Screening of MRSA colonization in patients undergoing total joint arthroplasty

Abstract

Background:

Periprosthetic infection is commonly caused by Staphylococcus aureus and, if resistant to methicillin (MRSA), is associated with increase in severity and costs to patient and healthcare systems. MRSA colonizes 1–5% of the population, therefore using a screening and decolonisation protocol the risk of periprosthetic infection could be reduced. The objective of our study is to report the results of a preoperative MRSA screening and management protocol utilised at our hospital.

Methods:

All patients undergoing a total joint arthroplasty at our hospital were preoperatively screened for MRSA colonization with swab samples of five different locations. Exposure to risk factors were investigated in colonised patients and they were treated for 5 days prior surgery with nasal mupirocin, chlorhexidine sponges and oral tablets.

Results:

During the 48 months of the study, MRSA colonisation was identified in 22 (1.01%) of 2188 patients operated. The culture was positive only in the nasal swab in 55 patients. In five patients the nasal culture was negative, but they had another positive swab culture (three in the groin and two perianal). None of the patients reported a history of recent antibiotic treatment or hospitalization.

Conclusion:

At our institution, the prevalence of MRSA colonisation is 1.01% in patients undergoing hip and knee arthroplasty. Interestingly, our screening protocol included samples from five different anatomic locations, and it is important to highlight that we found patients with negative nares culture and positive cultures in other locations. Therefore, the number of carriers may be underdiagnosed if only nasal samples are obtained.

Level of evidence:

IV

Introduction

Periprosthetic infection is a serious complication that can lead to severe consequences for the patient and the healthcare system. Staphylococcus aureus is the most common infecting organism (Bengtsson et al, 1979) and, if resistant to methicillin (MRSA), is associated with a higher morbid-mortality, longer hospitalisation and increase in readmissions during the first postoperative year and global costs (McGarry et al, 2004; Melzer et al, 2003; Whitehouse et al, 2002). According to the European Centre for the Prevention and Disease Control (ECDC/EMEA, 2009) MRSA infections in Europe cause one million excess days of admission, 5400 deaths and €600 million of additional cost per year.

MRSA can survive up to 20 days on non-biological surfaces (Sexton et al, 2006) and it colonises 1–5% of the population (Gorwitz et al, 2008). We have not found any paper that confirms the prevalence of MRSA in Spanish orthopaedic patients. However, other studies reflect prevalence in the Spanish population with risk ranging from 0.3 to 9.7%, and they all agree that it has increased in recent years (Alvarenga et al, 2019; den Heijer et al, 2013; Lozano et al, 2011). The nares are the most frequent location (Wertheim et al, 2005) followed by the respiratory tract, perineum, groins, axilla, urinary tract and open wounds (Boyce, 2001). Although it has been classically found in chronic, hospitalised or institutionalised patients, it is increasingly encountered in community patients that do not share the former risk factors (DeLeo et al, 2010). Malcom et al (2016) concluded that congestive heart failure, chronic renal failure and previous inpatient admission in the previous 12 months are associated with MRSA colonisation. Similarly, colonisation has been directly related to the patient ASA (American Society of Anesthesiologist) classification (Moroski et al, 2015). MRSA colonization is considered an independent risk factor for periprosthetic infection (Kalmeijer et al, 2000). Patients that have S. aureus nasal colonisation also have an up to nine times increased probability of developing orthopaedic surgical-site infection (Kalmeijer et al, 2000). Baratz et al. (2015) failed to demonstrate a diminution of MRSA periprosthetic infection risk when performing MRSA screening and colonisation possibly because of the very low rate of infection, these numbers did not reach statistical significance. They found that, to reach an 80% statistical power, they would have needed 72,000 patients in each group. Chen et al (2013b) on the other hand concluded that using a decolonisation protocol the risk of periprosthetic infection is reduced. Sporer et al (2016) showed a reduction of periprosthetic infection risk from 1.11 to 0.34% (p < 0.05) by using screening and decolonisation. In addition, Kim et al (2010) found similar results with the infection rate decreasing from 0.19 to 0.06% (p < 0.003).

The mechanism by which MRSA colonisation leads to periprosthetic infection is not fully understood. It could be the result of transient bacteraemia after endotracheal intubation or simply direct contamination of the surgical wound from the body reservoir (Goyal el al, 2013; Valdés et al, 2008). MRSA virulence is due to its resistance to beta-lactamase and its ability to produce a glycocalyx. The resistance to beta-lactamase is caused by the bacterial acquisition of a genetic segment called “mecA” that encodes the penicillin-binding protein 2a (PBP2a) in the presence of which the bacteria continues to produce peptidoglycans, necessary for cell-wall synthesis, even in the presence of anti-cell-wall synthesis antibiotics (Kocsis et al, 2010; Marcotte and Trzeciak, 2008;Rozgonyi et al, 2007). In addition, once these bacteria are attached to the surface of the implant, they are able to synthesize a glycocalyx layer, which forms the biofilm that places them out of the reach of antibiotics and immune cells (Ando et al, 2004).

Recently, using molecular techniques, Skråmm et al (2014) showed that in colonised patients that have developed S. aureus surgical infections, an identical organism can be identified in the nares and in the surgical wound, suggesting a direct relation between colonisation and periprosthetic infection. Therefore, preoperative MRSA screening and decolonisation could result in a decrease in postoperative infections in orthopaedic surgery patients.

The objective of our study is to report the results of a preoperative MRSA screening and management protocol utilised at our hospital between 2015 and 2017 on all patients undergoing arthroplasty surgery.

Material and methods

From January 2015 until December 2018, 2205 total joint arthroplasties were performed at our hospital, a192 bed public hospital. Of these 17 confirmed infected revisions were excluded, so we studied 2188 arthroplasties, mostly primaries (1426 knee and 762 hip).

All patients were preoperatively screened for MRSA colonisation. A trained nurse obtained swab samples of five different locations (axilla, groin, umbilicus, perianal, and nasal). These samples were cultured at the microbiology laboratory in CHROM Agar MRSA plates at 37°C (98.6°F) for 48 h. The cost of this test was €25 (€20 for the plate and €1 for each swab). The results were considered positive if at least one of the five samples demonstrated the presence of MRSA and those patients were classified as colonised.

All colonized patients were interviewed to elicit exposure to risk factors for MRSA colonisation (institutionalisation or recent hospitalisation, recent antibiotic treatment (3 months before screening), immunosuppression or previous MRSA colonisation). All colonised patients received a decolonisation kit (price €6.96) containing five sponges of 4% chlorhexidine (Dispomedic^® Sponge C), a 2% nasal mupirocin tube (Bactroban^® nasal) and 15 tablets of 5 mg. chlorhexidine (Deratin^® 5 mg); the kit contains instructions and an information leaflet. Patients were asked to follow the treatment for the five days prior to the surgery. They were advised to wash the head and body with the sponges daily, to apply the intranasal ointment daily and take one tablet every 8 h. In addition, they received Vancomycin as perioperative prophylaxis (1 g 1 h prior to surgery and 1 g every 12 h postop for 24 h).

Results

MRSA colonisation was identified in 22 patients (1.01%) of the 2188 operated on during the 48 months of the study. In 15 patients, the culture was positive only in the nasal swab. In five patients the nasal culture was negative, but they had another positive swab culture (three in the groin and two perianal). None of the patients had a history of recent antibiotic treatment or hospitalisation. The majority of the patients (62%) were classified as ASA II, 22% were ASA III and the remaining (16%) were ASA I.

Discussion

MRSA colonisation varies with geographical location, the prevalence varying between 1.8 and 5% depending on the study (Table 1).

Table 1.

Methicillin-resistant S. aureus (MRSA) colonisation prevalence.

Study	Geography	Colonization prevalence (%)
Kim et al (2010)	East USA	4.4
Price et al (2008)	West USA	1.8
Iqbal et al (2017)	UK	1.8
Baratz et al (2015)	East USA	5
Chen et al (2013a)	East USA	4.6
Moroski et al (2015)	West USA	4.2
Sporer et al (2016)	East USA	2.9
Ahmad et al (2019)	UK	2.2
Valverde et al. (this study)	Spain	1.01

At our institution, a MRSA colonization prevalence of 1.01% was observed in patients undergoing hip and knee arthroplasty. Although most studies in the literature include only this population, Kim et al (2010) also included patients undergoing spinal surgery and arthroscopic procedure as long as they were admitted as inpatients for one night. In addition, Price et al (2008) and Ahmad et al (2019) included all patients admitted to their orthopaedic department. However, Iqbal et al (2017) encountered a 1.8% prevalence of MRSA colonization when including only patients treated for fractures.

Samples are usually obtained by rubbing swabs dry or soaked in sterile saline in both nares for 5 s (Jernigan and Kallen, 2010). Nares is the most common anatomical area screened since it is known to be the largest reservoir. However, sensitivity of a single anatomical area has been questioned. El-Bouri and El-Bouri (2013) compared the sensitivity of checking one sample from one location to a combination of two to three samples from seven different locations (axilla, hairline, nose, perineum, throat, groin and nose). Checking only the nose resulted in a 50.5% detection rate, whereas when studying three locations detection rate increased to 92%, thus concluding that screening three locations (groin, nose and throat) resulted in the greatest sensibility. Ahmad et al. (2019) obtained the same results. In our study, five different body locations (nose, groin, axilla, umbilicus and perianal area) were screened. To our knowledge, no such study has been performed previously. In our study, no axillary culture was found positive. The isolated study of the axilla has been previously found to have a low sensitivity by El-Bouri and El-Bouri (2013) and Ahmad et al (2019). Therefore, in our current practice we have stopped taking samples from the axilla and instead obtain them from the throat.

Microbiological processes differ among laboratories. Routine culture shows a 90% sensitivity and 100% specificity, while polymerase chain reaction (PCR) reaches a sensitivity of 100%, but specificity drops to 98%. The results of PCR are available in less than 24 h, while cultures can take up to 72 h (Paule et al, 2004). However, PCR is more expensive (€98.82) than classical cultures (€39.53) (Bühlmann et al, 2008). Although PCR is the gold standard (Weiser and Moucha, 2015), for reasons of cost we, like many others, have used the usual culture technique for this study (Chen et al, 2013a; Iqbal et al, 2017; Sporer et al, 2016).

Usual MRSA decolonisation protocols include the intranasal application of mupirocin 2% twice a day, and there is no consensus about adding chlorhexidine washings to this. Some authors (Weiser and Moucha, 2015) recommend adding body washes with 2–4% chlorhexidine soap for 5 days as well as the use of 2% chlorhexidine wipes. The chlorhexidine wipes have recently been introduced to facilitate administration and have the same effectiveness as the soap (Edmiston et al, 2007). However, Chen et al (2013b) in a systematic review found no additional benefit when adding chlorhexidine washes to the isolated use of intranasal mupirocin. The studies analysed in this review showed great heterogeneity and no definite conclusion could be reached. Malcolm et al (2016) used only intranasal mupirocin and chlorhexidine wipes preoperatively, but they do not report decolonisation success. When 85% of the patients fulfilled the treatment with mupirocin and washes for 5 days, Kim et al (2010), Moroski et al (2016) and Baratz et al (2013) reached a 68, 92 and 78% decolonization success respectively. Chen et al (2013a) obtained a 100% decolonisation success with a 97% compliance with the same treatment. We have been testing decolonisation success in our environment only since June 2016 and due to our low colonisation prevalence, we need further study to reach statistical significance.

Perioperative antibiotic prophylaxis with Vancomycin or Teicoplanin, when available, extending 24 h postoperatively (Hansen et al, 2014) is the gold standard in patients with confirmed MRSA and decolonisation has not been achieved or demonstrated. In addition, isolation measures during admission should be followed. When decolonisation has been demonstrated, the usual prophylaxis with cefazolin (Hansen et al, 2014) can be used. We use Vancomycin in MRSA colonised patients because at the time of surgery we do not yet have the results of the culture after the decolonisation treatment. Baratz et al (2015) add Vancomycin to Cefazolin for 24 h in colonised patients, and Iqbal et al (2017) use routinely Teicoplanin combined with Ciprofloxacin in these patients.

Lee et al (2010) evaluated different screenings and decolonisation programmes and showed cost effectiveness provided the prevalence of MRSA colonisation is at least 1%, and the success of decolonisation is at least 25%. Following these criteria our protocol would then be cost effective since our prevalence is 1.01% and our decolonisation success nears 89% (preliminary number – we need more cases).

Cost of periprosthetic infection being very high, any measures showing a reduction in the incidence of this devastating complication could lead to important savings for healthcare systems: Sculco (1993) and Slover et al (2011) estimated the cost of a periprosthetic infection to range between US$50,000 and US$75,000. Considering this, universal screening would be cost-effective if it saved a few infections. In our hospital, total costs of universal screening for the 48 months of the study (2188 patients) were estimated at €54,700. Decolonisation treatment cost for the 22 patients with MRSA colonisation was €153.56. In Spain, the mean cost of treatment of an early periprosthetic joint infection has been estimated at €19,270 increasing to €60,257 for late infections (Garrido-Gómez et al, 2013). It follows then, that our universal screening had been cost-effective if we had saved at least one of the 22 colonised patients from developing an infection of their arthroplasty.

There were some limitations in our study. This was not a randomized study and there was no comparison group, as all patients in our practice are routinely screened in five locations and decolonised with the same treatment for MRSA since it is our hospital’s routine standard of care. We detect the presence of MRSA by culture swab, and not by PCR which could increase sensitivity. There was also no control rate of MRSA that could be present due to laboratory contamination.

Conclusions

At our institution, the prevalence of MRSA colonisation is 1.01% in patients undergoing hip and knee arthroplasty. Interestingly, in our screening protocol we obtained samples from five different anatomic locations, which let us identify some patients with negative nares culture showing positive cultures in other locations. Therefore, the number of carriers may have been underdiagnosed if only nasal samples were obtained.

In this study, universal screening and decolonisation protocols appear to be a cost-effective measure in our environment, although further research is necessary concerning the reduction of infection risk.