Abstract
Abstract
Background:
This prospective study investigated the role of mechanical and biological factors in aseptic implant loosening or presumed silent prosthetic joint infection (PJI).
Methods:
Thirty-seven patients were investigated. Microbiologic and molecular methods were used to detect bacteria on the surface of the failed implants removed during revision arthroplasty. Histopathologic analysis was performed. The influence of body mass index (BMI) and various co-morbidities on implant failure also was determined.
Results:
The results of sonicated fluid cultures were positive for bacteria in 29.7% and the results of intra-operative tissue and joint liquid cultures in 18.9% and 16.2%, respectively. Molecular detection with 16S rRNA sequencing revealed a large variety of bacteria. The most frequent organism was coagulase-negative Staphylococcus (CNS). The outcomes of histopathologic tests of peri-prosthetic tissue showed evidence of the infection type in all culture-positive joints and in 41.4% of the cases with negative culture results. Overweight status or obesity was present in 82.8% of the culture-negative patients.
Conclusions:
The results of this study proved the presence of micro-organisms on the surface of implants in both aseptic and presumed PJI cases. Inclusion of the sonication procedure in the diagnostic algorithm increased the ability to identify the pathogen. The results of our study suggest the co-existing roles of BMI and the time to implant loosening as well as biological agents in causing prosthesis loosening.
T
This prospective study was designed to investigate the microbiologic cultures of joint aspiration fluid and intra-operative tissue specimens and to compare the outcomes with sonication fluid cultures in cases supposed pre-operatively to be either aseptic loosening or presumed PJI.
Patients and Methods
Patients
A total of 37 patients aged 39–81 y (median 65 y) with presumed aseptic hip or knee prosthesis loosening were included in the study. Patients had attended the Department of Orthopaedic and Traumatology, Medical University of Silesia, School of Medicine in Katowice, Poland from 2012 to 2015. The average period between the first and second stage of revision arthroplasty was 110 months (approximately 9 years and 2 months). Body mass index (BMI) and co-morbidities were assessed for all patients (Table 1). The average followup on these 37 patients was 27.8 mos (range 21.3–36.1 mos). No patients were lost to followup. We defined treatment failure as recurrence of infection in the same hip or knee and the need for long-term antibiotics. Clinical examination and radiologic assessment were conducted, and laboratory markers of infection were determined.
Inclusion criteria: Patients with aseptic hip or knee prosthesis loosening or presumed PJI. Qualification criteria: Clinical signs of prosthesis loosening (positive tests in physical examination), radiologic features of loosened implants, normal or elevated concentrations of serum laboratory markers of infection (C-reactive protein [CRP], white blood cell [WBC] count) if there was a lack of or insufficient liquid volume recovered by pre-operative joint aspiration.
Exclusion criteria: Signs or symptoms of infection (local inflammatory signs, sinus tract, or systemic symptoms of infection), antibiotic administration two weeks before revision arthroplasty, rheumatoid arthritis, immunosuppression or chemotherapy, and lack of patient consent for participation in the study.
Study approval was obtained from the Ethics Committee of the Medical University of Silesia in Katowice, Poland. Informed consent was obtained from all participants.
Surgical management
The operative procedures included removal of loosened components of the prosthesis and cement if present; debridement; and partial or total revision exchange arthroplasties. In cases where there was a high suspicion of PJI (visible change in peri-prosthetic tissues, liquid in the joint), commercial antibiotic-loaded cement spacers were implanted.
Microbiologic methods
Culture
A sample of joint liquid obtained by pre- or intra-operative aspiration and three to six intra-operative tissue samples harvested during revision surgery were cultured. Directly after surgery, the removed components of the loosened prostheses were placed in a sterile container and subjected to sonication according to the diagnostic scheme presented in our previous [2] and other authors' [3,4] papers. The cultures were kept for as long at 14 days to recover slow-growing micro-organisms.
An etiologic agent of peri-prosthetic infection was considered to be the same micro-organism isolated from two examined materials (joint fluid, intra-operative specimen, sonicates). The results of tissue cultures were said to be positive if pathogen growth was obtained from at least two samples taken from the affected prosthetic joint [5].
Molecular methods
Molecular detection of micro-organisms was performed in the cases of aseptic loosening of a prosthesis as previously described [2].
Histopathologic Tests
Soft tissue surrounding the implants and peri-prosthetic interface membrane were taken for histopathologic testing. The outcomes were recorded according to the Krenn and Morawietz classification [6].
Statistical analysis
Distribution properties of studied characteristics were performed using the Shapiro-Wilk test and defined measures of dispersion. To evaluate the correlation between the variables, the χ2 and Fisher exact test were used. Significance tests (Student t-test and Mann-Whitney U test) were used to verify the hypothesis that there was no difference between variables. P values <0.05 were considered statistically significant. Calculations were performed using software from StatSoft, Inc. (STATISTICA, version 12, 2014; www.statsoft.com).
Results
Systemic inflammatory markers
The WBC counts ranged from 4,500/mm3 to 15,600/mm3, the average value being 5,890/mm3. The counts were normal in 35 cases; in two (5.4%), an elevated WBC count was recorded. In these two cases, the bacteria were recovered only by sonication (Table 2, cases 9 and 12).
C = cemented; CRP = C-reactive protein; N = normal value (CRP: N < 10 mg/L; WBC: N < 10,000/mm3); P-atb – prolonged antibiotic therapy, Pih = prolonged incision healing; UC = uncemented; WBC = white blood cell.
Elevated concentrations of CRP were reported in five cases (range 18.0–93.4 mg/L; median 32.5 mg/L). In patients with positive cultures, CRP was elevated in four cases. In the group of patients with negative cultures, CRP was above 10 mg/L in only one case (p = 0.005).
BMI
No organisms were more common in patients with higher BMI (Fig. 1).
The distribution of BMI in group of patients with positive (n = 8) and negative (n = 29) culture results.
Microbiologic results
The culture results of specimens are presented in Table 2.
The results of sonicated fluid cultures were positive in 29.7% of the cases (11/37); the intra-operative tissues and joint liquid cultures were positive in 18.9% (7/37) and 16.2% (6/37), respectively (Table 3). The growth of the same species—Staphylococcus aureus—was noted at both sites in two cases (patients no. 1 and 2). In the following six cases (patients no. 3–8), the growth of the same species was observed in at least two of the three examined specimens. In case 8, pathogens were isolated by culture from at least two intra-operative samples. In the next five cases (patients no 9–13), the cultures were positive for Staphylococcus epidermidis in two cases and for Ralstonia pickettii in three cases.
Molecular detection with 16S rRNA sequencing was performed in the remaining cases when micro-organism growth was not observed. Sequencing results revealed a large variety of bacteria in each case. The most frequent isolate was CNS and other representatives of human and environmental microflora, much as in our previous study [2].
Operative treatment options
One-stage arthroplasty was performed in 22 of 29 culture-negative patients (76%): partial exchange of prosthesis elements was done in 41% (12/29) and total exchange arthroplasty in 35% (10/29). A two-stage procedure with the use of a spacer was performed in 24% of culture-negative patients with probable peri-prosthetic infection. Two-stage arthroplasty was performed in 12 cases of the entire series and in five of eight patients (63%) with positive culture results. The details are presented in Tables 2 and 4.
Histopathologic results
The outcomes of histopathologic tests of peri-prosthetic tissue showed evidence of infectious type (type II or III) in all cases of positive cultures (cases no 1–8) and in 12 of the 29 patients with negative cultures (p = 0.004).
Discussion
Joint prosthesis loosening can be the result of either an aseptic or an infective process. Aseptic loosening is caused by the lack of stability of a prosthesis because of inadequate integration with the bone. The concept that infection is not involved is changing, as the presence of pathogens on the surfaces of implants has been proved [2,7]. Our study was performed to establish the usefulness of microbiologic tests in determining a possible role for infectious agents in supposedly aseptic loosening. Clinical characteristics are the main guide in the initial suspicion of the cause. Thus, the absence of local inflammatory signs or a sinus tract supports the diagnosis of an aseptic process.
C-Reactive protein as a non-specific marker can be useful during the first stage of diagnosis. The five patients with significantly elevated concentrations of CRP (but with no obvious local or general infection) were strongly suspected of having PJI. In our study, the higher concentration of CRP (>10 mg/L) was common in patients with bacterial growth in the sonicate fluid (S. epidermidis in one case) or in synovial liquid or intra-operative specimens (p = 0.005).
According to Berbari's research group, both the erythrocyte sedimentation rate (ESR) and a concentration of CRP within the physiologic range are evidence of non-inflammatory loosening of hip joint prostheses [8,9,10]. Some other investigators confirmed physiologic ranges of CRP in about 50% of patients with aseptic loosening [11]. It was stated that CRP is a non-specific indicator, and its concentration can remain elevated during the course of many diseases [10,11,12]. The WBC count, a multifunctional diagnostic parameter, cannot be associated with the implant loosening process only.
In our study, the number of patients with positive cultures increased after sonication, from 24.3% to 35.1%. Higher sensitivity of sonication in comparison with conventional peri-prosthetic culture was reported in our previous study (37.5% vs. 43.7%) and in other studies; e.g., that by Trampuz et al. (78.5% vs. 60.8%) [2,3]. Our study revealed that culture of the sonicate included in the diagnostic algorithm of peri-prosthetic infection allowed us to determine the etiologic agents in eight of 37 cases (21.6%). Among the bacteria isolated, besides typical etiologic agents of peri-prosthetic infection, R. pickettii was isolated in four cases after using sonication. The presence of R. pickettii has been noted in as many as 53% of patients in other clinical studies [15]. This bacterium is described as a non-fermenting gram-negative bacillus with low virulence living in wet environments (water, skin disinfectants, skin care products). It can colonize human tissues without causing any symptoms. It has been emphasized lately that even micro-organisms living in a natural environment can be serious pathogens, especially for patients with risk factors and immunodeficiency. It should be noted that the presence of such environmental bacilli may cause false-positive results associated with contamination.
Surgical findings (macroscopic pus or histologic change) together with cultures of surgical samples have identified some infections in patients with presumed aseptic loosening. This situation has been well defined by Tsukayama et al. as a particular type of infection, namely, intra-operative positive cultures [13] or, more recently, by other authors as subclinical PJI [14].
Some studies show that the peri-prosthetic membrane is the ideal material to characterize the histologic type of inflammation, thus providing valuable evidence for the underlying cause of implant loosening [16,17]. Our results revealed a significant correlation (p = 0.004) between positive microbiologic culture results and detection of type 2 or type 3 peri-prosthetic membranes. Infections types were revealed in 12 of 29 patients with negative culture results. Similarly, the study of Hischebeth et al. included 18 patients with histopathological evidence of infection but no bacterial growth in cultures of tissue samples [18]. This failure might have resulted from antibiotic treatment prior to surgery, which is a well-known reason for negative outcomes of culture [19].
In patients with negative cultures, the most important factor in the loosening of prosthesis probably was a higher body weight. In this group, a BMI above 25 was noted in 82.8% (24/29) compared with the group of patients with loosening and a positive culture result (62.5%; 5/8) (Fig. 1). Although this correlation was not statistically significant, in our opinion, further studies on a larger group of patients could well prove the importance of BMI in the loosening of a prosthesis.
When analyzing the co-morbidities, previous orthopaedic operations (not only of the examined joints) had been performed in 51%. The same percentage of patients suffered from cardiac diseases (mostly ischemic heart disease), which were significantly more common in patients with positive cultures (p = 0.04).
Most of patients with a pre-operative diagnosis of aseptic loosening underwent one-stage exchange arthroplasties (25/37; 67%), whereas all the patients with PJI highly suspected because of significant elevated CRP underwent two-stage revision arthroplasty. These treatment options are recommended by other authors [20, 21].
In our study, during an average of two years' followup, nine cases were judged to be failures. The complications were more frequent in the patients with positive cultures than in patients with negative cultures, 50% and 17.2%, respectively (p = 0.08). Failures were observed in 42.8% of the treated knees and 20% of the treated hips (Table 4). The most common complication was prolonged need for antibiotics caused by impaired wound healing. In cases with positive joint fluid cultures, the incidence of re-operation was higher. In two cases of likely PJI, the patients required re-operation. In the first case (patient no. 1), the patient needed prolonged antibiotic therapy as well as surgery because of obscure pain in the patellofemoral joint, and resurfacing of the patella was conducted. This operation was not considered a failure, because the mechanical problem was the reason for the operation. The second case (patient no. 2) finally underwent arthrodesis of the knee because of the general medical condition and the high risk of re-infection.
Some authors have proved a correlation between prosthesis age and failure risk, in that early loosening is more frequently caused by hidden PJI than is late loosening (>24 mos), 40% vs. 22%, respectively [14,22]. In our study, the period before prosthesis loosening was less than 24 mos in two of the eight patients with positive cultures. There were three patients with early loosening in group with negative cultures (3/29; 10.3%).
Although we had only a small number of patients, our results and the outcomes of other authors, the lack of clinical signs of infection, and the negative pre-operative joint aspirations and intra-operative specimens do not exclude the possibility of bacteria on the implants in cases of aseptic loosening of the prostheses. There is a need to follow procedures in learning about the role of uncommon bacteria from the biologic materials in peri-prosthetic infections, as in our series.
In summary, we found micro-organisms on the surface of implants removed because of clinical features of aseptic loosening of a prosthesis. Inclusion of the sonication procedure in the diagnostic algorithm of prostheses loosening increases the ability to identify pathogens. Previous operations and heart disease are more common than other co-morbidities and probably have a greater influence on implant loosening. The results of this study suggest the co-existing roles of the BMI and the period to implant loosening as well as biologic agents of loosening of prostheses. Our analysis proved better outcomes from operative treatment of asymptomatic PJI using one-stage operations than two-stage revision arthroplasty.
Footnotes
Acknowledgments
This study received funding from the Medical University of Silesia in Katowice (KNW-1-092/K/5/0). Equipment for molecular analysis was purchased under Silesian Bio-Farma Center for Biotechnology, Bioengineering and Bioinformatics Project no POIG.02.01.00-00-166/08 THE OPERATIONAL PROGRAMME INNOVATIVE ECONOMY FOR 2007-2013. Priority Axis 2.
Author Disclosure Statement
No competing financial interests exist.