is there any strategies that the authors implemented to deal with confounding variables?

Statistical analyses consisted of the ² test, Fisher exact test, or Pearson correlation coefficient performed using IBM SPSS version 22.0 (IBM-SPSS Inc, Armonk, NY) and Epi-info (Centers for Disease Control and Prevention, Atlanta, GA). A P value<.05 was considered to indicate statistical significance.

The purpose of active surveillance cultures for the detection of MRSA is to prevent SSIs by eradicating MRSA carriage and to manage MRSA carriers with contact precautions. During period A, preoperative MRSA active surveillance cultures from the nasal cavity were obtained from all patients. MRSA carriers underwent decolonization by mupirocin to prevent MRSA SSIs. Contact precautions were performed only for patients with MRSA infections. Despite the use of this prevention protocol, there was a high incidence of MRSA SSIs among MRSA carriers and noncarriers. A previous study¹² reported that MRSA carriers frequently contaminate their environment with these bacteria and a program of universal screening, contact precautions, and hand hygiene was associated with a decrease in health care-associated transmissions of MRSA.¹⁰ Therefore, the additional prevention protocol implemented during period B included contact precautions for all MRSA-positive patients and CEZ-based antimicrobial stewardship. This protocol mandated the increased use of alcohol-based hand hygiene solution. Additionally, because of the increased frequency of administration of CEZ per day at operative and postoperative days to almost all patients, the AUD of CEZ increased as well. The period B protocol also reduced the duration of AMP use. Its use decreased the incidence of MRSA SSIs, especially in MRSA noncarriers, despite an increase in MRSA carriers. These results suggest that hand hygiene; AMP stewardship; and contact precautions for MRSA-positive patients, including MRSA carriers, are an important measure to prevent MRSA SSIs in patients undergoing orthopedic surgery.

Previous studies have reported the effectiveness10, 13 but also the ineffectiveness14, 15 of active surveillance culture for MRSA. The effectiveness of these interventions seems to strongly depend on the prevalence of MRSA, compliance with general infection control measures, and the settings in which those measures are implemented.¹³ In Japan, the prevalence of MRSA is high and community-acquired MRSA has spread to outpatients.¹⁶ We previously reported that among MRSA strains isolated from orthopedic patients the incidence of community-acquired MRSA carrying SCCmec type IV was 27.5%.¹⁷ In the present study, there was a higher incidence of MRSA carriers during period B than during period A. Thus, active surveillance culture is an important method for recognizing patients hospitalized on wards with a high risk of MRSA SSI who thus require contact precautions and decolonization. We performed active surveillance culture from the nasal cavity because nasal MRSA carriage is a risk factor for MRSA SSIs.⁸ However, the sensitivity of nares screening only is reported to be 60%-70%¹⁸; therefore, the number of cases of colonization may be underestimated.

A meta-analysis¹⁹ demonstrated that MRSA isolation is significantly associated with previous CEZ use. However, in our study there was a significant negative correlation between the AUD of CEZ and nosocomial MRSA SSIs. In a rat model of intestinal MRSA carriage, CEZ had less influence on the intestinal flora, without facilitating MRSA proliferation, than FMOX or cefmetazole.²⁰ These results and those of our study suggest that CEZ prophylaxis does not increase the selective pressure on MRSA.

Prolonged AMP for >48 hours was a risk factor for MRSA SSI during periods A and B. The increased consumption of CEZ during period B reflected the frequency rather than the duration of its administration. Previous studies reported that prolonged AMP after surgery increases both the isolation of Enterobacteriaceae and enterococci with acquired resistance.²¹ The Society for Healthcare Epidemiology of America/Infectious Disease Society for America Strategy to Prevent Surgical Site Infections in Acute Care Hospitals recommended the discontinuation of AMP within 24hours after surgery.²² Prolonged AMP use should therefore be avoided.

The incidence of MRSA SSIs among MRSA carriers was high during periods A and B. In previous studies, topical decolonization of both nasal passages and skin sites using chlorhexidine bathing reduced the incidence of SSIs caused by S aureus,²³ and VCM prophylaxis reduced the incidence of those caused by MRSA.²⁴ Although all MRSA carriers were decolonized using mupirocin, VCM was administered only to a small proportion of them. None of the MRSA carriers who received both mupirocin decolonization and VCM prophylaxis developed an MRSA SSI. A prevention protocol for MRSA carriers that includes chlorhexidine bathing and VCM prophylaxis may therefore also be effective.

A limitation of our study was its retrospective design and the participation of a single institution. In this before-after study, several precaution methods were introduced at the same time and several data, including cefazolin AUD and hand hygiene, changed similarly after implementation. Therefore, these results do not allow us to rank how the approaches contributed to the reduction of MRSA infection rate. A previous report suggested that care bundles consisting of S aureus screening, decolonization, and AMP stewardship reduced the incidence of S aureus SSI in patients undergoing cardiac, hip, or knee surgery and adherence to the bundle was important.²⁵ However, we could not determine the improvement of hand hygiene measures and the compliance with contact precaution measures before and after the implementation, which can be possible biases that limit the interpretation of the results. Although our findings showed that CEZ AUD was negatively correlated with the rate of MRSA SSIs and suggested that prolonged (>48 hours) AMP use was a risk factor for MRSA SSIs, there might be several other important risk factors for MRSA SSIs such as American Society of Anesthesiologists physical status or comorbidity. Their identification needs further analysis, including multivariate analysis.