Post-discharge decolonization of patients harboring methicillin-resistant Staphylococcus aureus (MRSA) USA300 strains: secondary analysis of the CLEAR Trial

Abstract

The CLEAR Trial recently found that decolonization reduced infections and hospitalizations in MRSA carriers in the year following hospital discharge. In this secondary analysis, we explored whether decolonization had a similar benefit in the subgroup of trial participants who harbored USA300, using two different definitions for the USA300 strain-type.

Introduction

Despite a substantial decline in methicillin-resistant Staphylococcus aureus (MRSA) bloodstream infections between 2005-2012, recent progress has slowed and MRSA remains a major cause of morbidity and mortality in the U.S. ¹ MRSA was primarily healthcare-associated until the mid-1990s, when a specific clone began appearing in the community with distinctive antimicrobial susceptibility patterns and epidemiologic features, including a propensity for causing necrotizing infection. ² USA300 is the predominant community-associated MRSA strain in the U.S., and has emerged as a significant cause of healthcare-associated infection. ^3,4

There has been longstanding interest to examine novel strategies to prevent infections due to USA300. It is unclear whether USA300 prevention requires unique strategies compared to other strains (e.g., USA100) because of differences in virulence genes and clinical presentation. ^1–4 Research to identify the most effective strategies for preventing USA300 is warranted because evidence suggests that bloodstream infections from USA100 have declined while those due to USA300 have not. ^1,3

Risk for invasive MRSA infection is greatest in the first six months following hospital discharge and remains elevated for one year. ⁵ The CLEAR (Changing Lives by Eradicating Antibiotic Resistance) Trial found that decolonization reduced MRSA infection by 30% and all-cause infection by 17% among recently hospitalized MRSA carriers in the year post-discharge. ⁶ It is possible that response to decolonization differs when comparing USA300 to other MRSA strain-types. We conducted a secondary analysis of the CLEAR Trial to evaluate whether decolonization had a similar benefit among USA300 carriers as it did for the full trial population.

Methods

CLEAR was a randomized controlled trial of 2,121 recently hospitalized MRSA carriers from January 2011–January 2014. ⁶ Consented participants were randomized to hygiene education alone or hygiene education plus repeated decolonization consisting of a 5-day regimen of 4% chlorhexidine soap and 0.12% chlorhexidine mouth rinse plus twice-monthly nasal mupirocin for 6 months post-discharge. Nares, skin, throat, and, wounds, if present, were swabbed upon enrollment and at 1, 3, 6, and 9 months of follow-up. This trial was approved by the University of California, Irvine institutional review board.

MRSA strains from participants who completed all follow-up visits were processed for whole genome sequencing (WGS). ⁷ Sequencing reads were assembled de novo using SPAdes v3.12.09. Contigs <500bp were discarded. The in silico multi-locus sequence typing (MLST) was inferred from assemblies using PUBMLST database (https://pubmlst.org/saureus/) and custom scripts. Staphylococcal cassette chromosome mec (SCCmec) typing was performed using pairwise blast to SCCmec elements (ccr, mecA, mecI, mecR genes). Combination and orientation of SCCmec hits were parsed by custom scripts to infer SCCmec type as defined in (http://www.sccmec.org). Staphylococcal protein A (spa) typing was performed using Ridom database (http://www.spaserver.ridom.de). Panton-Valentine Leukocidin (PVL) LukS/LukF genes and the arginine catabolic mobile element (ACME-arcA) were assessed for descriptive purposes.

We evaluated 2 USA300 subgroups: a 2-criteria subgroup defined by presence of SCCmec type IV and MLST 8, and a 3-criteria subgroup additionally requiring spa type t008, the dominant spa type associated with USA300. ⁸ We allowed flexibility in spa type because previous work in Southern California, the location of the CLEAR Trial, identified several spa types closely related to t008 in the region, including t024. ⁸

We assessed USA300 subgroups for the primary trial outcome, time to first MRSA infection by Centers for Disease Control and Prevention criteria as determined by blinded chart review, using unadjusted Cox proportional hazard models. We performed adjusted proportional hazard models accounting for participant demographics, comorbidities, presence of wounds or devices, history of hospitalization or nursing home residence in the year prior to enrollment, and receipt of anti-MRSA antibiotics as a time-varying covariate. We performed “per-protocol” and “as-treated” analyses to account for adherence to the decolonization protocol.

Results

Complete follow-up was achieved for 1,191 (56.2%) of 2,121 trial participants. Among 783 participants with MRSA strains tested, 420 (53.6%) met the 2-criteria USA300 definition and 321 (41.0%) met the 3-criteria definition with spa type t008. Among 99 participants in the 2-criteria cohort without t008, 38 spa types were identified; 76 participants carried spa types highly related to t008, ⁴ including t622 (N=47) and t024 (N=15), and 23 carried spa types that were not highly related to known healthcare-associated spa clonal complexes. Presence of ACME-arcA, PVL lukF, and PVL lukS genes was detected in the majority of participants meeting 2-criteria (66.4%) and 3-criteria (86.7%) USA300 definitions.

Characteristics of USA300 carriers were similar to the full trial population (Table 1). ⁶ We note that the 2-criteria USA300 cohort had a greater proportion of male participants than the 3-criteria cohort. Frequency and rates of MRSA outcomes (first infection per person) are described in Table 2. For the overall trial population, the decolonization arm experienced a 0.041 absolute reduction (30% relative reduction) in MRSA infections per participant-year. In comparison, the 2-criteria USA300 cohort experienced a 0.050 absolute reduction (33% relative reduction) in MRSA infections per participant-year and the 3-criteria USA300 cohort experienced a 0.014 absolute reduction (10% relative reduction) in MRSA infections per participant-year. Compared to the overall trial population, USA300 carriers by either definition experienced higher proportions of skin and soft tissue infections, and lower proportions of bacteremia and bone and joint infections. Outcomes were similar when USA300 carriers by either definition were compared to those in the full trial cohort without USA300.

Table 1. Characteristics of USA300 Carriers in the Project CLEAR Trial

Abbreviations: CLEAR (Changing Lives by Eradicating Antibiotic Resistance), MLST (multilocus sequence typing), SCCmec (staphylococcal cassette chromosome mec), spa (staphylococcal protein A gene), ICU (intensive care unit), MRSA (methicillin-resistant Staphylococcus aureus).

^a Data reflect a response to a survey question or chart review. Not all participants responded to every question, and not all enrollment charts were received from recruiting hospitals despite a signed release request. Across the 2,121 trial participants, data were missing for 21 participants.

^b Scores on the Charlson Comorbidity Index range from 0 to 10, with higher scores indicating more coexisting illness.

Table 2. MRSA Infection Outcomes (First Infection per Person) per 365 Days of Follow-Up, According to CDC Criteria

Abbreviations: CDC (Centers for Disease Control and Prevention), MLST (multilocus sequence typing), SCCmec (staphylococcal cassette chromosome mec), spa (staphylococcal protein A gene), MRSA (methicillin-resistant Staphylococcus aureus).

Supplementary Table 1 provides estimates of “per-protocol” and “as-treated” decolonization effects according to Cox proportional-hazard models. “Per-protocol” point estimates of hazard ratios (HR) were similar between USA300 carriers by the 2-criteria definition (0.59 (0.32, 1.09)) and the full trial population (0.61 (0.44, 0.85)), but were higher for the 3-criteria cohort with spa type 008 (0.82 (0.42, 1.59)). “As-treated” HR point estimates decreased with increased adherence for USA300 subgroups, consistent with overall trial findings.

Discussion

The CLEAR Trial demonstrated that post-discharge decolonization significantly reduced MRSA infections and hospitalizations. We found that USA300 carriers assigned to decolonization experienced a 33% reduction in MRSA infections compared to USA300 carriers assigned to education when using a 2-criteria definition. This was similar to the overall trial findings. However, when using a 3-criteria definition for USA300, only a 10% reduction was seen with decolonization. Explanations may include the fact that compared to the 2-criteria cohort, the 3-criteria cohort had an approximately 30% greater proportion of ACME/PVL genes, which may confer greater virulence, and a gender disparity of 30%, which may be associated with differential quality when applying decolonization. In a prior evaluation, we found that chlorhexidine concentration on the skin was nearly two-fold greater among female participants, and that female participants were more likely to use a mesh-sponge when applying chlorhexidine in the shower. ⁹

This study has important limitations. First, ideal criteria for defining USA300 are still lacking. ¹⁰ To account for this, we evaluated 2 cohorts using different definitions for USA300. Second, while this study is limited to a 1-year follow-up period, it is known that MRSA infection risk is highest within 6 months of hospital discharge and normalizes after a year. ⁵ Finally, although the trial was not powered to evaluate a USA300 subgroup, it provides a valuable design for assessing the magnitude of strain-specific responsiveness to decolonization during a time when national rates of MRSA invasive disease have plateaued and USA300 is responsible for an increasing proportion of infections. ³ Larger studies may still be needed to examine differences in the outcomes and prevention of MRSA infections due to USA300 versus other strain-types.