Trends in microbiological epidemiology of orthopedic infections: a large retrospective study from 2008 to 2021

This single-center retrospective study provides information about the changes in the prevalence of pathogens and the distribution of pathogens among different types of orthopedic infections. This study also highlights the antimicrobial susceptibility characteristics of the main pathogens in orthopedic infections over the past 14 years and thus provides a basis for refining the choices of empirical antimicrobial strategies.

In this study, S. aureus was the major pathogen causing orthopedic infections. It is well established that S. aureus is the most common pathogen in many orthopedic infections, accounting for 31.7%-43.6% [6, 14, 15]. We then analyzed the distribution of pathogens in different orthopedic infections. The diversity of implants, local microenvironment, and pathogen adaptation are responsible for the different proportions of pathogens in different orthopedic infections. Moreover, the local microenvironment may vary with the anatomical sites of surgery and the tissues involved, the physical and chemical characteristics of the implanted biomaterial, and the type of individual tissue response [16]. In the comparisons between Group A and Group B, we found that S. aureus and CoNS were significantly more likely to be associated with orthopedic implants. Implant-associated infections can lead to biofilm formation, and the results may be due to the fact that S. aureus and CoNS have multiple mechanisms for attachment and biofilm formation that contribute to their association with implant infections [17, 18]. Among non-implant-related infections, gram-negative bacilli (GNB) were predominant, which may be related to the fact that most of these patients with these infections suffered an open fracture after direct trauma and/or infection caused by soft tissue injury [19]. It has been reported that a high incidence of GNB, mainly Enterobacteriaceae, is common in developing countries [20, 21]. These results are consistent with the findings in studies by Montanaro et al. [6] and Arciola et al. [16]., who reported that the frequency of staphylococci appeared higher in medical device (MDS) infections and that GNB were more frequent among non-MDS-associated infections.

After comparing the differences between the two major groups, we further analyzed the subgroups. Regarding sex, men were engaged in more manual labor and physical activity than women, while the incidence of osteoarthritis was higher in women, which may account for orthopedic infections being predominantly identified in males except for arthroplasty-associated infections. Arthroplasty-associated infections may be an influential factor in the lower proportion of males with implant-associated infections. In this study, we found that S. aureus was the dominant pathogen in cases with internal fixation-associated infection, whereas CoNS was the major pathogen in PJI. This discrepancy is likely due to the differences in the interstitial milieu of implants. However, Rosteius et al. [7] analyzed 937 patients with hip and knee PJI between 2003 and 2011 and found that S. aureus was the predominant pathogenic type. Moreover, Triffault-Fillit et al. [22] counted the pathogen differences in 567 PJI cases according to the time of occurrence from prosthesis implantation and revealed the following: in early/delayed infections, S. aureus was the main pathogen; in late acute infections, Streptococcus spp. was the most common, and CoNS were the most frequent pathogens only in the late chronic types of infection. Different regions, periods, races and infection types may result in different pathogen prevalences. Furthermore, lower proportions of Pseudomonas spp., Enterobacter spp. and Acinetobacter spp. in the arthroplasty group were observed in our study. The characteristics were consistent with the results reported by Hu et al. [23] and Wang et al. [8]. PJIs caused by GNB were mostly acute hematogenous infections that originated from urinary tract infections [22], and the low incidence of the mechanism of acquisition could explain these results. Exposure to the environment is a common feature of external fixation and musculoskeletal trauma, which may be responsible for the larger proportions of GNB [19]. Overall, there were significant differences in the distribution of pathogens between different types of orthopedic infections. These differences can be used to guide the empirical use of antibiotics if the pathogens are unknown.

Regarding the trend of the pathogen distribution over the past 14 years, the proportion of S. aureus in Group A did not change significantly, whereas that in Group B showed a significant increasing trend. Therefore, the significantly increasing proportion of S. aureus in the total cohort was mainly influenced by the change in Group B. The proportion of Actinobacteria spp. and Enterobacter spp. decreased in the two major groups; these bacteria are the main hospital-associated pathogens. Strict aseptic operation, hand hygiene, and cleaning and disinfection procedures have contributed to this change. Moreover, we found a decreasing trend of CoNS and an increasing trend of S. aureus in the arthroplasty group over time, which could explain the decreasing proportion of CoNS in Group A. However, Wang et al. [8] observed a decreasing percentage of S. aureus and a rising percentage of CoNS in a retrospective study of 10,768 patients who received primary total knee arthroplasty (TKA) from 2002 to 2014. In a retrospective study of 2524 patients with PJI from 19 hospitals between 2003 and 2012 [9], it was found that the proportion of S. aureus and CoNS was consistently stable and did not change significantly. The differences may be caused by the different antimicrobial usage in different regions, there is also heterogeneity in the prevalence of pathogens according to geographical areas.

Polymicrobial infection is an important constituent of orthopedic infection, which cannot be ignored. The management of patients with polymicrobial orthopedic infections requires the administration of a broad-spectrum antibiotic or often multiple antibiotics to ensure adequate coverage against the infecting organisms. The treatment of these patients may be associated with a higher failure rate, mortality, and treatment costs than those of patients with monomicrobial infections [24]. Thus, we then analyzed mixed orthopedic infections, and S. aureus remained the most common pathogen. However, the polymicrobial infection group had a higher proportion of GNB (57.6% vs. 44.4%) and Enterococcus spp. (10.1% vs. 5.2%). Peel et al. [25] and Tan et al. [24] found a similar association between polymicrobial infection and infection with GNB and Enterococcus spp. Musculoskeletal trauma was the main component in Group B, and this mechanism results in direct contact with the external environment. Musculoskeletal trauma wounds can be easily contaminated by the external environment, which may account for the high polymicrobial infection rates of the no-implant group and musculoskeletal trauma group [26, 27]. Previous studies asserted that polymicrobial infection usually occurs in the early stage and that dehiscence after surgery is more common in polymicrobial than in monomicrobial infections [28]. Regarding the trend of the polymicrobial infection rate, the total cohort, Group B,and subgroup with internal fixation demonstrated a decreasing trend, which may be due to improved postoperative wound care and antibiotic prophylaxis.

In this retrospective study, we focused on the antimicrobial susceptibility of the main pathogens in orthopedic infections. We found that the percentage of infections caused by MRSA decreased over time. In a study of 191,460 S. aureus isolates collected at 427 centers [29], the overall MRSA rates declined between 2005 and 2016, which is consistent with the results of the current study. A decreasing trend was observed in the overall prevalence of MRSA infection in China [30], and a marked decrease in the prevalence of MRSA among Chinese hospitals within recent years has already been reported by the China Antimicrobial Surveillance Network (CHINET) [31]. Moreover, we found that the susceptibility to some antibiotics of MRSA increased over time. Dickstein et al. [32] and Diekema et al. [29] also found that the susceptibility of MRSA isolates to some older antibiotics was increasing, a possible result of the epidemic spread of MRSA clones that are more susceptible to these agents, whereas antibiotic susceptibility was found to be stable in a retrospective study by Zhang et al. [33] of 61,045 S. aureus isolates from 226 centers between 2012 and 2017. Ongoing surveillance and further research are required to detect future waves of MRSA epidemics and resistance in orthopedic infections. The analysis of S. aureus revealed that the proportion of MSSA was slightly higher than that of MRSA. However, among CoNS, the proportion of MRCoNS was significantly higher than that of MSCoNS. Tsai et al. [34] also reported these differences, which may be due to the difference in the adaptive capability to changing environmental conditions between staphylococci. Regarding CoNS, Pseudomonas spp. and Actinobacteria spp., the susceptibility of these isolates to some common antibiotics was found to be stable over time. However, Enterobacter spp. showed a decreasing trend of susceptibility to carbapenem antibiotics. This finding is consistent with the finding that carbapenem-resistant Enterobacteriaceae (CRE) infection is highly endemic in China [35]. Although CRE is not the major pathogen in orthopedic infections, this increasing trend still deserves increased concern.

This study has several limitations. First, we assessed the microbiological epidemiology of only culture-positive orthopedic infections. Culture-negative orthopedic infections are also an important part, and it has been reported that culture-negative orthopedic infections are caused mostly by fungi and Mycobacterium tuberculosis [36], which may have led to their low proportions of these two pathogens in our study. In addition, the data represent the experience from a single center, which exposes the results to a risk of local epidemiologic bias and thus may not be indicative of the experience of others.