Study population and trial design

This trial was conducted at William Osler Health System (WOHS), an 870-bed tertiary-care hospital in Ontario, Canada. In a prior study,^{Reference Jamal, Mataseje and Brown6} 49 CPO-contaminated drains were identified at WOHS between October 2017 and January 2018: 34 harbored Enterobacterales, 11 harbored Pseudomonas spp, 3 harbored Acinetobacter spp, and 1 harbored Shewanella spp. These 49 drains were cultured again twice in August 2018. The 32 drains that remained CPO contaminated on both culture days in 2018 were deemed to still be contaminated and were thus eligible for this study. These drains were stratified by drain type (sink vs shower) and randomized to intervention or comparator.

Study intervention, randomization, and blinding

The intervention was combined chemical, mechanical, and heat cleaning of CPO-contaminated drains administered once on day 0 by a contracted plumber. The intervention protocol was developed by the infection prevention and control department at St Joseph’s Health Centre, Toronto, Ontario, in 2016. First, the shower or sink grate was removed and was scrubbed and rinsed. Second, the P-trap was filled with sodium hydroxide (White Magic, Chemfax, Calgary, Canada) and plugged to allow 10 minutes of contact time; the drain was then brushed and rinsed with water for 10 minutes. Third, the P-trap was filled with enzymatic solution (EmPower, Metrex Research, Orange, CA) and plugged to allow 10 minutes of contact time; the drain was then brushed and rinsed with water for 10 minutes. Fourth, the P-trap was filled with accelerated hydrogen peroxide (Rescue Gel Sporicidal, Diversey, Charlotte, NC) and plugged to allow 10 minutes of contact time; the drain was then brushed and rinsed with water for 10 minutes. Finally, a steamer tip was inserted into the drain for 10 minutes.

The comparator was the existing protocol for CPO-contaminated drains at WOHS administered by trained hospital housekeeping staff once on day 0 on regular wards and weekly for sinks in the intensive care unit (ICU). Accelerated hydrogen-peroxide sporicidal gel (300 mL, Rescue Gel Sporicidal, Diversey, Charlotte, NC) was poured into the drain and onto the sink basin surrounding the drain, allowed 1 minute of contact time, then rinsed with running water for 2 minutes.

Eligible drains were randomized with a 1:1 allocation ratio using proc RANUNI in SAS software (SAS Institute, Cary, NC). Randomization was performed by a research assistant not otherwise involved in the study, who communicated drain assignments to the contracted plumber or housekeeping staff administering the intervention or comparator. Study investigators, WOHS infection prevention and control department members, and personnel obtaining and processing drain specimens were blinded to drain allocation.

Collection and laboratory processing of drain specimens

One specimen was collected from each drain on day 0 (before administration of the intervention or comparator), day 1, day 2, day 3, day 7, day 14, month 1, month 2, month 3, month 4, month 5, and month 6. A 15-cm sterile cotton swab was inserted as far into the drain tailpiece as possible and all surfaces of the tailpiece were swabbed vigorously. Swabs were placed into brain heart infusion with 10% Dey-Engley neutralizing broth (Hardy Diagnostics, Santa Maria, USA) and were incubated at 37°C for 18–24 hours. Two hundred microlitres of incubated broth was centrifuged at 5,000 rpm for 5 minutes, then the pellet was resuspended in 100 µL lysis buffer (1% Triton X-100, 0.5% Tween 20, 1 mM EDTA, 100 mM Tris-HCl, pH 8) and boiled at 100°C for 10 minutes. The cell debris was removed by centrifugation and the supernatant was preserved. The PCR reaction mix (2× Qiagen Quantitect Probe PCR Master Mix, 10 μM forward and reverse primers, 2.5 μM Taqman Pro probes) was thawed, then the PCR was set up in a 96-well plate with 3 reactions per sample for 3 different sets of multiplex reactions targeting different carbapenemase genes. PCR was run on an ABI 7500 Real-time Sequence Detection System (Applied Biosystems, ThermoFisher Scientific, Waltham, MA).^{Reference Tijet, Patel and Melano7}

Broth from direct PCR-positive specimens from day 0, day 14, month 2, month 4, and month 6 were subcultured onto a Supercarba plate (CHROMagar, Paris, France). Each morphologically distinct colony was identified to the species level using matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS, bioMérieux, Marcy-l’Étoile, France), then was subcultured onto blood agar (Oxoid, Nepean, Canada). The mCIM test⁸ was performed on Enterobacterales isolates. PCR for carbapenemase gene detection was performed as described above for each mCIM-positive Enterobacterales isolate and all afermenters.

Inpatients with CPO colonization or infection at WOHS

Room locations were documented for all inpatients with CPO colonization or infection at WOHS during the trial. The hospital operates at >100% capacity, and patient room assignments were not influenced by the trial.

Study outcomes

The primary outcome was drain decontamination, defined as no detectable carbapenemase gene in the drain from day 1 to day 7 (inclusive).

Sample size

We estimated that 50% of drains in the comparator group would remain contaminated at day 7. With a power of 80% and a 2-sided α of 0.05, we would be able to conclude that the intervention was superior to the comparator if 10% or fewer drains in the intervention group remained contaminated.

Statistical methods

The proportion of drains that were decontaminated in intervention versus comparator groups were compared using the Fisher exact test. The relative risks (and 95% confidence intervals, CIs) of decontamination in the 2 groups were calculated. To determine whether direct PCR cycle threshold (Ct) value was associated with successful culture of a CPO, we used a logistic regression model with a generalized estimating equation with an exchangeable correlation matrix to account for cultures from the same drain over time. Analyses were conducted using SAS version 9.4M6 software (SAS Institute, Cary, NC).