Epidemiology of Prosthetic Joint Infection



Figure 2.1
Trends in hip (left) and knee (right) replacement surgery , 2000–2014, selected European countries (OECD Health Statistics) (From OECD/EU [3], with permission)



Most of the information about the incidence of prosthetic joint infection in the published literature has been hampered by methodological problems. These include a reliance on case series rather than well-designed cohort studies, the lack of explicit or standardized case definitions, incomplete case ascertainment, selection biases and, especially, differences in the length of follow-up [9]. Studies with longer follow-ups will report higher cumulative incidences (a percentage), even when the true incidence is low; failure to account for differences in length of follow-up between groups will lead to wrong conclusions [10]. Consequently, estimates of cumulative risk based on comparison should be made with caution unless the follow-up periods are the same. The denominator for the incidence rate is the prosthetic joint year.

The overall rate of prosthetic joint infections is highest in the first 2 years after surgery; the greatest risk of prosthetic joint infection occurs in the first 6 months after the operation and declines steadily after that [9, 11]; nevertheless, approximately 20–25% of all prosthetic joint infections occur after 2 years [11, 12]. According to a review by researchers from the Mayo Clinic in 2000, the combined incidence of total hip and knee arthroplasty infection was 5.9 (95% confidence interval [CI] 5.3–6.5) infections per 1000 joint years in the first 2 years following implantation and 2.3 (95% CI, 2.1–2.5) in years 2–10 [9]. The rates of late prosthetic joint infection (detected >2 years after the index operation) and of very late prosthetic joint infection (detected >5 years after the operation) for primary hip and knee replacements due to primary osteoarthritis performed between 1998 and 2009 were analysed from nationwide Finnish health registers [12]. The incidence rate of late prosthetic joint infection was 0.069% (95% CI, 0.061–0.078) per prosthesis year and was higher after knee replacement than after hip replacement (0.080 vs. 0.057). The rate of very late prosthetic joint infection was 0.051% (95% CI, 0.042–0.063) per prosthesis year, 0.058 for knees and 0.944 for hips [12]. The cumulative incidence of infection following primary hip or knee arthroplasty implanted between 1969 and 2007 was 0.5%, 0.8% and 1.4% at 1, 5 and 10 years, respectively, according to a population study in Olmsted County, Minnesota [13]. In the US Medicare population between 1997 and 2006, the incidence of prosthetic joint infection within 2 years of total knee arthroplasty was 1.55% and 0.46% between 2 and up to 10 years [11]. The higher early incidence of infection following implantation of a prosthesis followed by a decrease over time reflects the predominant effect of infection acquired during surgery, variable delays in symptom onset and diagnosis of infection after implantation and the decreasing susceptibility of prostheses to haematogenous seeding over time [9].

The establishment of national and international nosocomial surveillance networks has provided information on rates of surgical site infection after joint replacements. Fourteen European countries participate in the orthopaedic modules within the European Centre for Disease Prevention and Control surgical site infection surveillance network. The latest overall infection rates provided by the network for 2010–2011 were 0.7% for infection within 1 year of a knee replacement (95% confidence interval [CI] 0.7–0.8) (intercountry range 0.2–3.2%) and 1.0% for a hip replacement, including hip hemiarthroplasty (intercountry range 0.4–11.4%), with considerable variation in rates between countries [14]. A similar cumulative incidence of 0.9% (95% CI, 0.85–1.02) for prosthetic joint infections within 2 years of the index surgery was reported from Sweden’s hip arthroplasty register between 2005 and 2008 [15]. European infection rates for hip and knee arthroplasty within the first year of the index surgery are in line with estimates from the National Healthcare Safety Network of the Centers for Disease Control and Prevention in the USA: 0.9% for hip and 1.3% for knee prosthesis from 2006 to 2008 [16]. Data from the Healthcare Infection Surveillance Western Australia shows that the incidence of infection for hip and knee arthroplasty was 1.4 (95% CI 1.2–1.6) and 1.4 (95% CI 1.2–1.5), respectively, in 2008–2013 [17]. A surveillance study of surgical site infections in patients undergoing surgical procedures from 2005 to 2010, conducted in 30 countries across 4 continents (America, Asia, Africa and Europe), showed infection rates within 1 year of a hip and knee replacement of 2.6% and 1.6%, respectively [18]. A recent review of several national surveillance networks in Europe and the USA reported considerable differences in data collection methods and data quality, mainly in follow-up and post-discharge surveillance [19]. In that review, the cumulative incidence of prosthetic joint infection after total hip arthroplasty ranged from 1.3% to 2.9% and from 0.7% to 3.7% following total knee arthroplasty [19]. Conventional surgical site infection surveillance focuses largely on infections detected at the hospital where the operation was performed. Infections diagnosed and treated at other healthcare facilities may consequently be missed by conventional surveillance, which can lead to varying degrees of underestimation of the infection rate [20]. In a US study, 17% of infections would have been missed using operative hospital surveillance alone [20]. In that study, when infections diagnosed at other centres were included, the cumulative incidence of infection in the year following surgery was 2.3% for total hip arthroplasty and 2% for total knee arthroplasty [20]. Other sources of information on the incidence of prosthetic joint infections are the national arthroplasty registers, although it has also been found that the rate of prosthetic joint infection may be underestimated also in arthroplasty registers [21, 22]. An important weakness of the arthroplasty registers is that they are not designed for registration of infections. In the Nordic arthroplasty registers, the surgeon decides—based on a subjective assessment—whether or not the revision/reoperation is due to an infection. Positive cultures will not be available until 2–7 days after surgery, but once the revision diagnosis is reported to the register, it is probably never changed [22].

In most, but not all, of the studies, the rate of infection for total knee arthroplasty is higher than for hip arthroplasty [7, 9, 12, 16, 18]. The rate of surgical site infection is higher following hip hemiarthroplasty than total hip arthroplasty, ranging between 1.7% and 7.3% [23]. Hip hemiarthroplasty is the emergent surgical procedure indicated for displaced intracapsular femoral neck fractures, which are more frequent in the elderly population compared with total hip arthroplasty which is an elective procedure generally performed in younger people.

Some studies have reported an increasing incidence of prosthetic joint infections in hip and knee arthroplasties. Data extracted from the Nationwide Inpatient Sample in the USA showed that the incidence of prosthetic joint infection for hip arthroplasties increased from 1.99% (CI 95%, 1.78–2.21%) in 2001 to 2.18% (CI 95%, 1.97–2.39%) in 2009; after adjusting for other patient demographic factors, there was a significant year-to-year increase in the risk of hip infection over the study period [24]. The corresponding incidence rates for knee arthroplasties were 2.05% (CI 95%, 1.86–2.23%) in 2001 and 2.18% in 2009 (CI 95%, 1.99–2.37%). A more gradual but, nonetheless, significant increase in the risk of infection over time was found for knee compared to hip arthroplasty from 2001 to 2009 [24]. Similarly, the Nordic Arthroplasty Register Association (Denmark, Finland, Norway and Sweden) found an increase in cumulative 5-year revision rates due to infection in hip arthroplasties, rising from 0.46% (CI 95%, 0.42–0.50) during the period from 1995 to 1999 to 0.71% (CI 95%, 0.66–0.76) from 2005 to 2009 (6); the entire increase in risk of revision due to infection was within the first year of the primary surgery, and the risk increased in all four countries [25]. A population-based study, however, conducted in Olmsted County, Minnesota (USA), from 1969 to 2007, found no increase over the period of the study [13]. A recent study based on the Danish Hip Arthroplasty Register and several other national registers found that the relative risk of prosthetic joint infection in the year following primary total hip arthroplasty implantation did not increase over the 2005–2014 study period (the incidence was 0.53% [95% CI, 0.44–0.63] for 2005–2009 and 0.57% [95% CI, 0.49–0.67] for 2010–2014) [26]. In the study based on Finnish (nationwide) health registers of primary hip and knee replacements between 1998 and 2009, the incidence of late prosthetic joint infection (>2 years after the index operation) varied between 0.041% and 0.107% during the years of observation, with no temporal trend, while very late infections (>2 years after the index operation) increased significantly, from 0.026% in 2004 to 0.056% in 2010 [12]. By contrast, several national and international nosocomial surveillance networks have shown decreasing rates of surgical site infection after joint replacement in recent years [14, 17, 27, 28].

There is currently insufficient data to analyse the true incidence of arthroplasty infection at other anatomic locations, since the rates are based mainly on single-centre studies. After a shoulder arthroplasty, the rate of prosthetic infection appears to be similar to those for hip and knee prostheses, ranging from 0.98% to 1.3% in US series [29, 30]. The reported infection rate for elbow arthroplasty has been higher than for other joints: 3.3% in a systematic review [31]. The reasons for this may include an increased number of patients with rheumatoid arthritis (immunocompromised) receiving elbow arthroplasty and the fact that the elbow is a subcutaneous joint with a thin soft tissue envelope.

While it is still unclear whether the incidence of prosthetic joint infection is increasing, the absolute number of cases is growing. A further increase in the absolute number of arthroplasty infections is expected in the future, due to an increasing number of primary implantations carried out on a progressively elderly population with more associated comorbidities, the significant increase in the number of revision procedures, better methods of detection for the microbial biofilms involved in prosthetic joint infections, the increasing prevalence of microorganisms resistant to standard antibiotic prophylaxis and the accumulating number of arthroplasties that stay in place but remain at risk of infection during their implanted lifetime. Late infections occurring many years after prosthesis implantation may become more common, since the number of people living with some kind of joint arthroplasty is also increasing [12].



2.2 The Impact of Prosthetic Joint Infection


Prosthetic joint infections have a significant impact, not only on healthcare resources and economic costs but also on the morbidity, quality of life and mortality of patients. Research continues to quantify the impact of these infections.


2.2.1 Impact on Patient Mortality, Morbidity and Quality of Life


Prosthetic joint infection is widely depicted as a devastating complication with a potential impact on a patient’s mortality and quality of life . Nonetheless, there is very little information in the literature about the quantity and quality of life of these patients.

Berend et al. [32] studied 205 infected total hip arthroplasties treated with a two-stage reimplantation protocol and found that, in spite of the high degree of infection control, there was a 48% mortality rate over the study period (1996–2009). Choi et al. [33] performed the first analysis of septic and non-septic revisions by investigating mortality rates in 93 patients after revision total hip arthroplasty matched to 93 control subjects. They found that the mortality rate in the septic group was 33% (31/93) and 22% (20/93) in the aseptic group at the 5- and 6-year follow-up, respectively. Although this difference was not statistically significant, the septic patients were younger and died 6 years earlier. The same authors performed a similar study focusing on 88 infected total knee arthroplasties and 88 controls [34]. The overall mortality rate following revision total knee arthroplasty was 10.7% after a median follow-up of 4 years but was 6 times higher after septic revision (18%–16/88) than after aseptic revision (3%–3/88) [34]. In order to determine the effect of prosthetic joint infection on mortality, Zmistowski et al. [35] compared the outcomes of 436 infected revisions with 2342 patients undergoing revision arthroplasty for aseptic failure. Prosthetic joint infection was associated with a fivefold increase in mortality, even after controlling for other variables. Mortality in the prosthetic joint infection cohort was 3.7% at 90 days, 10.6% at 1 year and 25.9% at the 5-year follow-up. These figures compare unfavourably with some of the most commonly dreaded cancers, such as female breast and uterus and male prostate cancer [36]. Of course, the increased risk of mortality is not only due directly to the adverse effect of infection and treatment but also to the fact that prosthetic joint infection often reflects a poorer health status. Nevertheless, these figures should raise awareness of the systemic impact of disease among those doctors involved in the management of these infections and of the need to concentrate on two highly interconnected dimensions when dealing with prosthetic joint infections: infection eradication and general health status.

There is extensive evidence to show that a successful total joint arthroplasty greatly increases the patient’s quality of life in terms of function, pain and mobility, although surprisingly few studies in the literature about quality of life after prosthetic joint infections. Cahill et al. [37] were among the first to address this issue. They compared 62 uncomplicated total joint replacements and 34 cases of prosthetic joint infection, using a visual analogue scale for satisfaction, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Assessment of Quality of Life (AQoL) and the Short-Form 36 (SF-36). They found that infection reduced patient satisfaction and seriously impaired functional health status and health-related quality of life, but provided no information about the influence of infection control on health status. Helwig et al. [38] analysed 58 patients with prosthetic joint infection, applying the Short-Form Health Survey 12 (SF-12) to evaluate physical and mental status according to the outcome of infection, but did not differentiate between those treated with debridement and implant retention and one- or two-stage revision protocols. Surprisingly, when they compared successful and unsuccessful therapies, they found no significant differences on either scale, suggesting that even after a good clinical outcome, patients remained physically and mentally limited. As expected, when the infected cohort was compared with the general German population, they found that physical status was significantly disadvantaged.

There is some debate about whether one- or two-stage exchange is the best surgical option for chronic prosthetic joint infection. The impact on quality of life would be a major consideration for deciding which is the most appropriate option, although it is not possible to draw definitive conclusions from the available data. One study analysed functional outcome after revision surgery for prosthetic hip infection, and two-stage exchange was the first therapeutic option for more severe infections and patients with more comorbidities [39]. Studies that compare different surgical options with similar patients or in clinical trials should include functional and quality-of-life analysis. At the same time, there is also little information about the functional outcomes of patients treated with open debridement and implant retention, a common surgical strategy for treating acute prosthetic joint infection. Aboltins et al. [40] prospectively collected pre-and post-arthroplasty data of 2134 total joint arthroplasty patients, in which there were 41 prosthetic joint infections. The main conclusion was that prosthetic joint infection cases treated with debridement and implant retention had a similar improvement from pre-arthroplasty to 12 months post-arthroplasty as patients without prosthetic joint infection in terms of quality of life, according to the SF-12 survey. The analysis did not evaluate the potential influence of aetiological microorganisms. Núñez et al. [41] evaluated 24 patients with acute knee prosthetic joint infection who underwent debridement and implant retention and were in remission after 12 months of follow-up. Their health-related quality of life was analysed using WOMAC and SF-36 at baseline (before knee replacement) and at 12 and 24 months after antibiotic treatment ceased. There was a significant improvement in all items from baseline to 48 months after the operation, except for patients infected with Staphylococcus aureus who had significantly worse outcomes, most especially in terms of stiffness and function on the WOMAC index and the SF-36. Indeed, the only variables independently associated with worse outcomes were S. aureus, number of comorbidities and age. S. aureus is a virulent microorganism that causes severe soft tissue damage, which is potentially associated with a higher degree of fibrosis and scarring and would explain, at least in part, the worse functional outcomes.

Although more studies are needed to fully clarify the extent of the impact of prosthetic joint infection on a patient’s quality of life, there is no doubt that it should be a concern for professionals involved in the management of these patients. According to the literature, debridement and implant retention for early postoperative prosthetic joint infection is associated with a similar quality of life to subjects who do not have infection except those caused by S. aureus. Early diagnosis and treatment would probably improve the results in such cases. Meanwhile, revision surgery for infection has been clearly associated with a significant deleterious impact on health-related quality of life.


2.2.2 Economic Impact


Prosthetic joint infection management represents a substantial economic burden for hospitals, healthcare systems and patients alike. Infection is consistently one of the leading causes of total joint revision surgery [4245]. It is often the first or second most common indication for revision total knee arthroplasty [42, 44] and the third most common for revision total hip arthroplasty after aseptic loosening and dislocation [45]. It is also a leading cause of failure in other prostheses, specifically, shoulder, elbow and ankle prostheses [30].

The real cost of treating an infected joint is not easy to ascertain. It depends on many variables, such as the type of surgery, treatment, patient comorbidities and even bacterial factors, such as the antibiotic susceptibility profile. The full spectrum of economic impact includes not only the most commonly reported direct in-hospital costs but also direct outpatient costs (follow-up visits, rehabilitation, pharmacy), as well as indirect costs that are virtually impossible to gauge with any accuracy, such as loss of productivity or absence from work by the patient or his caregivers.

Kurtz et al. [24] included over 150,000 prosthetic joint infection cases and found that the average total hospital costs for infected hip revision were $72,700 US dollars (USD) in 2001, and $93,600 USD in 2009. The average charges for infected knee revision were $58,700 USD in 2001 and $74,900 USD in 2009. More recent studies from the USA included not only inpatient costs but also the cost of outpatient services. In 2014, Kapadia et al. [46] identified 21 infected total knee arthroplasties and matched them to 21 non-infected subjects who underwent uncomplicated primary surgery. Patients with prosthetic joint infection had significantly longer hospital stays, more readmissions and more clinic visits. The mean total episode cost (fixed and variable direct costs) for patients with surgical site infections was $116,383 USD (range, $44,416–$269,914), which was significantly higher than the mean $28,249 USD (range, $20,454–$47,957) in the matched group. Just recently, the same authors studied 16 consecutive infected total hip arthroplasties matched to 32 non-infected patients [47]. As before, the mean cost per episode was significantly higher in the infected group, $88,623 USD (range, $44,043–$158,202) than in the matched cohort, $25,659 USD (range, $13,595–$48,631).

The specific cost varies sharply from one setting to another, depending on the type of healthcare system and the corresponding economic standard. Fernandez-Fairen et al. [48] performed a systematic review of the literature and found significant disparities in absolute values between publications, depending on the country of origin. Nonetheless, the cost of a septic revision was consistently around 2–4 times more expensive than primary surgery and 1.5–3 times more expensive than aseptic revision surgery.

Cases of early postoperative and haematogenous prosthetic joint infection can be treated effectively with debridement and implant retention. The specific objective of an Australian study by Peel et al. [49] was to calculate the cost associated with this strategy. They focused on 21 prosthetic joint infections (12 total hip arthroplasties, 9 total knee arthroplasties) matched to 42 control patients with uneventful primary joint replacements. They included inpatient and also outpatient expenses, including readmissions, follow-up medical and nursing visits, medical imaging, pathology and pharmacy, including dispensed antibiotics. The total cost for patients with infection was $69,414 Australian dollars (AUD), compared with $22,085 AUD for the controls, with significant differences across almost all areas of patient care. The cost of an infection including the index operation and the costs of prosthetic joint infection management was 3.1 times that of an uneventful primary arthroplasty.

In summary, the economic cost of treating a case of prosthetic joint infection is two to four times higher than a primary replacement or aseptic revision.

More studies are needed that not only include a descriptive cost analysis but also take into account outcome measures, such as successful infection eradication, functional results and quality-of-life measurements after treatment. These cost benefit analyses will allow for more informed decision-making in all fields of prosthetic joint infection management. Prevention remains the best way to avoid the dire health-related and economic consequences of infection. Advances in the prevention of prosthetic joint infection will be needed to make an impact on the anticipated increase in the number of infections in the years to come.


2.3 Risk Factors


Several risk factors for the development of a prosthetic joint infection have been described, mainly derived from patients with total hip and knee arthroplasties. Some of the proposed risk factors should however be interpreted with caution because different studies used diverse methods or employed different classifications/scoring systems or focused on only one particular anatomic site [50]. Some studies have suggested that the risk factors could vary according to the particular anatomic joint [51]. Identifying the current risk factors that predispose patients to prosthetic joint infections after an arthroplasty will help the clinician establish strategies to prevent them. Risk factors for acquiring prosthetic joint infections can be categorized as patient characteristics, perioperative related factors and risk during bacteraemia [50].


2.3.1 Host Risk Factors


Nutritional status (mainly obesity), diabetes mellitus, rheumatic diseases and immunosuppressive therapy are among the most frequently reported risk factors for developing prosthetic joint infections [50, 52]. Smoking, coagulopathy, preoperative anaemia and previous joint surgery, mainly previous arthroplasty, have also been described as risk factors for prosthetic joint infection.

Obesity, defined as weight >20% above the ideal body weight or increased body mass index (BMI), has been associated with and increased risk of infection in several studies [52, 53]. Possible reasons include prolonged operative duration, increased allogenic blood transfusions and the presence of other comorbidities [54, 55]; however, obesity has remained an independent risk factor after adjustment for other covariates in some investigations [55]. In their study, Peel et al. described that for every 1 kg/m2 increase in BMI, there was an associated 10% increase in the risk of prosthetic hip infection [51]. It was suggested that the association between obesity and hip arthroplasty could reflect an increase in postoperative dead space and the excellent medium for microbiological growth provided by necrotic fat [51]. Conversely, suboptimal preoperative nutrition with BMI <25 or malnutrition with serum albumin <34 g/L has also been associated with an increased risk of prosthetic joint infection [52, 54, 56, 57].

Diabetes mellitus is a risk factor for infection after general surgical and arthroplasty procedures [52, 58]. Moreover, Mraovic et al. observed that postoperative morning hyperglycaemia (blood glucose > 200 mg/dL) increased the risk of surgical site infection, even in patients without diabetes [59].

Patients with rheumatoid arthritis are at higher risk for developing prosthetic joint infections, with a relative risk increased approximately two- to fourfold, compared with that of patients without rheumatoid arthritis [9, 52, 60, 61]. This risk increases further in the context of revision arthroplasty or when there has been a previous prosthetic joint infection. It is often difficult to separate the relative contribution of the underlying illness, the accompanying comorbid conditions and the therapy with immunosuppressive or immunomodulating agents used with the patients [55]. New treatment approaches for patients with rheumatoid arthritis, including earlier use of disease-modifying antirheumatic drugs and the advent of biologic drugs, such as antitumour necrosis factor inhibitors, may have significantly increased the risk of prosthetic joint infections in these patients [61]. Until further data is available, the discontinuation of chronic treatment with these drugs should be assessed on a case-by-case basis before undergoing elective orthopaedic surgery. In their study, Peel et al. demonstrated that systemic corticosteroid therapy remained a predictor of infection when controlling for underlying comorbidity; this association may be mediated, at least in part, by impaired wound healing [51].

A systematic review found that smokers were substantially more likely to have postoperative complications following total knee or hip arthroplasty [62]. Current smokers undergoing knee or hip replacement had more often surgical site infections than never smokers. This is likely related to the negative effects associated with vasoconstriction on surgical wound healing [57].

Greenky et al. analysed 15,222 patients who underwent total joint arthroplasties from January 2000 to June 2007. A percentage of 19.6% presented with preoperative anaemia; prosthetic joint infection occurred more frequently in anaemic patients at an incidence of 4.3% compared with 2% in nonanaemic patients (P < 0.01) [63]. The multivariate model confirmed the risk of prosthetic joint infections to be two times higher in anaemic patients vs. nonanaemic patients.

A mean international normalized ratio (INR) greater than 1.5 has been found to be more prevalent in patients who developed postoperative wound complications (such as haematoma) and later prosthetic joint infections [64].

A history of prior arthroplasty on the index joint has consistently been recognized as a risk factor for prosthetic joint infection, increasing the risk of infection by up to eight times compared with patients with primary implantation [56, 60, 65]. The risk increases with the number of previous joint arthroplasties [60]. Prolonged operating times during revision surgery, the presence of unrecognized infection at the time of revision and abnormal surrounding soft tissue could be contributing factors.

S. aureus colonization increases the risk of S. aureus surgical site infections. The risk for these infections may be decreased by screening patients for nasal carriage of S. aureus and decolonizing carriers during the preoperative period [66, 67].

The presence of distant infections previous to the joint replacement has also been related to a higher risk of prosthetic joint infections, presumably due to transient bacteraemia from a distant infection site during this high-risk period [55]. Therefore, it has been recommended to screen for the presence of active infection elsewhere (such as urinary tract infection, respiratory tract infection, active skin infection, abscess or infected ulceration) prior to an elective prosthetic replacement. If asymptomatic pyuria or bacteriuria is associated with the development of prosthetic joint infections is not completely clear [6870].


2.3.2 Perioperative Factors


It is considered that prosthetic joint infection is frequently acquired in the operating room during the arthroplasty procedure [57]. During arthroplasty, 50–67% of surgical gloves are estimated to be perforated, which is associated with increased infection rates. Handwashing, double gloving and changing gloves at regular intervals during the operation may be preventive strategies [7173]. Human traffic in the operating room is associated with increased bacterial air counts, while opening and closing the theatre door disrupts the airflow around the patient, allowing microorganisms to enter the airspace around the surgical site [74]. Hypothermia could also facilitate prosthetic joint infection by inducing peripheral vasoconstriction with a substantial reduction of subcutaneous oxygen tension and directly inhibiting the inflammatory response [75]. The use of alcohol-based antiseptic skin preparations, combined with povidone or chlorhexidine, in the operating room [76], skin drapes and clipping hair immediately before surgery rather than the night before are associated with a reduced risk of surgical site infections [77]. A timely and appropriate perioperative antibiotic, according to the current guidelines for antimicrobial prophylaxis, is one of the most effective agents for the prevention of prosthetic joint infection [78]. It is recommended to administer the antimicrobial 1 h prior to surgical incision, with a repeat dose if the operation extends beyond 2 or 3 h, or if there is substantial blood loss [75, 77, 79].

Another independent risk factor predictive of prosthetic joint infection is the duration of the operation. The risk increases significantly when a procedure lasts more than 120 min, which is a reflection of more complex surgery, with prolonged surgical exposure and tissue damage during the procedure [80].

In essence, all wound complications (such as delayed healing, drainage or persistent dehiscence, haematoma, seroma) increase the risk of infection. Several authors have shown that developing a superficial surgical site infection not involving the prosthesis is a significant risk factor for prosthetic joint infection [60, 65, 81]. In Berbari et al.’s study, surgical site infection correlated with a 35.9-fold increase in the risk of infection in multivariate analysis [60]. Several authors have described local haematoma and wound discharge as risk factors for infection [51, 53, 56, 60]. One case-control study with 63 cases observed that drainage tube implants reduced the risk of subsequent prosthetic knee infections. However, previous studies have found that drainage tubes reduce haematoma formation; they have not shown a reduction in infection [51].

There are risk factors for prosthetic joint infections not primarily associated with the surgical procedure or wound healing. These include developing postoperative atrial fibrillation and myocardial infarction as independent risk factors. One plausible explanation is that all patients with serious cardiac complications receive aggressive anticoagulation with heparin or similar, which has been reported to be an independent risk factor for the development of prosthetic joint infection [53, 64]; also, the patients are generally older and sicker with pre-existing medical conditions that delay wound healing [53].

An association has been reported between allogeneic blood transfusion and infection related to the immunomodulation effect of the transfusion [82]; these patients are 2.1 times more likely to develop prosthetic joint infections, compared to those who do not receive a transfusion [53].

Longer hospital stays are another adjusted independent risk factor for infection. These patients are more likely to be exposed to nosocomial organisms that can lead later to prosthetic joint infection [53]. For this reason, it is important to avoid unnecessary stays in hospital before elective joint implantation.


2.3.3 Risk of Haematogenous Prosthetic Joint Infection


Finally, it is important to remark that an arthroplasty implant is at risk of infection not only in the immediate postoperative period but during their implanted lifetime due to the risk of bacteraemia. Nevertheless, the incidence of haematogenous seeding to a joint from a remote infection is low (0.1%) [83]. The situation is different in the case of S. aureus, where the rate of prosthetic joint infection after S. aureus bacteraemia is approximately 35% [84]. This means that if bacteraemia (mainly due to S. aureus) occurs, patients with uninfected prosthetic joints should be carefully monitored clinically for the development of prosthetic joint infection. In this situation, early diagnosis may avoid exchange of the prosthesis since infection can be cured with debridement and implant retention.

Along the same lines, the question of whether dental procedures alter the risk of prosthetic hip or knee infection has been actively debated in the last few decades. Recent case-control and cohort studies have finally concluded that the risk of infection in patients with prosthetic joints does not increase after dental procedures and specific antibiotic prophylaxis is not required [85, 86].


2.3.4 Risk Scores


The American Society of Anaesthesiologists (ASA) score is a widely used grading system for preoperative health of the surgical patients based in five classes. The ASA score has been associated with an increased risk of prosthetic joint infection in several studies [5153].

The National Nosocomial Infections Surveillance (NNIS) System surgical score for identifying patients at a high risk of postoperative surgical site infection includes the ASA preoperative assessment score, the duration of the surgical procedure and surgical wound classification of each procedure (classification degree of microbial contamination of surgical wound at time of operation) [87]. The NNIS has been shown to be a better predictor of surgical site infection than individual components of the index. In one large case-control study, an NNIS score ≥1 was a significant risk factor for the development of prosthetic joint infections and an NNIS score of 2 correlated with a 5.2-fold increase in the probability of infection [60]. These findings remained in the multivariate analysis.

Two proposed Mayo prosthetic joint infection risk score models were developed using data from 339 cases and 339 controls of patients undergoing total hip or knee arthroplasty in the same period at a tertiary referral hospital; risk factors were detected using multivariable modelling [56]. The baseline Mayo prosthetic joint infection risk score included BMI (either high or low), a previous operation on the index joint, prior arthroplasty, immunosuppression, ASA score and procedure duration. This score has the potential to help identify high-risk individuals at the time of surgery. The 1-month-postsurgery score for risk of prosthetic joint infection contained the same variables, as well as postoperative wound drainage. The last score can be used in the postoperative period as an early workup in patients with early signs or symptoms suggestive of prosthetic joint infection. The two risk score models require external validation before they can be implemented in clinical practice [56].


2.4 Classification Schemes


Several classifications have been proposed for prosthetic joint infections. Their objective is to guide medical and surgical decisions in patients with prosthetic joint infections.

The Zimmerli classification divides prosthetic joint infections into three categories based on time to infection: early-, delayed- and late-onset infections. Early-onset infection occurs in the 3 months following arthroplasty. The microorganisms involved are usually more virulent and are inoculated into the surgical site during implant surgery. In this classification, delayed-onset infection occurs after 3 months and before 12 or 24 months. This type is usually caused by less virulent microorganisms that contaminate the surgical site during arthroplasty. Late infections occur between 1 and 2 years after arthroplasty and are considered to be mainly haematogenous in origin, although some are also caused by slow-growing bacteria acquired during the index surgery [88].

This classification is somewhat similar to an older one by Coventry et al., who defined three stages of prosthetic joint infection. Stage I is acute infection in the first 3 months after surgery; stage II is delayed infection occurring between 3 months and 2 years after arthroplasty and constant chronic pain after the operation; stage III is a haematogenous infection with a previously pain-free period [89].

Another important classification that is frequently used is the Tsukayama classification , which proposes four types of prosthetic joint infection [90]. Early postoperative infection occurs in the first month after arthroplasty. Late chronic infection occurs after this time and is generally associated with a more protracted clinical course. The third type is acute haematogenous infection , which is a late infection with a long, previously asymptomatic period and usually follows a more acute clinical course. The fourth type of prosthetic joint infection is a positive intraoperative culture, found in patients undergoing revision arthroplasty for presumed aseptic failure. This latter category and late chronic infection represent the same clinical scenario: a loosened prosthesis inserted months or years previously, although with the difference that the new prosthesis has already replaced the infected one at the time of diagnosis in the positive intraoperative culture category.

The Zimmerli and Tsukayama classifications are the most frequently used and are similar from a practical point of view. Except for the timing, a positive intraoperative culture and late chronic infection are equivalent to delayed infection in the Zimmerli classification. Tsukayama’s haematogenous category is defined in the same way as Zimmerli’s late category, except for the time limit, set at 2 years. In summary, early postoperative infections and haematogenous infections (Zimmerli’s late type) can be regarded as acute infections, whereas Tsukayama’s late chronic and Zimmerli’s delayed prosthetic joint infections correspond to chronic infections.

Most teams use these classifications for deciding how to manage prosthetic joint infections. Patients with early postoperative and haematogenous infections are candidates for debridement and implant retention with prolonged antimicrobial therapy in an attempt to cure the infection without removing the implant. Late chronic and delayed infections are frequently managed with a two-stage implant exchange. This approach is based on the possibility of curing acute infections while biofilm is still immature and the difficulty of treating chronic infections with mature biofilms without removing the implant [91].

Other authors have added some useful considerations to prosthetic joint infection classifications. According to Garvin and Hanssen, acute prosthetic joint infections occur in the 4 weeks following surgery and late chronic prosthetic joint infections 4 weeks after surgery with an insidious clinical onset [92]. This insidious clinical onset is useful in daily practice for distinguishing between late chronic and late (acute) haematogenous prosthetic joint infections. Senneville et al. mainly considered the duration of the symptoms and attached less importance to timing with respect to surgery. Their classification proposes acute infection as one with <1 month of symptoms. Other infections with symptoms duration of >1 month are late infections [93].

As some studies on the subject of prosthesis retention have suggested, successful management of prosthetic joint infections depends on factors other than the time when infection occurs [94, 95]. Hence, factors such as the condition of the host, the appearance of the soft tissue around the prosthesis and the virulence of the microorganism causing the infection should be taken into account when deciding on the therapeutic attitude. McPherson et al. categorized prosthetic joint infections not only in terms of timing but also in terms of the systemic status of the host [96]. Their classification included early postoperative infection (stage I), haematogenous infection (stage II) and late chronic infection (stage III). Late chronic infection was considered when symptoms arose 4 weeks or more after the index arthroplasty. Patients were classified, on the basis of age, the presence of neutropenia and low CD4 cell count, as non-compromised (A), compromised (B) or significantly compromised (C). The infection site was also graded according to the presence of chronic infection, fistula, tissue loss and similar factors, as grade 1, 2 or 3 (uncompromised, compromised and significantly compromised, respectively). The use of this classification for the management of prosthetic joint infections has yielded different results in prognosis [96, 97].


2.4.1 Microbial Aetiology


Much of our current understanding of the microbial aetiology of prosthetic joint infections comes from studies that have limitations due to small sample size, single-centre experiences, lack of uniform or standardized definitions of infection and a variety of selection biases [9, 98]. Most studies have focused on specific categories of infection (mainly early-onset or chronic infections) or on infections treated with particular surgical strategies (largely debridement with retention of the prosthesis or two-stage exchange). Few studies have systematically described the full microbial aetiology of these infections [98].

A wide range of bacterial and fungal microorganisms can cause prosthetic joint infection (see Table 2.1). Aerobic Gram-positive cocci are the most common group of causative microorganisms (65–78%) [9, 55, 98], driven largely by infection with staphylococci, both coagulase-negative staphylococci and S. aureus, which account for 50–65% of all infections [55, 98]. S. aureus is a virulent pathogen and prosthetic joint infection by S. aureus typically presents with acute infection, although chronic infections have also been reported [99]. The group of microorganisms referred to as coagulase-negative staphylococci includes many species, with Staphylococcus epidermidis being the most frequently identified member of this group. Many are ubiquitous members of the skin microbiota. This group of organisms is the most frequent cause of chronic infection. Since much of the literature on prosthetic joint infection tends not to refer to individual species, the role of different species is unclear. Streptococcus and Enterococcus species were involved in 9% and 8% of cases, respectively, in a recent large multicentre study [98]. Among the streptococci, Streptococcus agalactiae was the species most often isolated. More than 50% of Enterococcus species were involved in infections occurring in the first 90 days after prosthesis implantation, and more than 50% of cases were polymicrobial infections [100]. Enterococcus faecalis was isolated in more than 85% of enterococcal infections [98, 100]. In most past series, aerobic Gram-negative bacilli were implicated in less than 10% of cases of prosthetic joint infections [9, 55, 88]. Recently, however, studies in different geographical areas have reported higher frequencies of these pathogens, ranging from 17% to 42% [98, 101106]. The percentage of Enterobacteriaceae appears to be increasing, and the most common species isolated (in descending order) are Escherichia coli, Proteus spp., Enterobacter spp. and Klebsiella spp. [98, 105]. Anaerobic bacteria were involved in 7% of all cases of prosthetic joint infection in the recent large multicentre study referred to above, with Cutibacterium spp. (formerly Propionibacterium spp.) being the most commonly identified anaerobic bacterium (5% of all infections) [98]. Cutibacterium acnes is a low-virulence microorganism, generally found in the skin microbiota and sebaceous glands. This microorganism can be inoculated at the time of surgery, but most infections have an indolent clinical course and are usually diagnosed months after prosthesis implantation. Less frequently isolated in prosthetic joint infections are aerobic Gram-positive bacilli, such as Corynebacterium spp. (2%), fungi (1%) and Mycobacterium spp. (<1%) [98]. Even though fungi are not commonly involved in prosthetic joint infections, the proportion of infections has significantly increased in recent years [98]. In decreasing order, the following species are involved in more than 80% of all prosthetic joint infections: S. aureus, S. epidermidis, E. coli, Pseudomonas aeruginosa, E. faecalis and P. acnes (it should be remembered that coagulase-negative staphylococci are not often identified to the species level, so that S. epidermidis may be the most common species) [98].
Feb 8, 2018 | Posted by in ORTHOPEDIC | Comments Off on Epidemiology of Prosthetic Joint Infection

Full access? Get Clinical Tree

Get Clinical Tree app for offline access