Introduction
Worldwide, an increasing number of individuals have undergone joint-replacement surgeries, particularly at the hip and knee, either for elective reasons or following sustained trauma. Among all possible complications, prosthetic joint infection (PJI) is the one that is most feared; despite its low incidence, ranging from just 1 to 2% [
1,
2] for primary and up to 4% for revision surgeries [
3,
4], it boasts high morbidity and mortality rates. Gram-positive cocci (GPC), such as
Staphylococcus aureus and coagulase-negative
Staphylococci, are the major PJI-related microorganisms, followed by Gram-negative bacilli (GNB), with a prevalence ranging from 5 to 23% [
4‐
6]. In some case series, PJIs attributed to GNB have been reported at rates greater than 40% [
7,
8].
Acinetobacter is a genus of Gram-negative bacteria including a total of 31 different species. Due to its ability to spread in health care environments, Acinetobacter baumannii (Ab) is currently the most difficult species to control and eradicate [
9]. This microorganism is ubiquitous in the environment [
9] and has become one of the most successful pathogens associated with health care–related infections due to its ability to express a variety of antimicrobial resistance mechanisms and to form biofilms on both biotic and abiotic surfaces [
10]. On a global scale, approximately 50% of
Ab strains have been identified as multidrug-resistant (MDR). The World Health Organization (WHO) has declared carbapenem-resistant
Ab in particular to be one of the most important species among
Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Ab, Pseudomonas aeruginosa and
Enterobacter sp.; these organisms are referred to collectively as the ESKAPE group and are considered priority pathogens due to the threat they pose to global public health, necessitating urgent actions and the development of new antibiotics to combat them [
11,
12]. Unfortunately, in several Latin American countries,
Ab strains have shown resistance to virtually all classes of antibiotics, including carbapenems
. This worrisome microbial epidemiology has been identified at some Brazilian hospitals, where 77% of these isolates have exhibited resistance to carbapenems [
13]. The production of oxacillin-hydrolysing carbapenemase (carbapenem-hydrolysing class D enzymes) has been identified as the most common antibiotic resistance mechanism, and the global dissemination of OXA-type clones, including OXA-23, OXA-72 and OXA-58, is regarded as the most common mechanism of antibiotic resistance [
13,
14].
The emergence of musculoskeletal surgical site infections and orthopaedic implant–associated infections caused by
Ab has become a matter of urgent concern for health care providers due to the limited therapeutic arsenal available, particularly against carbapenem-resistant strains [
15]. Moreover, the treatment of PJIs caused by MDR and extensively drug-resistant (XDR) GNB, particularly
Ab, is hampered by its ability to be encased within biofilms. The resistance of
Ab against virtually all antimicrobials and its intrinsic capacity for biofilm formation may correlate with lower cure rates and increased disease morbidity since treatment usually requires a combination of highly toxic systemic antibiotics [
16,
17]. Despite this challenge, to our knowledge, no published studies have investigated independent risk factors (RFs) for
Ab-associated PJI. Indeed, few previous publications on
Ab-associated PJI have attempted to describe, in a case series report format, aspects of surgical and antibiotic therapy [
15,
18,
19]. Therefore, this study aimed to identify the independent RFs for
Ab-associated PJI and to assess the role of
Ab in treatment outcome.
Materials and methods
Study design
This study was performed as an observational, single-centre, retrospective, cohort study using data obtained from 2672 patients undergoing arthroplasties between January 2014 and July 2018, at a Brazilian orthopaedic referral centre. All patients diagnosed with PJI, either due to
Ab (
Ab-PJI) and other microorganisms (Non-
Ab-PJI), were identified from clinical and microbiological records and surgical description sheets. The primary study endpoint was the identification of independent predisposing factors associated with PJI caused by
Ab and secondary endpoint was to access if the
Ab-PJI have influence on treatment outcome. The study included individuals aged 18 years or older who met the diagnosis criteria for PJI according to the Musculoskeletal Infection Society (MSIS) [
20]. Inclusion criteria also required the same identified pathogen yielding in at least two peri-prosthetic tissue samples, and prospective follow-up period of a minimum one-year period. Patients who underwent arthroplasty at an institution other than ours, did not meet the criteria for PJI as defined by the MSIS or had culture-negative results were excluded. The study was reviewed and approved by the local ethics committee (approval no. 2,610,914 on April 20, 2018).
Definitions
The PJI onset date was defined according to the date of the first observation of typical infectious signs and symptoms. MDR-
Ab was defined as the nonsusceptibility of the identified pathogen to at least one antimicrobial agent from three or more different antimicrobial classes (e.g., aminoglycosides, cephalosporins with an anti-Pseudomonas effect, carbapenems, fluoroquinolones, penicillin + β-lactamase inhibitors, monobactams and polymyxin).
Ab that were extensively drug-resistant (XDR) to multiple antibiotics were defined as those lacking susceptibility to at least one antimicrobial agent from all but two classes of antimicrobials [
21].
Early-onset PJI was defined as those cases occurring < 3 months after the index surgery, whereas late PJI was defined as those cases in which the diagnosis occurred more than 3 months after the index surgery. The remission of infection was defined as the absence of clinical, laboratory, or radiological symptoms at the last medical follow-up (with a minimum follow-up time of 1 y). Therapeutic failure was defined as infection recurrence at a previously controlled site; requirement for new surgery, a second course of antimicrobial therapy, chronic antibiotic suppression, excision arthroplasty, or limb amputation; or death within the follow-up period [
22,
23].
Microbiological analysis
In the surgical ward, a minimal of three different periprosthetic tissue samples and synovial fluid were collected and processed for microbiology. Synovial fluid sample were aseptically inoculated into aerobic standard blood culture bottles. Tissue samples were homogenised in 3 ml of brain-heart infusion (BHI) broth for 1 min and inoculated onto aerobic sheep blood agar, chocolate agar, and anaerobic blood agar and into thioglycolate broth (BD Diagnostic Systems, Sparks, MD). The time limit for processing samples was 6 h. Aerobic were incubated aerobically at 35–37 °C in 5–7% CO
2 for 7 days, and anaerobic plates were incubated at 37 °C for 14 days. Additionally, 0.5 ml of tissue homogenate was inoculated in thioglycolate broth, incubated for 14 days, and sub-cultured on blood agar plates when the broth became cloudy. Colonies of microorganisms growing on plates were identified, and their susceptibilities to antibiotics were tested according to standard microbiological techniques. The bacteria were identified by conventional biochemical and metabolic tests in accordance with the international standards and definitions established by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [
24]
. Sensitivity tests were performed using the disk diffusion technique, and the determination of minimum inhibitory concentrations (MICs) was performed by automated means or by the e-test method, the results of which are presented according to standardised microbiological techniques.
Potential risk factors
Variables associated with the patient, surgery, and postoperative procedures were identified by reviewing the medical, intraoperative, and microbiological records to identify potential RFs for Ab-PJI. Demographic variables (sex and age), comorbidities (the presence and number of comorbidities, alcoholism, and smoking habits), the American Society of Anaesthesiologists (ASA) physical status classification, previous use of antibiotics during the past 3 m, and previous orthopaedic infections were assessed. Associated surgical aspects included the arthroplasty site (hip vs knee), total or partial arthroplasty, primary or revision surgery, and post-traumatic arthroplasty or elective arthroplasty. The factors related to the postoperative period that were considered included postoperative hematoma, the presence of sepsis at the time of diagnosis, concomitant infections diagnosed at different sites, and early or late infection. The surgical strategies, including debridement, antibiotics, and implant retention (DAIR) or any prosthesis removal (Non-DAIR), were assessed for survival and outcome analyses.
Statistical analysis
For the overall study population and the groups defined as Ab-PJI and Non-Ab-PJI, qualitative variables are reported as the mean and percentage, and quantitative variables are presented as the median and standard deviation (SD). Associations between qualitative variables were analysed using the Chi-square test and Fisher’s exact test, as indicated. The associations between quantitative variables were assessed by logistic regression. The risk estimate was calculated for the associated variables and reported as the odds ratio (OR) with a 95% confidence interval (CI). The logistic regression model was used to select significant variables from among those identified as significant in univariate analyses. Only variables with significance less than 0.20 (p < 0.20) were included in the logistic regression. Variables with significance less than 0.05 (p < 0.05) in the multiple regression were included in the final model. To estimate the probability of survival as a function of time, Kaplan-Meier (KM) analyses were performed, and the resulting curves were compared using the log-rank method. All data were analysed using SPSS, version 23 (IBM-SPSS Inc., Chicago, IL, USA).
Discussion
To our knowledge, this is the first study to investigate predisposing factors associated with
Ab-associated PJI. Interestingly the well-known predictors of PJI – which include revision surgeries, nonelective arthroplasties and late infections (PJI diagnosed after 3 m of index surgery) – were independently associated with
Ab infection. This likely reflects the particular epidemiology of a Brazilian orthopaedic referral centre, where the rates of nosocomial SSI caused by MDR-GNB are high [
25,
26]. In addition, the high selective pressure imposed by misuse of empirical broad-spectrum antibiotics is likely to have played a major role [
27]. Since then, a local antimicrobial stewardship program has been implemented as a tool to enhance the appropriateness of antibiotics prescriptions.
In the microbiological sample, a greater frequency (81.8%) of Ab strains causing PJI were XDR strains, whereas 12.1% were MDR strains and only 6.1% were sensitive to multiple antibiotics. The rate of susceptibility to carbapenems was worryingly low, with only 3% of cases sensitive to imipenem and only 6% sensitive to meropenem. The higher prevalence of Ab-associated PJI at our institution was not assumed to represent an outbreak but instead an endemic nosocomial pathogen typically identified in the intensive care unit environment that boasts an overwhelming ability to colonise the human skin. Furthermore, before 2018, immediate postoperative care protocols for patients who undergo arthroplasty in our hospital were usually enacted in the intensive care unit, which is likely to have increased the rate of skin colonisation by Ab strains.
Orthopaedic implant–associated infections have traditionally been considered difficult to treat due to the formation of bacterial biofilms on the implant surface and the low levels of antibiotic penetration into bone tissue and biofilms [
28,
29]. In addition, the higher levels of bacterial resistance commonly expressed by
Ab makes treatment of infections involving this organism even more challenging due to the scarcity of available drugs and the potential for antibiotic-related toxicity [
30]. Although
Ab is ubiquitous in nature and colonises the skin of healthy individuals, most human infections by this organism are health care–associated. A systematic review by Falagas et al. [
31], which included 55 articles describing
Ab infections, linked
Ab infections to prolonged hospital stays, intensive care unit treatments and the use of invasive devices.
Despite the poor availability of studies specifically describing
Ab-associated PJI,
the number of osteomyelitis and fracture-related infections caused by Ab seems to be on the rise worldwide, especially when considering those associated with high-kinetic energy trauma and open fractures [32].
Many studies have described a strong association between complex traumatic gunshot wounds resulting in fracture-related infections or osteomyelitis and Ab infection in various conflict-affected regions, such as Iraq, Afghanistan and Yemen [32‐
35]. However, whether
Ab is acquired during the act of injury itself from primary contamination or is acquired in the hospital during the trauma care and subsequent surgical procedures remains unclear. In addition to reports from Middle Eastern countries, a study by Vanegas et al. [
36] from Colombia addressed osteomyelitis, skin and soft tissue infections and reported an increased number of infections caused by
Ab, with a strong association with recent hospitalisation or surgery and previous use of antimicrobials in the past 6 m. Despite the increased risk of PJI following revision surgery [
37‐
39], the association between revision arthroplasty and
Ab-associated PJI has not yet been explored in the literature. However, we are aware that implant contamination during surgery is a primary source of infection, and patients previously colonised with
Ab may be at increased risk for PJI.
In our study, a 2.6-fold increase in the risk of
Ab-associated PJI was identified among patients undergoing an emergency arthroplasty, which suggests that, similar to in the case of fracture-related infections, posttraumatic arthroplasty may be a factor that predisposes patients to
Ab-associated PJI. The reasons underlying the association between trauma and
Ab infection were not elucidated in this study and require further investigation. In addition, we were unable to identify any independent associations between
Ab-associated PJI and closed proximal femoral fractures in the elderly population, recent hospitalisation history or recent use of antibiotics. However, nonelective arthroplasties may necessitate longer preoperative hospital stays due to the mandatory propaedeutic for assessing preoperative risks and the need to compensate for clinical comorbidities prior to performing the surgical procedure, which might increase the risk of colonisation by
Ab [
40]. However, the length of preoperative hospital stay, which could validate this hypothesis, was not a variable that was assessed in the present study.
In our study, late PJI was independently associated with
Ab infection, and a possible explanation may rely upon the lower level of virulence expression when bacteria express multiple antibiotic-resistant mechanisms. Several mechanisms associated with antimicrobial resistance, including pump efflux and biofilm organisation abilities, also reduce the bacterial replicative capacity [
41]. This stationary phase, associated with biofilm formation and maturation, is likely to cause
Ab infections to develop more slowly, increasing the likelihood of being diagnosed as late PJI [
42,
43]. Neither
Ab infection nor the surgical strategy used after the infectious diagnosis was independently associated with the final outcome or the risk of treatment failure. Few studies to date have assessed the prognostic factors associated with
Ab-associated PJI development, although some have reported high rates of therapeutic failure in
Ab-associated PJIs; for example, Vasso et al. [
19] described a 33.3% failure rate for the treatment of
Ab-associated PJI. However, their study was underpowered, assessing only nine patients in a group containing a mix of infections associated with both
Ab and
P. aeruginosa. Another study that assessed the outcomes of PJI caused by GNB -MDR reported that infections caused by MDR/XDR GNB were associated with high therapeutic failure rates when DAIR (52.2%) was performed as compared to when non-DAIR treatment strategies were employed (23.4%) [
44]. However, only three patients had
Ab-associated PJIs in this previous study, which is not a representative sample.
Importantly, the present study has potential limitations. First, it was performed as a retrospective investigation and took place at a single centre located in a major city in a developing country offering specialised orthopaedic care for the regional population. Consequently, the results obtained at our hospital may not apply to other hospitals. In addition, the identification and sensitivity tests were performed using nonautomated methods, and no molecular or genotypic analyses were performed to identify clonal variants or similar patterns of resistance mechanisms. Furthermore, no pairing was performed between the Ab-associated PJI and non–Ab-associated PJI groups to control for preoperative hospitalisation times or preoperative colonisation by Ab, which could support the hypothesis that PJI contamination occurred intraoperatively. However, this study explored the previously unexamined issue of PJI-predisposing factors and relied on the largest number of Ab infection cases described to date, with a high frequency of MDR/ XDR strains.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.