is only feasible in centers that have a cohort of prosthetic joints. The number of patients in our study is comparable to that in the other 2 studies.3,4 In conclusion, the rate of PJI after SAB is high, ranging from 30% to 40%. This rate is consistent in several studies,3,4 and clearly higher than rates reported for bacteremia with other pathogens.2 PJIs were observed in community-onset bacteremia, in which there is a delay from symptoms to antimicrobial treatment. In our view, these data indicate the following for clinical practice. First, considering the high incidence of haematogenous seeding is the key to early diagnosis of PJI. Second, in case of SAB in patients with arthroplasties that had no previous signs of infection, prosthetic joints should be clinically and radiologically (e.g. ultrasound) closely monitored for signs of infection. If PJI is suspected, diagnosis should be ascertained by synovial cell count21 and microbiological and histological examinations of obtained specimens.22 Confirmed PJI should be treated as described previously.1,5 Rapid diagnosis of PJI is important, because it may avoid an exchange of the prosthesis; these infections can be cured by debridement and retention of the implant.7,23

22

Acknowledgment Presented in part: 50th Interscience Conference on Antimicrobial Agents and Chemotherapy, Boston, M.A., 12e15 September 2010 (abstract K-1172). Financial disclosure: This study was supported by the ‘Infektiologie beider Basel’ foundation. Conflict of interest: none, for all authors.

References 1. Del Pozo JL, Patel R. Clinical practice. Infection associated with prosthetic joints. N Engl J Med 2009;361:787e94. 2. Uckay I, Lubbeke A, Emonet S, Tovmirzaeva L, Stern R, Ferry T, et al. Low incidence of haematogenous seeding to total hip and knee prostheses in patients with remote infections. J Infect 2009;59:337e45. 3. Murdoch DR, Roberts SA, Fowler Jr VG, Shah MA, Taylor SL, Morris AJ, et al. Infection of orthopedic prostheses after Staphylococcus aureus bacteremia. Clin Infect Dis 2001;32:647e9. 4. Lalani T, Chu VH, Grussemeyer CA, Reed SD, Bolognesi MP, Friedman JY, et al. Clinical outcomes and costs among patients with Staphylococcus aureus bacteremia and orthopedic device infections. Scand J Infect Dis 2008;40:973e7. 5. Lyytikainen O, Ruotsalainen E, Jarvinen A, Valtonen V, Ruutu P. Trends and outcome of nosocomial and community-acquired bloodstream infections due to Staphylococcus aureus in Finland, 1995e2001. Eur J Clin Microbiol Infect Dis 2005;24:399e404. 6. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101: 1644e55. 7. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med 2004;351:1645e54. 8. Blot SI, Vandewoude KH, Hoste EA, Colardyn FA. Outcome and attributable mortality in critically ill patients with bacteremia involving methicillin-susceptible and methicillin-resistant Staphylococcus aureus. Arch Intern Med 2002;162:2229e35. 9. Jacobsson G, Dashti S, Wahlberg T, Andersson R. The epidemiology of and risk factors for invasive Staphylococcus aureus infections in western Sweden. Scand J Infect Dis 2007;39:6e13. 10. Espersen F, Frimodt-Moller N, Thamdrup Rosdahl V, Skinhoj P, Bentzon MW. Changing pattern of bone and joint infections due to Staphylococcus aureus: study of cases of bacteremia in Denmark, 1959e1988. Rev Infect Dis 1991;13:347e58.

P. Sendi et al. 11. Lautenschlager S, Herzog C, Zimmerli W. Course and outcome of bacteremia due to Staphylococcus aureus: evaluation of different clinical case definitions. Clin Infect Dis 1993;16: 567e73. 12. Ghanem GA, Boktour M, Warneke C, Pham-Williams T, Kassis C, Bahna P, et al. Catheter-related Staphylococcus aureus bacteremia in cancer patients: high rate of complications with therapeutic implications. Medicine (Baltimore) 2007;86:54e60. 13. Zimmerli W, Waldvogel FA, Vaudaux P, Nydegger UE. Pathogenesis of foreign body infection: description and characteristics of an animal model. J Infect Dis 1982;146:487e97. 14. Schafroth M, Zimmerli W, Brunazzi M, Ochsner PE. Infections. In: Ochsner PE, editor. Total hip replacement. Berlin: Springer-Verlag; 2003. p. 65e90. 15. Cuijpers ML, Vos FJ, Bleeker-Rovers CP, Krabbe PF, Pickkers P, van Dijk AP, et al. Complicating infectious foci in patients with Staphylococcus aureus or Streptococcus species bacteraemia. Eur J Clin Microbiol Infect Dis 2007;26:105e13. 16. Ringberg H, Thoren A, Lilja B. Metastatic complications of Staphylococcus aureus septicemia. To seek is to find. Infection 2000;28:132e6. 17. Uslan DZ, Dowsley TF, Sohail MR, Hayes DL, Friedman PA, Wilson WR, et al. Cardiovascular implantable electronic device infection in patients with Staphylococcus aureus bacteremia. Pacing Clin Electrophysiol 2010;33:407e13. 18. Chamis AL, Peterson GE, Cabell CH, Corey GR, Sorrentino RA, Greenfield RA, et al. Staphylococcus aureus bacteremia in patients with permanent pacemakers or implantable cardioverter-defibrillators. Circulation 2001;104:1029e33. 19. Khatib R, Johnson LB, Fakih MG, Riederer K, Khosrovaneh A, Shamse Tabriz M, et al. Persistence in Staphylococcus aureus bacteremia: incidence, characteristics of patients and outcome. Scand J Infect Dis 2006;38:7e14. 20. Wyllie DH, Crook DW, Peto TE. Mortality after Staphylococcus aureus bacteraemia in two hospitals in Oxfordshire, 1997e2003: cohort study. BMJ 2006;333:281e4. 21. Trampuz A, Hanssen AD, Osmon DR, Mandrekar J, Steckelberg JM, Patel R. Synovial fluid leukocyte count and differential for the diagnosis of prosthetic knee infection. Am J Med 2004;117:556e62. 22. Trampuz A, Zimmerli W. Prosthetic joint infections: update in diagnosis and treatment. Swiss Med Wkly 2005;135: 243e51. 23. Zimmerli W, Widmer AF, Blatter M, Frei R, Ochsner PE. Role of rifampin for treatment of orthopedic implant-related staphylococcal infections: a randomized controlled trial. Foreign-Body Infection (FBI) Study Group. JAMA 1998;279: 1537e41.