A Novel Protocol for Nasal Decolonization Using Prolonged Application of an Alcohol-Based Nasal Antiseptic Reduces Surgical Site Infections in Total Joint Arthroplasty Patients: A Retrospective Cohort Study

Patients and Methods

This study received approval by our Institutional Review Board. Because of the success of a pilot study, standard of care at our institution during total joint arthroplasty was changed in 2016 to include alcohol-based nasal decolonization. Therefore, primary total joint arthroplasties from July 1, 2016 through December 31, 2017 constituted the intervention group. Patients undergoing elective, primary total joint procedures in the 18 months prior to the study period (January 1, 2015 through June 30, 2016) provided the pool from which the historical control group was identified. Exclusion criteria included patients requiring revision surgery or prior joint replacement surgery, active infections, patients who did not attend our pre-operative education class, minors, those with citrus or alcohol allergies, and mentally disabled patients.

Patients in the historical control cohort were identified using our total joint database, National Healthcare Safety Network (NHSN) reporting, and electronic medical record system. Exclusion criteria for the historical cohort were identical to the experimental cohort. Prior to the use of the protocol described in this study, neither routine nasal decolonization nor screening were performed. With the exception of nasal decolonization, our peri-operative protocols were standardized among the five surgeons during the study period and during the control cohort time periods.

During the intervention period, patients underwent skin decolonization with an antiseptic shower (HibiClens) the night prior and the morning of surgery, and chlorhexidine wipes (Stryker) were used in the pre-operative area. Patients received intravenous cefazolin prophylaxis prior to incision, patients at elevated risk for methicillin-resistant Staphylococcus aureus (MRSA) carriage including recent hospitalization or surgeries (i.e., within the past 12 months), previous MRSA infections, or patient-specific or environmental-exposure risk factors for MRSA, received vancomycin in addition to cefazolin, and patients with a cefazolin anaphylactic reaction documented in their chart received clindamycin instead. Our surgical closure and dressing changes were standardized among all surgeons, which included a deep layer with braided absorbable sutures, a subcuticular layer with braided absorbable sutures, a monofilament layer for the skin, and a topical skin adhesive. The dressing of choice was a foam absorbent dressing.

On the day of surgery, patients were instructed on the proper use of an alcohol-based nasal decolonization product (Nozin) in the pre-operative area by the nursing staff according to the manufacturer's instruction. It was applied by the patient in the pre-operative area and every eight hours for a total duration of 14 days after surgery. Application of the nasal antiseptic was supervised by nursing staff during the admission to ensure proper compliance. Additionally, the patient's principal post-operative caregiver (family member, friend, etc.) underwent nasal decolonization in the same manner as the patient beginning when the patient was discharged from the hospital. Adherence to the nasal decolonization protocol was recorded in a cohort of 293 patients using a tracking card filled out by the patient or caregiver.

Patients were monitored for infection for 90 days after surgery at post-operative appointments that were scheduled two weeks, six weeks, and three months post-operatively. Rates of SSI were obtained by referencing our total joint database, cross referenced with our institution's NHSN reporting, and verified manually by retrospective chart review. Diagnosis of SSI was based on the U.S. Centers for Disease Control and Prevention and the NHSN guidelines. ¹⁴ Investigators were blinded to rates of infection prior to completion of data collection. In patients developing SSI, standard cultures were taken where applicable.

Data analysis was performed using JMP statistical software (SAS Institute, Cary, NC). Categorical analysis was performed where applicable. Statistical significance was set at p < 0.05. Power analysis was performed prior to the collection of data. We considered an 80% reduction in SSIs as the effect size, and with an established infection rate of 2.5% prior to data collection, a sample size of 579 patients was needed.

Results

In all, 1,426 patients were included in the primary analysis. There were 779 patients in the nasal decolonization arm of the study and 647 historical control patients. No significant differences between body mass index (BMI), gender, duration of surgery, American Society of Anesthesiologists (ASA) score, or procedure location (hip vs. knee) were identified. Significant differences between groups were noted regarding discharge disposition (home vs. rehabilitation facility or skilled nursing facility) and age (Table 1).

Patients receiving nasal decolonization had a SSI rate of 0.64% (5/779), whereas the rate for the control group was 1.55% (10/647; Fig. 1). After adjusting for age, gender, BMI, ASA classification, and discharge disposition, multivariable analysis demonstrated that the alcohol-based nasal sanitization protocol reduced the rate of infection by 41.3% (p = 0.048; odds ratio, 2.43).

OPEN IN VIEWER

FIG. 1.

Surgical site infection rates.

Adherence to the protocol was tabulated using patient tracking cards for a cohort of 293 patients. Patients self-administered a mean of 32.4 doses (range, 0–42; standard deviation [SD], 12.29) of the required 42 doses, with 9.21% (27/293) taking zero doses after they were discharged. Adherence during admission was 100%. Of the primary caregivers undergoing decolonization, 38.9% (114/293) took at least one dose for a mean of 28.9 doses (range, 0–42; SD, 12.29). Adherence diminished as the time from surgery increased (Fig. 2).

OPEN IN VIEWER

FIG. 2.

Adherence rates for self-administered nasal decolonization.

Discussion

Nasal colonization with infectious organisms is considered a modifiable risk factor for SSI after orthopedic surgery.^4,15,16 Several studies have demonstrated that a large proportion of Staphylococcus aureus strains detected from SSIs match those swabbed from the nares, suggesting a potential auto-inoculation mechanism.^4,5,17 These studies illustrate the importance of decreasing endogenous bacterial load to prevent SSI. The literature is rife with studies advocating the use of antibiotic decolonizing agents such as mupirocin for the prevention of SSI; such protocols have become policy at many institutions.^18,19 Although they appear somewhat effective, approaches utilizing decolonization only in the pre-operative setting do not address the potential for recolonization or de novo colonization in the post-operative setting.

Additionally, widespread use of antibiotic agents has led to an increased risk of antibiotic resistance. The blanket use of intra-nasal mupirocin for nasal decolonization has been accompanied by the emergence of Staphylococcus aureus strains exhibiting resistance.^11,20,21 The increasing incidence of multi-drug–resistant organisms highlights the need for a non-antibiotic alternative for nasal decolonization.

Widely-used mupirocin has minimal effect on gram-negative and anaerobic bacteria. ²² A study of 1,878 community volunteers found that, in addition to a 41.0% carriage rate of Staphylococcus aureus, 33.4% of volunteers were nasally colonized with Enterobacteriaceae. ²³ This finding was further supported by a study by Micheel et al. ²⁴ in which samples from 1,541 students and healthcare workers were obtained and found 52.95% colonized with gram-negative bacteria. Moreover, a substantial number of patients are discharged to nursing homes. Roghmann et al. ²⁵ demonstrated a difference in femoral skin colonization with gram-negative bacteria compared with the general population.

The antiseptic effects of ethanol are a result of its ability to induce leakage of the plasma membrane. ²⁶ Because ethanol does not rely on interactions with cellular receptors, it has strong and immediate bactericidal effects. ²⁷ Additionally, no acquired bacterial resistance of clinical isolates to ethanol has been reported. ²⁸

The present study evaluated the effectiveness of an alcohol-based nasal antiseptic agent in decreasing SSIs in elective total joint arthroplasty patients. Utilization of this agent in the immediate pre-operative period and two weeks post-operatively resulted in a 41.3% reduction in acute SSIs from 1.55% to 0.64% (p = 0.048). Our decolonization protocol was unique in two aspects: the use of an alcohol-based, non-antibiotic agent for carriage reduction and its daily administration for a prolonged period (two weeks) post-operatively.

We selected a two-week period of administration after surgery to prevent re-colonization and mitigate the risk of new colonization during the time the surgical site is vulnerable to infection from topical inoculation. Several studies investigated the persistence of decolonization and found that these rates decrease with time.^29,30 Immerman et al. ²⁹ found that three months after decolonization, only 70% remained decolonized and a small number became newly colonized. They concluded that decolonization should be repeated prior to re-admission, regardless of prior colonization status. Baratz et al. ³¹ found that 22% remain colonized with Staphylococcus aureus despite a decolonization protocol. Although a short course of mupirocin prior to surgery has the potential to decrease nasal carriage of Staphylococcus aureus for a prolonged period of time, it does not do so consistently in a substantial proportion of patients. Conversely, a nasal antiseptic, when applied repeatedly in the post-operative period, continues to provide reproducible reduction of nasal carriage without fear of developing resistance.

The results of the present study are comparable to those evaluating the effects of alcohol-based nasal sanitization on rates of SSIs performed in surgical spine patients. Mullen et al. ³² evaluated 673 patients undergoing spine surgery using an alcohol-based nasal decolonization product pre-operatively and five to seven days after surgery. The mean SSI rate was decreased by 81%. ³² The present study shows favorable rates of infection reduction despite no attempt to decolonize nursing or surgical staff.

The rate of SSI reduction from 1.55% to 0.64% observed in the present study is similar to reported reduced infection rates found after decolonization with mupirocin and povidone-iodine. For example, Gernaat-van der Sluis et al. ⁸ reported a decrease in overall SSI rate from 2.70% to 1.34% (50.4% decrease) after treatment with pre-operative mupirocin in patients with known Staphylococcus aureus colonization. In another screen and treat study, Rao et al. ⁷ demonstrated improvements in infection rates from 2.6% to 1.5% (42.3% decrease) after decolonization. Phillips et al. ³³ observed a lower rate of SSI in patients receiving intra-nasal povidone-iodine compared with mupirocin.

Furthermore, the use of alcohol-based nasal sanitization provided excellent MRSA coverage by eliminating nasal carriage as a vector compared with the historical control group (Table 2). These findings are listed in chronological order. Additionally, unlike mupirocin, which has minimal gram-negative coverage, our protocol resulted in complete elimination of gram-negative organisms from SSIs during this study period.

Reported adherence to the prescribed protocol was surprisingly high. Despite the requirement for three times daily use over two weeks, more than 50% of patients reported using 40 or more of the required 42 doses. Several studies have examined patient satisfaction with nasal decolonization agents and found low rates of adherence and reports of local irritation.^34,35 Further investigation is warranted to determine how well patients tolerate the present protocol. We observed an inverse relation between the number of applications and the time from surgery, and the required duration likely contributed to an eventual reduction in adherence. Despite these shortcomings, we noted excellent compliance while the patient was admitted, and we believe this period to be especially critical for preventing colonization with nosocomial organisms.

The present study must be interpreted knowing its limitations. First, the difference found regarding infection rates between groups was found when performing a χ ² analysis, but not when performing regression analysis. This is likely because of confounding variables that were adjusted for. The study was underpowered for that analysis, and we would need a much larger patient population in both groups to demonstrate a difference in that regard after adjusting for these variables, because the overall infection rates is so low to begin with, which would not be realistic to perform at this time at our institution.. Additionally, we universally decolonized all patients without screening. Previous studies have shown that 20% to 30% of the population are nasal carriers of methicillin-sensitive Staphylococcus aureus and 1% to 5% are carriers of MRSA, and blanket inclusion of non-colonized individuals likely diluted our results with regard to the rate of SSI reduction in colonized individuals.^17,36 This limitation makes comparison of SSI reduction rates with other studies difficult. Second, we did not perform a prospective controlled study. The present study used a historical control group, which has sometimes been associated with confounding variables and potentially false outcomes. ⁴

We attempted to minimize these effects by selecting a sequential cohort of patients and adhering to strict protocols regarding surgical closure, peri-operative antibiotic agents, drain placement, skin decolonization, and dressing changes. The intervention and control groups had notable differences in age and discharge disposition, and we cannot rule out any effects these differences may have had on our infection rate. We also noted a difference in disposition at discharge, but there were no differences in infection rates regardless of where the patient was discharged. Last, although we used previously published figures to extrapolate approximate cost savings, we did not quantify the direct savings to the patient or our institution. Prior studies evaluating cost effectiveness of nasal decolonization with other agents have demonstrated large savings. ³⁷ With emerging payment bundles and capitated payments, further studies should assess the cost effectiveness of this approach to SSI prevention.

Conclusions

In conclusion, the results of the present study demonstrate that SSI rates were markedly reduced in total joint arthroplasty patients who followed a nasal decolonization protocol utilizing an ethanol-based antiseptic. The integration of the pre- and post-operative protocol into existing infection control and prevention measures was readily implemented and demonstrated a high level of acceptance and application compliance.