Oral Presentations

doi:10.3138/jammi_4.s1.abst-01

. 2019 Jul 11;4(Suppl):1–38. doi: 10.3138/jammi_4.s1.abst-01

Oral Presentations

PMCID: PMC9603425

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):1.

A01. Epidemiology of Clostridioides difficile Infection in Canada: A Six-Year Review of Provincial Surveillance Data

Y Xia ¹, M Tunis ², C Frenette ³, K Katz ⁴, K Amaratunga ^2,⁵, S Rhodenizer-Rose ⁶, A House ², C Quach ^7,⁸

Background

Two vaccines against Clostridioides difficile infections (CDI) are currently in phase III trials. To enable decision-making on their potential use in public health programs, national disease epidemiology is necessary. This study aims to determine the epidemiology of hospital-acquired CDI (HA-CDI) and community-associated CDI (CA-CDI) in Canada, using provincial surveillance data from ten Canadian provinces and provide a comprehensive discussion of current provincial surveillance programs.

Methods

We used publicly-available CDI provincial surveillance data from 2011 to 2016 with the most common surveillance definition for each province. Pooled HA-CDI incidence rates and CA-CDI proportions (%) were calculated for each province. Both HA-CDI and CA-CDI incidence rates (IRs) were examined for trends.

Results

Data from Manitoba were excluded given the substantial differences in the surveillance definition. HA-CDI rates from other provinces ranged from 2.1–6.5/10,000 inpatient-days, with a decreasing trend over time, while available data on CA-CDI show that both rates and proportions have been increasing over time. The absence of a common case definition for CDI surveillance is problematic and impacts both the number of cases and denominators reported. Discrepant denominators were identified as a major problem.

Conclusion

There is a need for a nationally adopted common case definition for CDI, CDI classification, and common methods for total inpatient-days determination. We also need a quality assessment system to ensure that standardized and quality data are reported. This study highlights the limitations of current provincial CDI surveillance, in particular, when attempting to calculate a pan-Canadian burden of illness.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):1–2.

A02. Detection of Pneumocystis jirovecii with qPCR to Differentiate between Colonization and Pneumonia

A Hadzic ¹, P Jayaratne ^1,², D Leto ^2,³, M Smieja ^1,², D Yamamura ^1,²

Background and Objectives

Pneumocystis pneumonia (PCP) is an opportunistic infection caused by the fungus Pneumocystis jirovecii (Pj). Diagnosis of PCP remains difficult and relies on microscopic methods with low sensitivity (direct immunofluorescent assay-DFA). Our study objective was to evaluate the use of qPCR to differentiate Pj infection from colonization and to investigate clinical and laboratory characteristics of patients with PCP.

Methods

We screened 415 consecutive bronchoalveolar lavage (BAL) specimens collected between January 2013 and September 2017 for Pj using a qPCR targeting the mitochondrial large subunit (mtLSU) region. Pj-positive specimens were further assayed with a qPCR targeting a cyclin-dependent kinase (CDC-2) gene. Two infectious diseases physicians, blinded to qPCR values, performed standardized retrospective chart review to classify patients as PCP or Pj colonized. Cycle threshold values (Ct) estimating fungal burden were compared with clinical diagnosis as the reference standard.

Results

Pj was detected by mtLSU-qPCR in 59 patients, in whom PCP was diagnosed in 13 patients by a combination of clinical and laboratory criteria excluding qPCR. DFA diagnosed PCP in 8 and qPCR was positive in all 13 patients. Ground-glass and interstitial changes on chest radiography were more frequent in PCP (p<.01). Higher fungal burden was found in PCP compared with colonized patients using mtLSU target (median Ct 26 and Ct 35 respectively; p<.01) and using CDC-2 target (median Ct 36 and Ct 39 respectively; p<.01). The thresholds corresponding to a specificity of 100% were 21 for mtLSU and 32 for CDC-2 target.

Conclusions

qPCR showed value diagnosing PCP in 5 additional patients compared with DFA. Clinical context is important in interpretation of indeterminate qPCR results. Further work to identify Ct values for diagnosis of PCP is needed.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):2.

A03. Novel Metabolomics Approach for the Diagnosis of Respiratory Viruses Directly from Nasopharyngeal Specimens

CA Hogan ^1,², AT Le ^1,³, TM Cowan ^1,³, BA Pinsky ^1,²

Objectives

Respiratory virus infections, including influenza A and B, are important causes of morbidity and mortality among pediatric and adult patients. These viruses infect respiratory epithelial cells, where they may induce metabolite alterations. As a proof-of-concept, we investigate the novel use of liquid chromatography (LC) combined with mass spectrometry (MS) for the study of host cell metabolite alterations to diagnose and differentiate respiratory viruses.

Methods

We studied nasopharyngeal swab samples positive for respiratory viruses by multiplex reverse transcriptase-polymerase chain reaction assay (RT-PCR) (GenMark Diagnostics, Carlsbad, CA). Banked, frozen samples (-80°C) stored in viral transport media were retrieved and thawed. Aliquots of 100μL were centrifuged at 13.3 × g for 15 minutes and analyzed by Agilent 6545 Quadrupole Time-of-Flight (Q-TOF) LC/MS (Agilent Technologies, Santa Clara, CA). Agilent Mass Profiler 3D principal component analysis was performed, and compound identification was completed using the METLIN metabolite database.

Results

A total of 130 samples were tested by Q-TOF LC/MS, including 120 positive samples [10 samples of each including adenovirus, coronavirus, influenza A H1N1 and H3N2, influenza B, human metapneumovirus, parainfluenza viruses 1, 2, 3, and 4, respiratory syncytial virus (RSV), and rhinovirus] as well as 10 negative clinical specimens. Q-TOF LC/MS allowed identification of key metabolites that distinguished all virus positive samples compared with the negative group, as well as differentiating these respiratory viruses from one another including between influenza A H1N1 and H3N2 subtypes.

Conclusions

Preliminary data from our Q-TOF LC/MS analysis show that respiratory viruses exhibit different host cell metabolomic profiles that allow viral differentiation to the species level, and for influenza A virus, the subtype level. This metabolomic approach has substantial potential for diagnostic applications in infectious diseases directly from patient samples and may be eventually adapted for point-of-care testing.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):2–3.

A04. Applying a Next-Generation Sequencing Pipeline in a Clinical Microbiology Laboratory for Surveillance and Outbreak Investigations: Defining the Genomic Epidemiology of Klebsiella aerogenes and Identifying Determinants of Carbapenem Resistance and Virulence

A Malek ¹, K McGlynn ¹, S Taffner ¹, L Fine ², B Tesini ³, J Wang ¹, H Mostafa ¹, S Petry ¹, A Perkins ¹, P Graman ³, D Hardy ^1,⁴, N Pecora ¹

Background

Next-generation sequencing (NGS) is a powerful resource to clinical microbiology laboratories for surveillance and outbreak investigations. Two temporally spaced clusters of carbapenem-resistant K. aerogenes (CR-KA) isolates (n = 15, 2017; n = 3, 2018) from patients admitted to a cardiac intensive care unit (CICU) were prospectively investigated to identify genetic relatedness, population structure, and determinants of carbapenem resistance and virulence.

Methods

31 patient strains (18 CICU-associated, 13 other wards) were sequenced (Illumina Miseq) and analyzed by a modular, in-house bioinformatics pipeline designed to flexibly address multiple clinical applications. A modified CFSAN-SNP module performed Hq-variant calling and markers corresponding to MLST, antibiotic resistance and virulence factors were identified using curated databases. Comparative genomics of local and global strains (publicly available K. aerogenes genome assemblies, n = 111) was assessed using Harvest and Kleborate suites.

Results

Phylogenomic analyses indicated that every CICU-associated CR-KA isolate was part of a clonal cluster (<9 SNPs apart), indicating protracted intra-ward transmission. No clonal relationships were observed between the CICU isolates and those from other hospital wards. Genes encoding carbapenemases were not detected and carbapenem resistance was attributed to mutations impacting AmpD activity and membrane permeability. The CICU strains harbored an integrative conjugative element (ICEKp10) that is associated with hypervirulent Klebsiella pneumoniae lineages. Comparative genomics showed the outbreak-associated strains to group closely with global ST4 strains, which may represent a dominant K. aerogenes lineage associated with human infections.

Conclusion

Along with high-resolution tracking of transmission events and assessment of the effectiveness of infection control measures, NGS facilitates investigation of cryptic drug-resistance and virulence. For poorly characterized pathogens, scaling analyses to include sequenced genomes from public databases offers tremendous opportunity to identify emerging trends and dominant clones associated with specific attributes, syndromes and geographical locations.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):3.

A05. Investigation of the Presence and Impact of Environmentally-Adapted Escherichia coli in Private Well Water in Southeastern Ontario

E Tropea ¹, M Kelly ¹, S Brown ¹, A Majury ^1,²

Background

The fecal contamination of drinking water causing acute gastrointestinal illness remains a threat to public health in Canada, including Ontario. To determine potability, water is assessed for the presence of fecal indicator bacteria (FIB) to indicate contamination with harmful fecal pathogens. Escherichia coli remains the most frequently employed FIB, given it is a natural inhabitant of the gastrointestinal tract (GIT) of warm-blooded animals. The use of E. coli as a FIB has been questioned in recent years with the identification of populations of environmentally adapted E. coli capable of surviving outside of the GIT and thus do not represent a recent fecal contamination event.

Methods

A total of 325 E. coli isolates from southeastern Ontario private wells were characterized by comparing their phylogenetic origin using Clermont phylotyping, and identifying unique genetic fingerprints through qPCR. Six accessory genes previously shown to distinguish environmentally adapted from fecal E. coli in surface water were investigated, including iutA, ccdB, clpXET1, hra1, phd and tratT. For comparison, 234 isolates from animal feces were also characterized.

Results

Of the 76 private wells analyzed, 66% had evidence of environmentally adapted E. coli. Further, 37% of these same wells had only environmentally adapted strains in the subset of E. coli isolates selected. Environmentally adapted E. coli were determined to have originated from several distinct phylogroups, including B1, B2, A, D and cryptic clades III-V.

Conclusions

Environmentally adapted E. coli are common in private well water in southeastern Ontario and come from several phylogroups. This has potential implications for E. coli as an FIB and for water quality testing methodologies and interpretations. The potential health risk associated with drinking water containing environmentally adapted populations of E. coli is not understood. Current work includes assessing these environmentally adapted strains in the context of virulence, disease, and as potential reservoirs of antimicrobial resistance genes.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):3.

B01. Withdrawn

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):3–4.

B02. Evaluation of Five Commercial Nucleic Acid Tests for Bacterial Gastroenteritis in Alberta, Canada

BM Berenger ^1,^2,³, L Chui ^1,⁴, C Ferrato ¹, T Lloyd ², DR Pillai ^2,³

Objective

Compare the performance of commercial nucleic acid tests (NATs) for enteric bacteria in Alberta, Canada.

Methods

BD Max Enteric Bacterial Panels, Fast-Track Diagnostics Bacterial Gastroenteritis, Prodesse Progastro SSCS, RidaGene Bacterial Stool Panel I and EHEC/EPEC, and Seegene Allplex Gastrointestinal Panels were used to test both patient feces and contrived specimens representing circulating serotypes and species in Alberta (see Table B02-1 for organisms in contrived specimens). Patient fecal samples that were culture-negative (n = 46) and positive for Campylobacter (n = 40), Salmonella (n = 40), Shigella (n = 15), Shiga toxin (stx)-producing Escherichia coli (STEC) (n = 20), or Yersinia enterocolitica (n = 10) were tested. Clinical specimens with a false negative NAT result were retested with the DNA extract; if the repeat result was negative, the original specimen was re-extracted and tested. Contrived specimens with false-negative results on a NAT were repeated with the specimen and/or isolate suspension.

Table B02-1:

Isolates in contrived specimen panel

Organism	Isolates in contrived specimen panel
Salmonella (n = 25)	Predominant serotypes from human (n = 20) and food sources (n = 5).
Shigella (n = 19)	S. flexneri, dysenteriae, and boydii serotypes isolated >twice in Alberta 2009–2016 (n = 18). S. sonnei (n = 1).
STEC (n = 10)	1a/2a;1c/2b;1d;2a;2b;2b/2c;2d;2e;2f;2g
Campylobacter (n = 4)	C. coli, upsaliensis, jejuni subsp jejuni and doylei.
Yersinia (n = 9)	Y. rohdei, federiksenii, intermedia, and pseudotuberculosis (n = 4). Different Y. enterocolitica serotypes (n = 5).

Open in a new tab

Results

Based on consensus (minimum two positive NATs), the percent agreement of the NATs for each organism result were as follows: Shigella 100%, Salmonella 97.6%–100%, Campylobacter 95%–100% (n = 5 C. coli, n = 35 C. jejuni), STEC 90%–100%, Y. enterocolitica 70%–90%, negative for these five targets 78.5%–100%. The following isolates in the contrived specimen panel were missed by the denoted assays: C. upsaliensis (all NATs), Y. non-enterocolitica spp (all NATs), stx_2f (all except RidaGene), stx_1d (Prodesse).

Conclusions

Commercial NATs reliably detect Salmonella, Shigella, C. jejuni, C. coli and the most prevalent STEC subtypes with the exception of Yersinia spp.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):4.

B03. Comparison of Prevalence of Pathogenic Protozoa in Stool Samples Using Molecular Testing Compared to Traditional Microscopy

H Almohri ¹, D Lungrin ¹, D James ¹, D Leto ¹, V Tran ¹

Background

Giardia intestinalis, Cryptosporidium spp and Entamoeba histolytica are the most common diarrhea-causing parasitic protozoa. Diagnosis of these parasites is usually performed by microscopy; however, at our institution, a molecular platform was implemented as a primary diagnostic test for detection of parasitic protozoa.

Objective

To compare the positivity by PCR testing in the detection of pathogenic protozoa to the positivity using Microscopy.

Method

Data for Microscopy results for the period 2016–2018 were compared with the results of PCR using Seegene Allplex PCR technology. Microscopy was performed on stool samples received in SAF. Concentrated specimen and direct smears were performed for the detection of protozoan cysts, helminth eggs and larvae in small numbers. A Kinyoun/Hematoxylin stain was used for the detection and identification of protozoan cysts and trophozoites including Cryptosporidium and Cyclospora. PCR testing was performed using Allplex Seegene technology. Stool samples for molecular testing were collected in Copan swabs with Cary Blair.

Results

A total of 190,817 specimens were tested using microscopy over 3 years from 2016 to 2018. Blastocystis hominis had the highest prevalence at 5.22% positivity, followed by Dientamoeba fragilis at 2.84%, Giardia intestinalis at 0.90%, Entamoeba histolytica/dispar at 0.62%, Cyclospora spp at 0.19% and Cryptosporidium at 0.11%. PCR testing as primary diagnostic testing started in late 2017. Total number of specimens tested by PCR was 89,928. The prevalence of pathogenic protozoa was as follows: B. hominis 10.92%, D. fragilis 5.55%, G. intestinalis 1.45%, E. histolytica 0.03%, Cryptosporidium 0.67%, and Cyclospora spp 0.17%.

Conclusion

Detection of pathogenic protozoa in stool samples is higher using molecular testing for primary testing compared with the microscopy for most protozoa except for Cyclospora spp. The most noticeable increase was detected for Cryptosporidium spp. Entamoeba detection was lower since molecular testing detects only E. histolytica while microscopy does not differentiate between histolytica/dispar.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):4–5.

B04. Molecular Detection of Non-O157 Shiga Toxin-Producing Escherichia Coli (Stec) Directly from Stool Using Real-Time Multiplex PCR Assays

M Bording Jorgensen ¹, J Szelewicki ¹, B Parsons ¹, C Lloyd ¹, L Chui ^1,²

Objectives

Non-O157 Shiga-toxin producing Escherichia coli (STEC) can cause outbreaks that have great economic and health impact. Since implementation of all STEC screening in Alberta in 2018, it is also essential to have a molecular serotyping method with faster turn-around time for cluster identification and surveillance purposes. This study sought to 1) evaluate 2 molecular assays for prediction of the top 6 non-O157 [O26, O45, O103, O111, O121, and O145] serotypes and 2) to perform molecular serotyping directly from stools.

Methods

Two multiplex serotyping PCR assays were used to determine the sensitivity and specificity. Sensitivity was assessed by extracting DNA from 10-fold cell dilutions of the top 6 serotypes. PCR assays were performed as singleplex and multiplex for comparison. Specificity was determined using a panel of bacteria (n = 27), non-top 6 STEC isolates (n = 45) and top 6 STEC (n = 157). The final phase was to test blinded stool specimens (n = 60) or broth samples (n = 44) submitted from frontline laboratories for STEC investigation. Our serotyping results were further confirmed by the Public Health Agency-National Microbiology Laboratory using conventional serotyping method.

Results

Both singleplex and multiplex assays were comparable and we observed 100% sensitivity and specificity. Direct molecular serotyping from stool specimens correlated with conventional serotyping from cultured isolate serotype. Furthermore, top 6 non-O157 STEC mixed infections were identified and confirmed by culture and serology.

Conclusion

We have shown that detection of non-O157 STEC can be done directly from stool specimens using multiplex PCR assays with the ability to observe mixed infections, which would otherwise remain undetected by conventional serotyping of a single colony. This method can be easily implemented into a frontline diagnostic laboratory as we move to culture independent assays in the near future.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):5.

B05. Inferring Amoxicillin-Clavulanate Susceptibility from Penicillin Susceptibility for Clinical Isolates of Streptococcus pneumoniae—Rethinking the Cut-Off Value

P Lagacé-Wiens ^1,², A Walkty ^1,², H Adam ^1,², M Baxter ², J Karloesky ^1,², GG Zhanel ²

Objectives

Current CLSI guidelines indicate that a penicillin minimum inhibitory concentration of ≤0.06 mg/L can be used to infer susceptibility to oral amoxicillin and amoxicillin-clavulanate in Streptococcus pneumoniae. However, given the high bioavailability (80%) and higher susceptibility breakpoint of amoxicillin, this relatively low penicillin MIC cut-off has the potential of over-calling amoxicillin resistance thus limiting the use of these oral agents for the treatment of respiratory infections. We sought to compare penicillin and amoxicillin-clavulanate MICs to establish an optimal penicillin MIC cut-off value to predict amoxicillin-clavulanate susceptibility.

Methods

Isolates were collected from the annual, ongoing CANWARD study between 2007 and 2017. Antimicrobial susceptibility testing and interpretation of susceptibility was performed using broth microdilution panels prepared following CLSI recommendations. Sensitivity, specificity and accuracy of penicillin MIC cut-offs ranging from 0.06 mg/L to 8mg/L for accurately predicting amoxicillin-clavulanate resistance were determined.

Results

2,577 S. pneumoniae isolates were included. 2,466 (95.7%) isolates were penicillin-susceptible (MIC≤0.06 mg/L) and 111 (4.3%) were penicillin-non-susceptible. 54 (2.1%) of the 2,577 isolates were non-susceptible to amoxicillin-clavulanate. The sensitivity, specificity, and accuracy of various penicillin cut-offs for predicting amoxicillin-clavulanate resistance were as follows:

Table B05-1:

Penicillin MIC cut-offs

Penicillin MIC cut-off (mg/L)	Sensitivity (for resistance)	Specificity (for resistance)	Accuracy
≤0.06	100%	84.4%	84.7%
≤0.12	100%	90.5%	90.7%
≤0.25	100%	93.4%	93.5%
≤0.5	100%	95.4%	95.5%
≤1	96.3%	97.7%	97.7%
≤2	46.3%	99.4%	98.3%
≤4	5.6%	100%	98.0%
≤8	1.9%	100%	97.9%

Open in a new tab

Conclusions

The current penicillin cut-off of ≤0.06 mg/L will misclassify 15.6% of amoxicillin-clavulanate–susceptible isolates as resistant. A penicillin MIC cut-off of ≤0.5 mg/L will accurately predict amoxicillin-clavulanate resistance in 100% of resistant isolates and accurately predict amoxicillin-clavulanate susceptibility in 95.4% of susceptible isolates. Despite better accuracy of higher cut-off values, ≤0.5 mg/L represents a better cut-off given the importance of accurately classifying resistant isolates.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):5–6.

C01. Outcomes of Administering Cefazolin vs. Other Antibiotics in Penicillin Allergic Patients with Anaphylactic Reactions for Surgical Prophylaxis at a Major Canadian Teaching Hospital

W Song ¹, TT Lau ², S Shajari ², A Aulakh ², L Forrester ³, N Partovi ², J Grant ²

Background

Approximately 10% of patients report a history of penicillin allergy. Recent literature suggests cross-reactivity between cephalosporins and penicillins are due to side-chain similarities. Since cefazolin has a unique side-chain from other beta-lactams, it can be safely administered in penicillin allergic patients for surgical prophylaxis. Since October 2018, our hospital updated all surgical prophylaxis pre-printed orders to use cefazolin in penicillin allergic patients, except in those with histories of cefazolin-specific allergy or delayed skin reactions (e.g., Stevens-Johnson syndrome).

Objectives

This study aims to retrospectively determine outcomes and safety of cefazolin as compared with other antibiotics for surgical prophylaxis in penicillin allergic patients with anaphylactic histories prior to implementation of cefazolin pre-printed orders.

Methods

All patients with reported anaphylactic reactions to penicillins prescribed surgical prophylaxis from October 2017 to October 2018 were included. Patients were stratified based on antibiotic received (i.e., cefazolin, clindamycin, vancomycin, other antibiotic) and a retrospective chart review was performed to assess for outcomes and safety.

Results

1,073 prescriptions for prophylactic antibiotics were identified. Of these, 240 cases met inclusion with histories of anaphylaxis to pencillins: 75 (31%) cefazolin, 75 (31%) clindamycin, 39 (16%) vancomycin, and 51 (21%) other antibiotics. General and spinal surgeries used the most cefazolin in penicillin allergic patients, while orthopedics the most clindamycin and thoracics the most vancomycin. Amongst those receiving cefazolin, no critical incidents of allergic reactions were reported and no delays in surgery occurred, as compared with clindamycin and vancomycin.

Conclusions

Cefazolin appears to be a safe option for surgical prophylaxis in patients with history of penicillin anaphylaxis. No differences in incidences of allergic reactions, complications, or surgical delays were reported, as compared with alternate antibiotics. Further larger studies are needed to confirm our findings and determine rates of adverse events associated with the various antibiotic regimens.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):6.

C02. Prescribing and Care by Pharmacists for Uncomplicated Urinary Tract Infections in the Community: Antimicrobial Utilization and Stewardship Results of the RxOUTMAP Study

NP Beahm ¹, DJ Smyth ^2,³, RT Tsuyuki ⁴

Background

Urinary tract infections (UTI) are common infections that often result in suboptimal antibacterial use. In some provinces, pharmacists have the authorization to prescribe medications for the treatment of uncomplicated UTI. Pharmacists are accessible primary care professionals who have an important role to play in antimicrobial stewardship. Our objective was to evaluate the appropriateness of antibacterial prescribing by pharmacists for patients with uncomplicated UTI.

Methods

We conducted a prospective registry trial in 39 community pharmacies across New Brunswick. Adult patients were enrolled if they presented to the pharmacy with either symptoms of UTI with no current antibacterial treatment (Pharmacist-Initial Arm) or if they presented with a prescription for an antibacterial to treat UTI from a physician (Physician-Initial Arm). Pharmacists assessed patients and, if they had complicating factors or red flags for systemic illness or pyelonephritis, they were excluded from the study. Pharmacists either prescribed antibacterial therapy, modified antibacterial therapy, provided education only, or referred to physician, as appropriate. Antibacterial therapy prescribed was compared between the study arms.

Results

There were 750 patients enrolled (87% Pharmacist-Initial Arm). The most commonly prescribed agents in the Pharmacist-Initial Arm were nitrofurantoin (88%), sulfamethoxazole-trimethoprim (TMP-SMX) (8%), and fosfomycin (2%); versus nitrofurantoin (55%), TMP-SMX (26%), and fluoroquinolones (11%) in the Physician-Initial Arm. Nitrofurantoin was prescribed for 5 days in 97% of Pharmacist-Initial orders as compared with Physician-Initial orders, where 65% were for greater than 5 days. TMP-SMX was prescribed for 3 days in 88% of Pharmacist-Initial compared with Physician-Initial, where 63% were for greater than 3 days. Therapy was guideline-concordant in 95% of Pharmacist-Initial compared with 35% of Physician-Initial (p<0.001). For guideline-discordant therapy from physicians, pharmacists prescribed to optimize therapy for 46% of patients.

Conclusions

Treatment was more guideline-concordant when initiated by pharmacists, with longer treatment durations and more fluoroquinolones prescribed by physicians.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):7.

C03. The Urine Culturing Cascade: Variation in Nursing Home Urine Culturing Practices and Association with Antibiotic Use and C. difficile Infection

KA Brown ^1,^2,³, N Daneman ¹, KL Schwartz ¹, B Langford ¹, A McGeer ⁴, J Quirk ¹, C Diong ³, G Garber ¹

Objectives

Rates of antibiotic use vary widely across nursing homes; a large portion of antibiotic prescribing may be due to variation in diagnosis and inappropriate treatment of asymptomatic bacteriuria. We aimed to obtain a complete portrait of urine culturing practices in Ontario and their association with antibiotic use.

Methods

We conducted a retrospective, multilevel evaluation based on quarterly nursing home assessments between April 2014 and January 2017 in 591 nursing homes covering over 90% of nursing home residents in Ontario. Home urine culturing was measured as the proportion of residents with a urine specimen submitted for culture in 14 days before the assessment. Adjustment covariates included: resident age, sex, history of acute care stay, Charlson comorbidity index, functional status, and history of urinary catheterization. Outcomes included receipt of any systemic antibiotic and any urinary antibiotic (defined as ciprofloxacin, norfloxacin, nitrofurantoin, fosfomycin, or trimethoprim and/or sulfonamides) in 30 days and Clostridium difficile infection in 90 days following the assessment.

Results

875,297 quarterly assessments, corresponding to 131,218 unique residents in 591 nursing homes were included. Urine cultures had been submitted in the prior 14 days for 7.9% of assessments. Home-level urine culturing varied substantially (10^th percentile = 3.4%, 90^th percentile = 14.3%). Within 30 days after the assessment, 17.1% of residents received an antibiotic and 5.4% received a urinary antibiotic. Urine culturing prevalence predicted total antibiotic use (adjusted risk ratio for homes at 90^th versus 10^th percentile of urine culturing [aRR]: 2.23; 95% CI: 2.00–2.39), urinary antibiotic use (aRR: 3.22; 95% CI: 2.82–3.86), and C. difficile infection rates (aRR: 2.00; 95% CI: 1.33–3.11).

Conclusions

Homes that submit more urine specimens for culture have significantly higher antibiotic use and more C. difficile infection. These findings suggest the broad applicability of antibiotic stewardship interventions aiming to improve the diagnosis and treatment of urinary tract infection.

Figure C03-1: — Urine culturing and antibiotic use

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):7–8.

C04. A “NIMBLE” Approach to Antimicrobial Stewardship: Nudging in Microbiology Laboratory Evaluation Scoping Review

B Langford ^1,², R Haj ², E Leung ², M Downing ², K Brown ¹, M McIntyre ³, LR Taggart ^2,⁴, LM Matukas ^4,²

Background

Antimicrobial stewardship programs (ASP) are most successful when multifaceted. Opportunities to influence prescriber behaviour by modifying the microbiology report may be promising strategies. Nudging guides decision-making through the strategic placement of architecture of choice, while maintaining prescriber autonomy. The purpose of this scoping review was to determine if nudging strategies in microbiology can improve antimicrobial use.

Methods

A search in the databases: Ovid MEDLINE: Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Ovid MEDLINE® Daily and Ovid MEDLINE, 1946–August 2018, Embase Classic+Embase 1947–August 2018, PsycINFO 1806–August 2018, and All EBM Reviews 2005–August 2018 was conducted. Simulated and vignette studies were excluded. Two independent reviewers were used throughout screening and data extraction.

Results

Of a total of 1,334 citations screened, 13 relevant studies were identified. Study types included pre/post intervention (n = 9), retrospective cohort (n = 3), and randomized controlled trial (n = 1). The majority of studies were performed in acute care settings (n = 11) and the remainder were in primary care (n = 2). Anatomical sites included urinary (n = 5), multiple sites (n = 5), respiratory (n = 2), and blood (n = 1). The majority of studies used a default choice nudging strategy (i.e., selective reporting). All studies reported at least one outcome of antimicrobial use (utilization n = 9, appropriateness n = 5, de-escalation n = 2, cost n = 1). Eleven studies reported a significant change in antimicrobial use with either introducing a nudging strategy or removing the strategy. Four studies evaluated the association of nudging on antimicrobial resistance, with two studies noting improvement in susceptibility rates.

Conclusions

There are a limited number of heterogeneous studies evaluating the impact of microbiology laboratory nudging. Opportunities for further study include identifying the optimal design of a microbiology report, performing prospective studies, evaluating the impact of nudging strategies (desired agents at eye level and framing), and determining the impact of nudging on patient clinical outcomes.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):8–9.

C05. Antimicrobial Use Among Adult Inpatients at Canadian Nosocomial Infection Surveillance Program (CNISP) Hospital Sites Across Canada, 2009 to 2016

W Rudnick ¹, L Pelude ¹, K Abdesselam ¹, M Science ², D Thirion ³, K Amaratunga ¹, J Comeau ⁴, B Dalton ⁵, J Delport ⁶, R Dhami ⁶, J Embree ⁷, Y Émond ⁸, G Evans ⁹, L Forrester ¹⁰, C Frenette ¹¹, S Fryters ¹², G German ¹³, D Gilby ¹, J Grant ¹⁰, D Gravel ¹, K Katz ¹⁴, P Kibsey ¹⁵, J Langley ⁴, B Lee ¹², M Lefebvre ¹⁶, J Leis ¹⁷, H Neville ¹⁸, A McGeer ¹⁹, S McKenna ⁹, A Simor ²⁰, K Slayter ⁴, K Suh ²¹, K Weiss ²², J Conly ^5,²³; CNISP¹

Background

Antimicrobial resistance (AMR) is a serious threat to global public health. Antimicrobial use (AMU) is a major contributor in accelerating the emergence and spread of AMR.

Objective

Identify trends and patterns of AMU in acute-care hospitals in Canada.

Methods

The Canadian Nosocomial Infection Surveillance Program (CNISP) has conducted sentinel surveillance for adult inpatient AMU since 2009. Data are collected on all prescribed J01 systemic antimicrobials as well as oral vancomycin and oral metronidazole. AMU was analyzed from dispensed medications using Defined Daily Doses (DDD) per 1,000 patient-days (pd) as per the World Health Organization. From 2014–2016, AMU was collected by ward type (intensive care unit [ICU] versus non-ICU). Linear regression was used to test for temporal trends.

Results

Between 2009 and 2016, 21–23 hospitals per year participated in surveillance (across 10 provinces). On average, 3.0 million patient-days were included in surveillance annually. Between 2009 and 2016, the total rate of AMU decreased from 645 to 554 DDD/1000 pd (P = 0.02). The rate of fluoroquinolone use decreased from 126 to 71 DDD/1000 pd (P<0.0001) (Figure C05-1). The rate of clindamycin use decreased by 50% between 2009–2012 (~14 DDD/1000 pd) and 2016 (7 DDD/1000 pd; P = 0.003). Between 2014 and 2016, the rate of AMU was higher on ICU wards (1276–1370 DDD/1000 pd) compared with non-ICU wards (461–513 DDD/1000 pd). In 2016, the 5 most commonly used antimicrobials were cefazolin (73 DDD/1000 pd), piperacillin/tazobactam (51 DDD/1000 pd), ceftriaxone (44 DDD/1000 pd), ciprofloxacin (41 DDD/1000 pd), and vancomycin (41 DDD/1000 pd). The top 5 antimicrobials used on ICU wards were piperacillin/tazobactam (177 DDD/1000 pd), vancomycin (173 DDD/1000 pd), cefazolin (140 DDD/1000 pd), ceftriaxone (116 DDD/1000 pd), and meropenem (114 DDD/1000 pd).

Figure C05-1: — Antimicrobial use for the top 8 antibiotic classes/subclasses for adult inpatients at CNISP hospitals across Canada, 2009 to 2016 (represents 74% of total AMU in 2016)

Conclusions

AMU has decreased among CNISP acute-care hospitals most notably among fluoroquinolones. These results support the need for ongoing surveillance of AMU within CNISP to monitor trends and provide Canadian benchmarks.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):9–10.

D01. Identifying Opportunities for Laboratory Stewardship through Prospective Review of Clostridium difficile Testing Indications at a Tertiary Care Hospital

N Reich ¹, M Payne ^1,^2,³, S Champagne ^1,², MG Romney ^1,², M Hinch ⁴, V Leung ^1,^2,^3,⁴, C Lowe ^1,^2,³

Background and Objective

Clostridium difficile infection (CDI) is a leading cause of nosocomial infections in Canada. Molecular testing for C. difficile is highly sensitive but cannot differentiate infection from colonization. Inappropriate test-ordering can lead to increased laboratory costs and unnecessary antibiotic treatment. Our objective was to determine the prevalence of and reasons for inappropriate CDI test-ordering.

Methods

From April 26 to May 28, 2018, hospitalized patients with a CDI order at a tertiary care hospital were prospectively reviewed, capturing clinical history, symptoms, vital signs, laboratory data (e.g. WBC and Cr), alternative reasons for diarrhea, ordering provider, and C. difficile test results. CDI orders were categorized as appropriate or inappropriate based on clinical presentation and identification of alternate reasons for diarrhea.

Results

89 charts were reviewed. The median length of stay prior to CDI test was 3 days. Physicians ordered 68(76.4%) tests and nurses initiated 21(23.6%). The order location was distributed: medicine (51), surgery (12), critical care (14), and emergency (11). Overall, 41(46%) tests were considered appropriate, while 48(54%) tests were assessed as inappropriate (Figure D01-1). Nurse-initiated tests were more likely than physician’s to be inappropriate (85.7% versus 44.1%, p = 0.00083). The two primary aetiologies of inappropriate testing were asymptomatic patient and alternative reason for diarrhea. The alternative reasons for diarrhea were laxative only (16), laxative and NG-feed (4), NG-feed only (7), and chemotherapy (3). The excess, direct laboratory costs associated with unnecessary testing was estimated at $2,347.68 (48 inappropriate tests at $48.91/test) for the 31-day period.

Figure D01-1: — CDI test appropriateness categorization

Conclusions

We identified that 54% of CDI tests ordered were inappropriate; 34% of patients had an alternative reason for diarrhea and 20% of patients tested were asymptomatic. Inappropriate CDI test-ordering was prevalent in our hospital. Diagnostic stewardship interventions to optimize CDI diagnosis should be explored.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):10.

D02. Development of a Tool to Assess the Evidence that Hospital Sinks are a Reservoir for Gram-Negative Bacterial Infection

C Volling ¹, J Johnstone ^1,², D Mertz ^3,⁴, H Maltezou ⁵, R Stuart ^6,⁷, A Jamal ¹, C Kandel ¹, BL Coleman ^1,², A McGeer ^1,²

Background and Objective

Decades of studies suggest that sinks in acute-care facilities are a reservoir for gram-negative bacterial infection. We sought to develop a tool to assess the strength of evidence for causality.

Methods

We adapted the Causal Analysis/Diagnosis Decision Information System (CADDIS) developed by the United States Environmental Protection Agency for use in ecosystems research. We used a modified Delphi process, with input from individuals with expertise in hospital infection epidemiology.

Results

Through 4 rounds of feedback and revision, and 2 tests of tool application to score evidence from published articles, we developed the Modified CADDIS Scoring for Assessing Causality in Studies of Hospital Sinks as a Reservoir for Gram-Negative Bacterial Infection or Colonization. For assessment of inter-rater reliability, 8 reviewers scored 4 randomly chosen articles, across 6 domains: Spatial/Temporal Co-occurrence, Temporal Sequence, Stressor-Response Relationship, Causal Pathway, Evidence of Exposure and Biological Specificity, and Manipulation of Exposure. Mean percent agreement was 69.6, 81.3, 87.5, 54.5, 85.7, and 80.4 per respective domain in Test 1. In Test 2, 6 reviewers scored 4 articles selected to be difficult to assess, with mean percent agreement 83.4, 65, 83.4, 46.7, 83.4, and 50 per domain. There were 16 (20%) cases of 100% agreement, and the Gwet’s AC statistic (adjusting for chance agreement) ranged from 13.4–73.5 (median 61.3). Areas of disagreement were felt to result from lack of a priori knowledge of causal pathways from sinks to patients and uncertain influence of co-interventions to prevent organism acquisition. Modifications were made until consensus was achieved that further iterations would not improve the tool.

Conclusion

Our modified CADDIS Scoring appears to be a promising tool that we are now applying in a systematic review on sink causality in hospital-acquired infection. Similar processes using a modified CADDIS may assist in assessing causality in other infection prevention and control applications.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):10–11.

D03. Candida auris Prevalence in Canadian Acute Care Hospitals, 2018

F Garcia Jeldes ¹, R Mitchell ², A McGeer ³, W Rudnick ², K Amaratunga ², C auris Interest Group ², S Vallabhaneni ⁴, SR Lockhart ⁴, A Bharat ⁵

Objectives

As of July 2018, there were 14 known cases of the emerging drug-resistant yeast Candida auris in Canada. We aimed to estimate the prevalence of C. auris among Canadian patients deemed to be at-risk in order to inform the development of surveillance and infection prevention and control policies.

Methods

High-risk patients from 14 Canadian Nosocomial Infection Surveillance Program (CNISP) and 7 non-CNISP Canadian Hospital Epidemiology Committee (CHEC) hospitals from six provinces were screened for C. auris between September 4, 2018 and November 6, 2018. For each patient, two pooled axilla/groin swabs were collected and a data collection form completed. One swab was inoculated onto chromogenic agar (CandiSelect® or Brilliance Candida®), and the second swab shipped to the U.S. CDC for culture. C. auris was identified by MALDI-TOF and isolates subjected to Illumina whole genome sequencing.

Results

A total of 475 at-risk patients were screened: 35 (7.4%) hospitalized outside Canada; 53 (11.1%) with travel to the Indian subcontinent; 105 (22.1%) colonized with carbapenemase producing organism (CPO), and 282 (59.4%) present on a ward with high antifungal use. Risk groups overlapped, for example, 49% of CPO-colonized patients reported travel to the Indian subcontinent. Two patients were found to be colonized with C. auris (prevalence = 0.42%). Both C. auris-colonized patients were recently hospitalized in India and both patients were co-colonized with NDM-1 and OXA-48-like producing organisms. Whole genome sequencing showed that the isolates were highly related to the South Asian clade, consistent with the patients’ recent overseas hospitalization.

Conclusions

We found the prevalence of C. auris in hospitalized patients in Canada to be low and associated with health care exposure abroad. The rapid emergence of C. auris in the United States and other countries suggests that ongoing surveillance for C. auris colonization in patients recently hospitalized outside Canada would be prudent.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):11.

D04. Carbapenemase-Producing Enterobacteriaceae (CPE) in Household Contacts and Household Environments of CPE-Colonized/Infected Persons

AJ Jamal ¹, L Farooqi ¹, A Faheem ¹, Z Zhong ¹, I Armstrong ², E Borgundvaag ¹, B Coleman ¹, K Green ¹, K Jayasinghe ¹, J Johnstone ³, K Katz ⁴, P Kohler ¹, A Li ¹, R Melano ³, M Muller ⁵, S Nayani ¹, S Patel ³, A Paterson ¹, S Poutanen ¹, A Rebbapragada ⁶, D Richardson ⁷, A Sarabia ⁸, S Shafinaz ¹, A Simor ⁹, B Willey ¹, L Wisely ¹, A McGeer ¹

Objectives

CPE have emerged rapidly worldwide. This prospective cohort study aimed to assess the risk of CPE transmission from colonized patients (index cases, ICs) to household contacts (HCs).

Methods

ICs were identified by population-based surveillance. Groin and rectal swabs, and urine specimens were collected from ICs/HCs at home visits every 3 months; swabs of 10 environmental surfaces (ESs) were also collected. Swabs/specimens were incubated in BHI broth overnight, followed by direct PCR to detect carbapenemase genes and cultures of PCR-positive samples.

Results

95 households and 175 HCs participated. 20 (11%) HCs in 20 (21%) households were CPE-colonized. 8/20 (40%) were CPE-colonized at first visit. The probability of initially non-colonized HCs becoming CPE-colonized by month 12 was 11% (Figure D04-1). Overall, 15/20 (75%) CPE-colonized HCs probably/possibly acquired CPE from the IC; 5 (25%) probably acquired CPE from another source (4 from travel). HCs were more likely to be older, be the IC’s spouse (OR 17; 95% CI: 4.8–63), have an underlying chronic medical illness (OR 4.5; 95% CI: 1.7–12), and have travelled abroad (OR 3.8; 95% CI: 1.2–12). Households with CPE-positive ESs were more likely to have CPE-colonized HCs (OR 7.89; 95% CI: 1.7-36). 299/2887 (10%) ES samples representing 203 unique ESs yielded CPE. CPE yield per ES type was: kitchen sink drain (19%), shower drain (18%), bathroom sink drain (17%), toilet drain (13%), pillow (12%), sofa/chair (10%), bathroom sink tap (9%), toilet handle (6%), telephone (5%), and kitchen sink tap (4%). Of 203 CPE-positive ESs, 70%, 14%, 7%, 2%, and 1% yielded CPE for 1, 2, 3, 4, and 5 visits, respectively. Of the persistently positive samples (positive at ≥3 visits), 14/23 (61%) were from drains.

Figure D04-1: — Cumulative probability of becoming CPE-colonized over time

Conclusions

After 12 months, transmission occurred to 8.6% of HCs of CPE-colonized patients. CPE contamination of household surfaces, particularly drains, is common.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):11–12.

D05. Randomized Controlled Trial of Chlorhexidine Gluconate, Intranasal Mupirocin, Rifampin, and Doxycycline Versus Chlorhexidine Gluconate and Intranasal Mupirocin Alone for the Eradication of Methicillin-Resistant Staphylococcus aureus Colonization: A Preliminary Analysis

L Eum ¹, S Materniak ², P Duffley ³, D Webster ⁴

Background

To prevent recurrent methicillin-resistant Staphylococcus aureus (MRSA) infections, both topical and systemic antibiotic therapies have been studied for decolonization. Clinical equipoise remains with regards to the role of decolonization. Furthermore, the advantages of systemic versus topical decolonization strategies remain unclear.

Objectives

To compare initial MRSA clearance and subsequent MRSA re-colonization rates over a 12-month period, systemic therapy (topical chlorhexidine gluconate, intranasal mupirocin, oral rifampin, and oral doxycycline) was compared with standard topical therapy at our centre’s Ambulatory MRSA Clinic (topical chlorhexidine gluconate and intranasal mupirocin).

Methods

MRSA-colonized patients were randomly assigned (3:1 allocation) to receive either systemic or standard therapy. Follow-up for MRSA screening swabs were obtained at 3, 6, and 12 months after completion of therapy, from the anterior nares and rectum, and where relevant additional specimens from skin lesions, Foley catheter urine (excluding lines), medical device exit sites, and any prior MRSA positive sites. Kaplan-Meier survival curves were calculated using the end date of therapy as time (0) and either the date of next positive MRSA sample or censored at either the last date of follow-up for those lost to follow-up or 12 months from the last day of therapy. Survival curves were assessed for significant differences using log-rank tests.

Results

Of 98 enrolled (73 systemic; 25 standard), 21 patients (16 systemic; 5 standard) were unable to complete the 12-month follow-up. Initial MRSA clearance, defined as 3 sets of negative MRSA cultures immediately following therapy, was achieved in 72.6% (systemic) versus 52.0% (standard). There was a marginally significant difference in the likelihood of remaining MRSA-negative post-therapy (p = 0.043, Figure D05-1). Among those who achieved initial clearance, there was no significant difference in the survival curves over a 12-month period (p = 0.970).

Figure D05-1: — Probability of remaining MRSA-negative post clearance

Conclusion

Standard topical decolonization was similar in efficacy compared with systemic decolonization after a 12-month follow-up.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):12–13.

E01. Emergence of a Novel ST1478 VRE in Canadian Hospitals Associated with Daptomycin Non-Susceptibility and High-Level Gentamicin Resistance

SW Smith ¹, R Mitchell ², K Amaratunga ², J Conly ¹, J Ellison ¹, J Embil ³, S Hota ⁴, J Johnstone ⁵, M McCracken ⁶, G Al-Rawahi ⁷, J Tomlinson ⁸, J Wong ⁵, G Golding ⁶

Objective

Infection with vancomycin-resistant enterococci (VRE) causes significant morbidity and mortality in hospitalized patients with few antimicrobial treatment options. A novel strain of VRE recently described in Australia was non-typeable by MLST methodology due to loss of the pstS gene. We describe the emergence of this strain in Canadian hospitals from 2013–2017 and compare the clinical characteristics of this strain to other VRE strains.

Methods

The Canadian Nosocomial Infection Surveillance Program prospectively collects epidemiologic data on all VRE bloodstream infections (BSI) from 65 hospitals in 10 provinces. Isolates are sent to the National Microbiology Lab for typing and antimicrobial susceptibility testing. Whole genome sequencing (WGS) of ST1478 isolates was carried out to determine interrelatedness.

Results

There were a total of 45 VRE BSI caused by ST1478 in 15 hospitals across six provinces accounting for an increasing proportion of VRE BSI (2.7% in 2013 to 27.6% in 2017). Patients bacteremic with ST1478 VRE were similar with respect to age, sex, ICU admission and mortality to those with non ST1478 VRE. Patients with ST1478 VRE were more likely to have had a solid organ transplant (30.2%) compared with patients with non-ST1478 VRE (11.9%; p = 0.002). There was a higher proportion of daptomycin non-susceptibility in the ST1478 strains (11.1%) compared with non-ST1478 strains (4.5%; p = 0.06). High-level gentamicin resistance was also more common in the ST1478 strains (93.3% versus 10.7%; p<0.001). WGS revealed close genetic relatedness, varying from 0–56 SNVs using AUS004 as the reference strain.

Conclusions

ST1478 VRE appears to be increasing and may be transmissible in Canadian hospitals. This strain is more prevalent in the solid organ transplant population and is associated with non-susceptibility to daptomycin and high-level gentamicin resistance. Further investigation is required to understand the emergence and transmission dynamics of this novel strain.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):13.

E02. Burkholderia cenocepacia ET12 Transmission in Adults with Cystic Fibrosis

A Blanchard ¹, L Tang ², M Tadros ³, M Muller ⁴, T Spilker ⁵, V Waters ¹, J LiPuma ⁵, E Tullis ⁶

Background

Burkholderia cenocepacia is a well-known pathogen in the cystic fibrosis (CF) population. Despite enhanced infection prevention and control (IPAC) precautions, 11 CF patients have been identified as having incident B. cenocepacia ET12 strain (ET12-Bc) infection since 2008 at the Toronto Adult CF centre.

Objectives

The objective of this study was to describe the investigation of ET12-Bc acquisition in CF patients in this centre.

Methods

A retrospective chart review of 11 cases of incident ET12-Bc infection was performed at St. Michael’s Hospital, in Toronto, Canada. Patient demographic and clinical characteristics were reviewed and an epidemiologic investigation was conducted. ET12-Bc isolates were analyzed by multilocus sequence typing (MLST) and whole genome sequencing (WGS).

Results

Ten patients had a hospital admission within the two months preceding their first ET12-Bc–positive sputum culture; ET12-Bc was detected 12 months following hospital admission in the eleventh patient. In all but one isolate, the seven MLST loci had a 100% DNA sequence match to these loci in ET12-Bc strain J2315, which represents Burkholderia MLST sequence type 28. The remaining isolate had a single nucleic acid polymorphism in one MLST locus. Epidemiologic investigation, together with WGS analysis, suggested that transmission occurred between patients during hospitalization in 8 of 11 patients. To date, 10 of 11 patients with new acquisitions have died (median survival of 379.5 days).

Conclusions

We identified ET12-Bc nosocomial transmission between CF patients in hospital, despite enhanced IPAC precautions. This led to a change in our hospital policy wherein ET12-Bc–positive patients are no longer cared for on the same unit as ET12-Bc–negative patients with CF.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):14–15.

E03. Still Rising: A Quasi-Experimental Study of Vancomycin Resistant Enterococcus (VRE) Bacteremia Rates in Ontario, 2009–2018

J Johnstone ¹, E Shing ¹, A Saedi ¹, K Adomako ¹, E Truong ¹, Y Li ¹, K Brown ¹, G Garber ¹

Objective

In June 2012, some Ontario hospitals discontinued VRE screening and isolation programs; additional hospitals have since followed. In 2015, we observed a significant increase in the incidence of VRE bacteremias in hospitals that had stopped screening. The aim of this study was to update these results through to 2018.

Methods

All Ontario hospitals are mandated to publicly report VRE bacteremias on a quarterly basis and we used these data for the study period January 2009 to September 2018. An interrupted time series Poisson regression was used to assess the slope change in VRE bacteremia incidence rate (primary outcome) after versus before the discontinuation of screening. Hospitals that continued to screen were the comparison group. Incidence rates were adjusted for hospital type and clustering within hospital site; slope changes were presented as incidence rate ratios (IRR) with 95% confidence intervals (CIs).

Results

97 hospital facilities publicly reported at least 1 VRE bacteremia during the study period. Overall, rates of VRE bacteremia increased from 0.74/100,000 in 2009 to 3.54/100,000 in 2018. In ceased-screening hospitals (n = 23), there was an increase in slope after screening was discontinued when compared with before (slope change IRR 1.44 [95% CI: 1.22–1.71]) (Figure E03-1). In screening hospitals (n = 74), the slope was not significantly different after June 2012 compared with before (slope change IRR 1.02 [95% CI: 0.70–1.50]). The results were similar when restricted to acute teaching hospitals; ceased-screening hospitals (n = 11, slope change IRR 1.77 [95% CI: 1.44–2.18]) compared with screening hospitals (n = 10, slope change IRR 0.83 [95% CI: 0.46–1.52]).

Figure E03-1: — VRE bacteremia rates in Ontario, January 2009 to September 2018

Conclusions

Since 2009, there has been over a 4-fold rise in the rate of VRE bacteremias in Ontario. The association between discontinuation of VRE screening and isolation and rates of VRE bacteremia is impressive and persistent. Hospitals aiming to minimize VRE bacteremias should consider VRE screening and isolation programs.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):15.

E04. Active Charcoal Filters in Hospital Ice Machines as a Source for Contamination of Mycobacterium fortuitum: Update of a Pseudo-Outbreak

N Sant ^1,^2,³, J Gosal ^1,^2,³, I Gorn ¹, K Suh ^2,³, M Desjardins ^1,^2,³

Background

We report a follow-up to a nosocomial pseudo-outbreak of respiratory tract colonization with Mycobacterium fortuitum believed to have originated from a contaminated ice machines. We wanted to further delineate the source of contamination and our hypothesis was that the charcoal filters were the ultimate source.

Methods

Ice machines from the affected hospital campus and unaffected campus were compared in terms of type of filtration system. We increased the scope of our investigation to include environmental samples collected from different hospital sites as well as additional sites in the ice machines. Environmental sampling was repeated after removal of the charcoal filters. Specimens were cultured with use of a continuously monitored broth system for the isolation of mycobacteria. Samples positive for mycobacteria were sent to the regional Public Health Laboratory for identification or by MALDI-TOF MS once implemented in our laboratory. Environmental isolates were compared by genotyping using ERIC-PCR. Data regarding the water quality including pH, temperature, turbidity and anionic makeup between hospital campuses was collected by PHL representatives.

Results

The affected campus ice machines involved a Pentek GS-10ALS filtration system which was not included in the ice machines of the unaffected campus. When sampling the ice machines, M. fortuitum was consistently isolated from the water sampled after charcoal filtration. Of note some ice machines remained positive other species of non-tuberculosis Mycobacterium. Repeat sampling from the ice machines post cleaning and filter removal were negative for M. fortuitum. Further investigations will involve measuring M. fortuitum rates in our patient population after filtration removal.

Conclusions

Water treatment using a charcoal filtration lead to a pseudo-outbreak of M. fortuitum. Active charcoal filters may allow for M. fortuitum biofilm formation and contamination of the water supply.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):15.

E05. The Economic Burden of Vancomycin Resistant Enterococcus (VRE) Bacteremia: A Population-Based Matched Cohort Study

J Johnstone ¹, C Chen ¹, E Shing ¹, K Adomako ¹, G Garber ¹, B Sander ²

Objective

To evaluate the costs attributable to VRE bacteremia from the health care payer perspective.

Methods

We conducted a population-based matched cohort study of hospitalized patients with confirmed VRE bacteremia in Ontario, Canada, between January 2009 and December 2013. Infected (i.e., exposed) patients were identified and hard-matched to up to three unexposed patients based on age, sex, comorbidities, admission date, rurality, neighborhood income, hospital type and pre-hospitalization resource utilization. All patients were followed until December 31, 2017. The primary study outcomes were costs up to 1-year post index date (C$ 2014).

Results

We identified 217 exposed patients and in preliminary analysis, we matched 127 exposed to 344 unexposed patients. In the exposed group, mean age was 62.9 years (SD 17.1), 40% were female, and common comorbidities were cancer (33%), heart disease (21%), and renal failure on hemodialysis (10%). In the unexposed group, mean age was 62.8 years (SD 16.5), 39% were female, and common comorbidities included cancer (24%), heart disease (4%) and renal failure on hemodialysis (5%). Length of stay was 63 days for exposed versus 8 days for unexposed patients. Mortality within 30 days was 19% for exposed patients 68% within 1 year, and 83% within the follow-up period, versus 7%, 17% and 35% for unexposed patients, respectively. Approximately half (53%) of exposed patients died during index hospitalizations versus 6% of unexposed patients. Mean cost for index hospitalization was $134,542.69 for exposed versus $10,838.46 for unexposed patients. Mean unadjusted costs for exposed patients at 30 days, and 1 year were $63,762.56 and $190,931.86 versus mean unadjusted costs for unexposed patients of $17,722.31 and $51,153.35.

Conclusion

VRE bacteremia is associated with increased health care costs, extending well beyond the index hospitalization.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):16.

F01. Metronidazole-Associated Neurologic Toxicity: A Nested-Case Control Study

N Daneman ¹, Y Cheng ¹, T Gomes ¹, J Guan ¹, M Mamdani ¹, F Rahman ¹, D Juurlink ¹

Objectives

There have been numerous case reports of peripheral and central nervous system toxicity among metronidazole users. We sought to confirm an association of metronidazole with neurologic toxicity, to quantify the magnitude of these risks, and to assess whether these phenomena are dose dependent.

Methods

We conducted a nested case control study in Ontario from April 1, 2003 to March 31, 2017. The cohort nest included all older adults age ≥ 66 years, who received at least one outpatient prescription of metronidazole or clindamycin, without a history of neurologic disorders or recent prolonged hospitalizations prior to index date. The cases were patients who experienced new cerebellar dysfunction, encephalopathy, or peripheral neuropathy, as identified in province-wide emergency department and hospital datasets. Ten control patients were matched to each case based on age, sex and recent hospital exposure. The primary exposure of interest was metronidazole treatment in the 100 days preceding the index date. Conditional logistic regression was used to test the association of metronidazole (versus clindamycin) exposure with neurologic toxicity.

Results

Among 140,556 older Ontarians with incident outcomes of neurologic dysfunction, there were 1,212 cases with metronidazole or clindamycin exposure, but not both, in the 100 days preceding the outcome event; these patients were matched to 12,098 controls. Metronidazole use was associated with increased neurologic toxicity compared with clindamycin (OR: 1.72; 95% CI: 1.53–1.94), which persisted after accounting for patient demographics, comorbidities, and other medication exposures (adjusted OR [aOR] 1.43; 95% CI 1.26–1.63). The increased neurologic toxicity was detected among patients with low, medium, and high cumulative doses of metronidazole.

Conclusions

Metronidazole is associated with a statistically significantly increased risk of peripheral and central nervous system toxicity; clinicians and patients should be aware of these rare but serious adverse events.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):16.

F02. Cutaneous Leishmaniasis: A 10 Year Experience in a Canadian Reference Centre for Tropical Diseases

A Lemieux ¹, F Lagacé ², K Billick ^3,⁴, M Libman ^4,⁵, S Barkati ^4,⁵

Background and Objectives

Cutaneous leishmaniasis (CL) is one of the “Neglected Tropical Diseases.” Migration and travel has led to increasingly encountered CL cases in non-endemic country such as Canada. The objectives are to describe the epidemiology, clinical presentations, diagnostic methods and treatments used for CL in our centre. Clinical differences between Old World (OW) and New World (NW) CL were evaluated as well as the sensitivities of the diagnostic methods.

Methods

A retrospective observational study was performed with all the laboratory confirmed diagnoses of CL between January 2008 and October 2018. Demographic, epidemiologic and clinical characteristics, diagnostic methods, treatments and outcomes were abstracted from computerized patient records. Clinical response was determined based on the degree of re-epithelialization at 1 year after treatment.

Results

A total of 52 cases were identified, clinical data was available for 48 cases (NW: 33, OW: 15; average age: 41 years (range: 1–75); male: 28 [58%]). Lower extremities (n = 15 [31%]) and face/neck (n = 13 [27%]) were the most common locations. At initial presentation, NW CL presented more often as ulcers (n = 28 [85%]) compared with OW CL that mostly presented as plaques (n = 9 [60%]); p = 0,006. Adenopathy was seen in 9 (27%) NW CL but not in OW CL (p = 0,0248). PCR had the highest estimated sensitivity with 98% compared with 68% for smear, 64% for histopathology and 65% for culture. Among all cases, 38 patients had documented treatment, 36 of them (95%) received systemic treatment as first-line. Liposomal amphotericin B was the most commonly used in 20/38 (53%). Of all CL cases, clinical response was achieved in 32 patients (67%).

Conclusions

This study represents the largest Canadian case series of CL. Non-endemic countries encounter a diversity of CL species. PCR is an essential tool for diagnosis and treatment leads to complete re-epithelialisation in most cases.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):17.

F031. Management of Mycobacterium tuberculosis Prosthetic Joint Infection: A Case Report and Literature Review

CK Lo ¹, EW Wilson ²

Background

Mycobacterium tuberculosis prosthetic joint infection (TBPJI) is a rare complication but can be seen in immunocompromised patients or those at-risk of tuberculosis (TB). Lacking clinical suspicion and experience with TBPJI often lead to delays in diagnosis. We report a case of left hip TBPJI in a Hungarian-Canadian immigrant, treated with concurrent surgical and medical therapy. We also performed a literature review on TBPJI case reports outlining their diagnosis and management.

Methods

A comprehensive search was conducted on English language literature published from 1980 to July 2018 on TBPJI, using EMBASE, OvidMEDLINE, PubMed, and Google Scholar. Data analysis focused on patient demographics (age/gender, risk factors), time elapsed from arthroplasty to symptom onset, diagnosis, therapy, and outcomes.

Results

Our literature review identified 53 cases of TBPJI from 38 published articles, most involving hip or knee infections. Symptom onset from arthroplasty ranged from 2 weeks to 38 years. Delays in diagnosis up to 3 years were reported, often after failing empiric antibiotics and/or repeated surgeries with no bacterial growth from intraoperative cultures. Diagnosis was confirmed via fluid/tissue acid-fast bacillus testing, TB culture/PCR, or pathology. Except for 2 cases, all patients received anti-mycobacterial therapy (AMT). AMT drug combinations and duration varied greatly. Eleven cases received AMT alone, while most required surgery (13 with DAIR [Debridement-Antibiotics-Implant-Retention], 18 staged revisions, 11 hardware removals). Of the 11 medically managed cases, 2 were declined surgery. Our case was unique as the surgeon suspected an infected native joint intraoperatively and opted for cemented spacer, until diagnosis of TBPJI was made 3 months later, where she successfully responded to 12 months of AMT followed by second-stage revision.

Conclusion

Though medical management alone is possible our literature review and experience recommend managing TBPJI with both AMT and surgical consultation/intervention as a safer option.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):17.

F04. Hit Or Miss? Comparison of Neisseria gonorrhoeae Detection Limits of Automated Molecular Methods used Across Nova Scotia

C Phillips ¹, D Haldane ^2,³, T Mailman ^2,^3,⁴, Y Hussein ², D Fontaine ², TF Hatchette ^2,³, JJ LeBlanc ^2,³

Objectives

Testing for Neisseria gonorrhoeae (GC) is often performed concurrently with Chlamydia trachomatis (CT). Due to the low prevalence of GC in Nova Scotia, confirmatory testing is reflexively performed following an initial positive (or indeterminate) GC result. Given the different commercial assays are used for these purposes in NS, the analytical sensitivity of methods used for GC testing was assessed.

Methods

Panels were sent to each NS laboratory performing CT/GC testing consisting of 10-fold serial dilutions of GC, with the final dilution in the recommended buffers for each method. GC screening methods included the Hologic Panther, BD Viper, BD Max, and Roche 4800 systems. GC confirmatory testing was evaluated on the Hologic Panther system, BD Max, or Cepheid Xpert. To assess analytical specificity, non-GC strains of N. meningitidis (serotypes B and C), N. lactamica, N. cineria, N. sicca, and N. subflava were also tested at high concentrations (>10⁷ CFU/mL).

Results

In the specificity analyses, the BD Viper was the only assay that showed cross reactivity (with N. lactamica). In the analytical sensitivity analyses, the Hologic Panther and BD Viper systems were the most sensitive screening methods for GC with LoD of 11±1 CFU/mL, followed by the BD Max at 27±4, and the least sensitive method was the Roche 4800 at 295±37 CFU/mL. GC confirmation on the Hologic Panther had equivalent sensitivity to the primary testing method, whereas testing algorithms using BD Max or Xpert for GC confirmation were less sensitive than their screening method at 27±4 and 31±4 CFU/mL, respectively.

Conclusions

The most sensitive methods for GC detection and confirmation were the Hologic Panther assays. Given the lower sensitivity of other methods used in Nova Scotia for GC screening and confirmation, testing algorithms may need to be revised.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):18.

F05. Understanding Why PrEP Uptake is Low Among the Most At Risk Individuals

B Goodall ¹, J Boudreau ², N Mishreky ², T Brothers ², F Gallagher ², R Booth ², E Mailman ¹, L Barrett ^1,^2,³

Background

In July 2018, there were 25 new HIV seroconversions in 6 months compared with 15 in 2017. These infections occurred in people who use drugs, within a discrete geographic area. PrEP (Pre-exposure prophylaxis) has been publicly funded in NS since July 2018, however there has been no uptake among the most at risk individuals. The barriers to access were unclear. Our objective is to determine self-perceived risk of HIV transmission, awareness of prevention methods, and PrEP knowledge in a high incidence region.

Methods

Using an ethics-approved questionnaire, 210 individuals were surveyed. Questionnaires were collected from pedestrians within the geographic area of the outbreak. Participants were asked “Yes” or “No” questions about their (1) awareness of PrEP, (2) self-perceived risk of contracting HIV, (3) knowledge of HIV prevention methods, and (4) willingness to use a daily medication to reduce HIV risk. The aggregate questionnaire responses were reported as percent answering “Yes” or “No” to each question.

Results

40/210 individuals self-identified as “at-risk” for HIV infection, 100% (40/40) of these individuals were aware of an HIV prevention method, 52% (21/40) were aware of PrEP, and 90% (36/40) would consider PrEP if told they were at high risk of HIV infection. 170/210 individuals self-reported as “not-at-risk” for HIV infection, 94% (160/170) were aware of an HIV preventative method, 29% (49/170) were aware of PrEP, and 81% (138/170) would consider PrEP if told they were at high risk of HIV infection.

Conclusions

Less than 20% of individuals surveyed in this high incidence area consider themselves at risk of HIV infection. Willingness to use a daily medication to reduce the risk of contracting HIV was high among all participants, however awareness of PrEP was low. These findings support efforts to increase harm reduction, as well as specific community awareness and PrEP prescribers.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):18–19.

G01. Assessing Adolescent Immunization Options for Pertussis in Canada: A Cost-Utility Analysis

K Anyiwe ^1,², M Richardson ^1,², J Brophy ^3,⁴, O Baclic ⁵, B Sander ^1,^2,^6,⁷

Objective

In Canada, adolescent pertussis vaccination helps prevent transmission. Trade-offs with respect to disease prevention and effectiveness can be associated with immunization timing. The objective of this study is to assess the cost-utility of different adolescent pertussis immunization strategies.

Methods

A cost-utility analysis was conducted using a Markov model, with adolescents (beginning at age 10 years) as the cohort of interest. The model assessed three vaccination strategies: 1) immunization of 10-year-olds, 2) removal of adolescent vaccination, or 3) immunization of 14-year-olds (status quo comparator). The analysis was conducted from a health care payer perspective and used a lifetime time horizon. Primary outcomes included life years, quality-adjusted life years (QALYs), health system costs, and incremental cost-effectiveness ratio (ICER). Costs and outcomes were discounted at 1.5% annually. Deterministic and probabilistic sensitivity analysis was conducted to assess parameter uncertainty.

Results

The current recommended adolescent immunization strategy (at 14 years old) resulted in an average of 40.4432 expected QALYs at a cost of $34.36 per individual. This strategy was dominated by immunization at 10 years and no immunization. Compared with no immunization, immunizing at 10 years of age had an ICER of $108,703.01 per QALY. Results were robust across a series of deterministic and probabilistic sensitivity analyses; findings were most sensitive to infection probability, vaccine cost, vaccine effectiveness, probability of mortality, and cost of inpatient care. At a cost-effectiveness threshold of $50,000/QALY, removal of the adolescent vaccine represented the most cost-effective strategy in 97% of simulations. However, at a threshold of $100,000/QALY, immunization at 10 years of age is marginally cost-effective relative to no immunization, with a 58% probability of being cost-effective.

Conclusion

Findings suggest that alternatives to the current Canadian adolescent pertussis vaccine schedule—especially no immunization—are more cost-effective relative to current immunization of 14-year-olds.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):19.

G02. The Lower Saint-Lawrence River of Québec, a Hot Spot for Sheepfold-Associated Q Fever (Coxiella burnetii Infection) in Canada: Review of 254 Cases

A de Beaumont-Dupont ^1,², P Jutras ¹, P Dolcé ¹, MA Rosca ¹, D Bolduc ³

Background

The Lower Saint-Lawrence region (LSLRR) has a 15-fold higher incidence Q fever (Coxiella burnetii infection) compared with the Québec provincial rate, frequently associated with sheepfolds. This study aims to review the Q fever clinical cases in the LSLRR.

Methods

All Q fever cases were retrieved from the microbiology logs, the medical records of Rimouski hospital (250-bed acute-care community centre) and Public Health records between 1991 and 2018. Confirmed acute cases included positive PCR, antibody titers greater than four-fold rise (CF or IFA). Probable acute cases included titers >1/40 CF or >1/128 IFA IgG phase 2. Chronic cases had positive PCR, antibody titers >1/320 CF or >1/1024 IFA IgG phase 1. Data were analyzed using EPI-INFO 7.2.2.6.

Results

Of 295 screened cases, 243 were acute (239 confirmed, 4 probable), 11 chronic. Median age was 46 years (range 3–84), 75% were male. For acute Q fever, prominent symptoms were fever (99%), headache (81%), chills (79%), sweating (72%), myalgia (68%), fatigue (66%). Acute cases included hepatitis 83%, pneumonia 4% and endocarditis 1%. antimicrobial was provided in 92%, mostly doxycycline (93%). Seasonal peak was observed from May to July, with 56% of acute cases. Chronic cases included 4 hepatitis, 3 endocarditis, and 1 aortitis. Comorbidity was observed in 22% of patients and 37% were hospitalized. Among the 8 counties in LSLRR, most cases (56%) were within the 2 counties with higher ovine production. Exposure to sheepfold was prominent 81%, including 34% shepherds, 20% sheepfold visitors, and 37% indirect.

Conclusions

To our knowledge, this is the largest retrospective study of Q fever in Canada. Fever with hepatitis were the most common manifestation of C. burnetii in the Québec LSLRR. Most patients (81%) were exposed to sheepfold. Protective measures should be implemented in the sheep industry to reduce Q fever in our region.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):19–20.

G03. Schistosoma and Strongyloides Screening in Immigrants as Part of HIV Care in Alberta

J McLellan ¹, MJ Gill ², S Vaughan ², B Meatherall ²

Background

People who previously lived in regions endemic to Schistosoma and Strongyloides are at high risk of chronic infections, even when they have immigrated to non-endemic regions. These parasitic infections can have serious and sometimes fatal consequences for people co-infected with HIV. While screening guidelines and data exist pertaining to parasitic prevalence for immigrant populations, these do not exist for HIV-positive populations.

Objectives

Assess the prevalence of chronic parasitic infections in immigrant/refugee HIV-positive individuals and identify epidemiologic and laboratory characteristics to enable more focused screening.

Methods

243 HIV-positive individuals, born outside of Canada and receiving care at a centralized HIV clinic in Alberta between 2015 and 2018 were screened for schistosomiasis and strongyloidiasis using serology and stool analysis. Epidemiologic and laboratory values were analyzed using univariate logistic regression.

Results

Defined by serology, the prevalence of schistosomiasis was 19.9% and strongyloidiasis 4.8%. Stool microscopy identified no Schistosoma or Strongyloides parasites. Age between 40 and 50 years (OR 2.50; 95% CI: 1.13–5.50), being a refugee (OR 3.55; 95% CI: 1.72–7.33), country of origin within Africa (OR 10.11; 95% CI: 1.25–82.00) or within East Africa (OR 14.93; 95% CI: 1.94–114.89), eosinophilia (OR 3.56; 95% CI: 1.25–10.16) and CD4 count less than 200 cells/mm³ (OR 2.46; 95% CI: 1.017–5.92) were associated with positive Schistosoma serology. Eosinophilia was associated with positive Strongyloides serology (OR 11.57; 95% CI: 2.81–47.65). Eosinophilia had poor sensitivity for identification of chronic parasitic infection.

Conclusion

In the HIV positive immigrant population, schistosomiasis and strongyloidiasis are found at a similar prevalence as in the refugee population and at significantly high rates to warrant targeted screening with serology. Schistosoma serology should be performed for all HIV-positive individuals originating from Africa. Stool microscopy and eosinophil counts are not useful for parasitic screening in this context.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):20.

G04. Decreasing Unnecessary Antibiotics for Clostridium difficile Colonization with a Nudge from Microbiology Reporting

D Herman ¹, A Sarabia ^2,³, H Chan ³, CM Graham ^2,³

Objective

Patients who test positive for Clostridium difficile by polymerase chain reaction (PCR) testing, with a negative toxin immunoassay (EIA) are commonly colonized and do not require treatment, yet clinicians often treat in the setting of a “positive” result. We evaluated the clinical impact of a nudge, changing from reports that included both assay results along with treatment recommendations, to the current “Clostridium difficile organism present but toxin not detected by EIA. Consider C. difficile colonization or early infection.”

Methods

We conducted a retrospective cohort study of all adult patients admitted to our multisite community hospital with a positive C. difficile PCR result and negative toxin EIA from January 1, 2016–June 30, 2018, including 15 months pre- and post-intervention. Patients dying within 24 hours of testing were excluded. Recurrent episodes were included if they were off treatment and >7 days from initial testing. The primary outcome was total days of therapy (DOT) for TID metronidazole, oral vancomycin and fidaxomicin and statistical process control charting determined special cause variation. Secondary outcomes included subsequent toxin positive disease (TPD), colectomy, all-cause mortality, and length of stay (LOS).

Results

369 episodes were identified in total, 169 occurring after the intervention. Mean DOTs/episode decreased from 13.7 to 7.7 (p<0.05) post-intervention, with SPC indicating special cause variation. Patients who did not receive any treatment increased from 6.0% to 23.7% post intervention (P<0.05). Pre and post-intervention, no significant change in adverse outcomes including subsequent TPD (9.9% versus 6.5%), colectomy (0% versus 0.6%), or mortality (8% versus 11.8%) occurred (all p values >0.05). LOS was statistically unchanged at 52 days pre-intervention and 43 days post-intervention (P>0.05).

Conclusions

A significant decrease in antibiotic utilization and exposure occurred after introduction of a reporting nudge raised the possibility of C. difficile colonization, without increasing subsequent disease.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):20–21.

G05. Recurrent Endocarditis in Persons Who Inject Drugs

L Rodger ¹, M Shah ¹, E Shojaei ², S Koivu ¹, M Silverman ^1,²

Objective

The incidence of infective endocarditis (IE) and recurrent IE is increasing among persons who inject drugs (PWID); however, literature studying recurrent IE in PWID is limited. Our objectives were to understand and compare the microbial etiology, clinical characteristics and variables associated with mortality between initial and recurrent IE episodes in PWID.

Methods

A retrospective cohort study based on chart review was conducted between February 2007–March 2016. We included adult inpatients (>18) at tertiary care centers in London, Ontario who met a diagnosis of definite IE based on the Modified Duke’s Criteria.

Results

390 patients had endocarditis with 212/390 in PWID. 68/212 (32%) PWID had a second episode with 28/212 (12%) having additional recurrences. Second episode IE is more common in PWID (14/179 [6.2%] versus 68/212 [24.3%]; p<0.001) with injection drug use increasing the risk of second episodes more than fourfold (RR 4.46; 95% CI: 2.21–8.99; p<0.001). There are few clinical differences between first and second IE episodes in PWID; however, the microbial etiology varies. There are significantly less Staphylococcus aureus infections (165/212 [78%] versus 43/68 [63%]; p = 0.03) and a wider variety of infectious organisms are seen causing second episode IE (Figure G05-1). In particular, fungal IE is more common in second episodes (1/212, 0.5% versus 5/68, 7.4%, p = 0.002). Additionally, fungal infection in recurrent endocarditis was associated with increased mortality in PWID with an adjusted HR of 4.43 (95% CI: 1.27–15.5; p = 0.02).

Figure G05-1: — Microbiology of (a) first episode and (b) second episode of endocarditis in PWID

Conclusions

PWID are at significantly higher risk of recurrent IE. Fungal endocarditis is more common in recurrent endocarditis and is associated with increased mortality, suggesting that providers should consider empiric antifungal therapy in PWID with suspected IE and a history of previous IE.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):21.

G06. Cost Analysis of Outpatient IV Clinic Daptomycin use Compared to Inpatient Vancomycin use in the Treatment of Patients with Resistant Gram-Positive Infections

Ap Sonpar ^1,², S Fryters ^2,³, A Chandran ^3,⁴, I Chiu ^2,⁵

Objectives

Assess the most cost-effective method of delivering treatment to patients with resistant gram-positive infections who are not eligible for Home Parenteral Therapy (HPT) and identify patient factors leading to barriers to discharge with HPT.

Methods

A six-month retrospective chart review at an inner city, tertiary care hospital in Edmonton, AB, identified patients receiving daptomycin as an inpatient or outpatient for serious MRSA or VRE infections. A cost analysis was done for the patients with serious MRSA infections who either started or completed their course of daptomycin therapy in the outpatient antimicrobial therapy (OPAT) clinic because they were not eligible for HPT with vancomycin due to no fixed address, no refrigeration in home, no medication coverage, or were denied by the HPT Program for other reasons. Hospital stay, OPAT visit, antibiotic drug (based on an 80 kg patient with normal renal and liver function) and administration costs were provided by Alberta Health Services.

Results

21 patients started or completed their daptomycin treatment courses in the OPAT clinic. Seven patients (33.3%) received daptomycin without an appropriate indication and were excluded from the cost analysis. Four patients (19.0%) received daptomycin due to allergy to vancomycin or infection with vancomycin resistant organisms. Ten patients (47.6%) received daptomycin via OPAT as they could not be discharged on vancomycin via HPT. Compared with inpatient vancomycin, using daptomycin in OPAT resulted in a 45% cost savings and using vancomycin in OPAT would have resulted in a 20% cost savings for these 10 patients.

Conclusions

OPAT clinic daptomycin use is a cost-effective way to facilitate discharge for patients requiring vancomycin who are not eligible for HPT.

Table G06-1:

Cost of therapies for resistant gram-positive infections

	Cost (Canadian dollars)
Vancomycin (inpatient)	131,965
Vancomycin (IV clinic)	98,916
Daptomycin (IV clinic)	72,630

Open in a new tab

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):22.

H01. The Ontario Program to Improve Antibiotic Use (OPTIMISE): Defining appropriate Antibiotic Prescribing in Primary Care—A Modified Delphi Panel Approach

J Wu ¹, R Ha ¹, B Langford ¹, G Garber ², N Daneman ¹, J Johnstone ¹, W McIsaac ³, S Sharpe ⁴, K Tu ⁵, K Schwartz ¹

Background

More than 90% of antibiotic use in Canada occurs outside hospital settings; however, there is currently no consensus or benchmark defining appropriate antibiotic use in Canadian primary care settings. The objective of this study was to define expected appropriate outpatient antibiotic prescribing rates for 23 clinical conditions using a modified Delphi method.

Method

An initial online questionnaire on percentage of visits for which an antibiotic should be prescribed for each condition was sent to a nine-member multidisciplinary expert panel; including community and academic family physicians, adult and paediatric infectious disease physicians, and antimicrobial stewardship pharmacists. Conditions that did not reach consensus were discussed during face-to-face meetings with anonymous voting. This process was repeated until 100% consensus was reached.

Results

Three of 69 clinical scenarios (23 conditions by three age groups) reached consensus online. The remaining 66 were discussed face-to-face. The average number of rounds required to reach consensus was 2.6 (min: 1; max: 5). Appropriateness rates, some of which differed by age groups (i.e., <2, 2–18, >18 years) where appropriate, were: pneumonia (100%); pyelonephritis (100%); non-purulent skin and soft tissue infections (SSTI) (100%); other bacterial infections (100%); reproductive tract infections (100%); urinary tract infections (95%–100%); prostatitis (95%); epididymo-orchitis (85%–88%); chronic obstructive pulmonary disease (50%); purulent SSTI (35%–50%); otitis media (30%–40%); pharyngitis (18%–40%); acute sinusitis (18%–20%); chronic sinusitis (14%); bronchitis (5%–8%); gastroenteritis (4%–5%); dental infections (4%); eye infections (1%); otitis externa (0%–1%); asthma (0%), common cold (0%), influenza and other non-bacterial infections (0%).

Conclusions

Using a rigorous method, this study defined the levels of appropriate antibiotic prescribing among community primary care providers. These results can be applied to community antimicrobial stewardship programs to define the level of inappropriate use and therefore set targets to optimize outpatient antibiotic use.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):22–23.

H02. Marked Reduction in Urine Cultures Collected from Long-Term Care Residents After Targeted use of Intensive Education and a Long-Term Care Urinary Tract Infection Medical Directive

GJ German ^1,², S Lutes ¹, K MacPhee ³, J Boswell ^1,⁴

Background

Asymptomatic bacteriuria is exceedingly common in the elderly population, particularly in residents of long-term care (LTC) facilities. Urine cultures are frequently ordered leading to unnecessary and often harmful treatment. Our primary objective was to determine if there was a reduction in urine cultures before and after implementing two different targeted strategies at LTC sites in the province. The first strategy was a medical directive at two public LTC sites, the second strategy being intensive education to private LTC sites.

Methods

A Urinary Tract Infection (UTI) Medical Directive was implemented at one of 9 public LTC sites in March/April 2016 and at the second site in September 2017. Intensive education on asymptomatic bacteriuria was given at 5 of 9 private facilities and to all of the private LTC directors starting in November 2016 and finishing in April 2017. All urine cultures collected were evaluated from January 2015 to September 2018 using the Cerner laboratory information system.

Results

The first public site saw a reduction from an average of 6.5 to 2.2 urine cultures per month (66% reduction), while the second site showed a similar decrease from 21.0 to 11.3 cultures per month (54% reduction). Similarly, there was a 60% reduction in urine cultures observed at 6 months across all private LTC sites, with a retained decrease of 50% noted at 12 months post-education. In comparison, the public LTC sites as a whole did not see a noticeable difference in urine cultures collected during the same 32-month period.

Conclusions

Both a nursing-led UTI Medical Directive and targeted educational efforts resulted in a marked reduction in urine cultures sent for lab testing from LTC facilities. Facilities not receiving either of these two interventions did not see a change. Future analysis will include studying the potential impact on antibiotic microbial use and resistance.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):23.

H03. Evaluating the Impact of Prospective Audit and Feedback on the Use of Clindamycin and Quinolones in Medicine Clinical Teaching Units

I Karsan ¹, S Elsayed ^1,², Z Popovski ², R Dhami ^1,^2,³

Background

Antimicrobial usage of quinolones and clindamycin was noted to be greater on the general medicine units of a large academic health care institution. The Antimicrobial Stewardship Program (ASP) implemented a prospective audit and feedback (PAF) strategy to optimize prescribing of these agents given their propensity to cause Clostridium difficile infection (CDI). This study evaluated the impact of PAF interventions on quinolone and clindamycin use and the incidence of CDI on clinical teaching units (CTU).

Methods

A PAF strategy was introduced in April 2015 at a large two-site academic health care institution housing six Clinical Teaching Units for general medicine patients. Using a face-to-face PAF model with medical teams (senior medical resident and team pharmacist), the ASP conducted twice daily reviews of all patients receiving antimicrobials. Clindamycin and quinolone utilization were compared using defined daily doses (DDD) per 1,000 patient days pre and post-intervention. CDI rates were also monitored.

Results

There was an overall reduction in quinolone use by 68.2% and 75.4% at Site 1 and Site 2, respectively. Clindamycin use decreased in the first year at both sites by 50.5% and 30.8%, respectively. As the use of both antimicrobial classes decreased, the corresponding CDI rates at both sites also decreased.

Conclusion

The implementation of targeted PAF interventions successfully reduced usage of clindamycin and quinolones, with resultant decreases in CDI incidence.

Table H03-1:

Antimicrobial consumption and CDI rates on CTU

		Antimicrobial Use (DDD/1000 Patient Days)
			Quinolones
Site		Clindamycin	Ciprofloxacin	Levofloxacin	Total	CDI Rate/1000 Patient Days
Site 1	Pre-PAF (2014–15)	129.80	742.43	1369.70	2241.93	1.54
	Year 1 PAF (2015–16)	64.22	579.71	951.21	1595.14	0.52
	Year 2 PAF (2017–18)	73.31	320.19	350.20	743.70	0.42
Site 2	Pre-PAF (2014–15)	68.37	801.84	1126.21	1996.42	0.43
	Year 1 PAF (2015–16)	47.32	508.66	759.59	1315.57	1.05
	Year 2 PAF (2017–18)	46.18	195.51	262.34	504.03	0.40

Open in a new tab

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):23–24.

H04. Provision of Antibiotic Prescribing Feedback to High-Volume Primary Care Physicians: Design and Implementation of a Randomized Controlled Trial

Kl Schwartz ^1,^2,³, N Ivers ⁴, B Langford ¹, V Leung ¹, KA Brown ¹, N Daneman ⁵, M Silverman ⁶, S Elsayed ⁶, J Wu ¹, E Shing ¹, J Leis ⁵, G Garber ¹

Objectives

We partnered with provincial health system organizations to conduct a pragmatic randomized controlled trial, mailing letters to the highest antibiotic-prescribing primary care physicians across Ontario, Canada. Herein, we describe the methodology and present early findings on implementation and acceptability of the intervention.

Methods

We used the IQVIA Xponent^TM database to identify the top 3,500 (27%) of 13,102 Ontario primary care physicians prescribing the highest number of antibiotics by volume dispensed. These physicians were randomized 3:3:1 to letter 1, letter 2, and control (no letter). The letters incorporated persuasive communication and provided normative social comparisons and change ideas for appropriate antibiotic initiation for acute respiratory infections (letter 1) or recommended treatment durations for uncomplicated infections (letter 2). The letters were drafted with extensive stakeholder engagement. Waiver of consent was granted by the ethics research board and all physicians, including controls, will receive an additional letter with a debrief after one year. An email address was provided within the letter for those wishing to ask questions or provide feedback. These responses were organized into common themes.

Results

A total of 3,000 letters were sent in December 2018. Participants prescribed more than 442 (average = 1031) antibiotics in the prior year. We received 31 (1%) responses within one month. The most common themes were i) concerns regarding lack of case-mix adjustment in the measurement of performance (n = 30) and ii) interest in ongoing reflection about antibiotic prescribing (n = 10). Only three (0.1%) physicians opted out.

Conclusions

Provision of unsolicited antibiotic prescribing feedback to primary care physicians is feasible and inexpensive particularly for jurisdictions that have antibiotic use data. Adjusting for patient volume may enhance acceptance of feedback. Changes in antibiotic prescribing rate and proportion of prolonged treatment durations will be evaluated as trial outcomes at 6, 12, and 24 months. NCT03776383.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):24–25.

H05. The 2018 Global Point Prevalence Survey of Antimicrobial Usage (AMU) and Resistance in 42 Canadian Hospitals

GJ German ¹, J Grant ², M Lefebvre ³, D Mertz ⁴, J Boswell ¹, S Lutes ¹, J Roberts ⁵, K Afra ⁶, L Valiquette ⁷, Y Émond ⁸, M Carrier ⁹, A Lauzon-Laurin ¹⁰, TT Nguyen ¹¹, H Al-Bachari ¹², J Kosar ¹³, S Peermohamed ¹³, D Landry ¹⁴, T MacLaggan ¹⁵, P Daley ¹⁶, G McDonald ¹⁷, A Ang ¹⁸, S Chang ¹⁹, V Leung ²⁰, K Katz ²¹, I Pauwels ²², H Goossens ²², A Versporten ²², C Frenette ²³

Objectives

The Global Point Prevalence Survey (PPS) is a well-established initiative for monitoring AMU and resistance. We report the results of a survey conducted in 42 hospitals in 2018.

Methods

The electronic PPS questionnaire was completed by each site for all inpatients receiving antimicrobials on a selected day between January and December 2018. Data collected included ward type, demographics, antimicrobial prescribed, indication(s), local guideline compliance, and antimicrobial-resistant organisms. A web-based application was used for data entry and reporting (www.global-pps.com).

Results

Canadian survey participation increased by 200% from 2017. Of the 42 hospitals, 25 were teaching institutions, including 14 tertiary care centres. The survey screened 11,748 patients on 702 wards. One-third of patients (n = 3924) received at least one antimicrobial, and 31% were on at least one antibiotic (AB). Of the 4,779 antibiotic courses, 49% were for community-acquired infections, 30% for hospital-acquired infections, 11% for surgical prophylaxis (SP), and 7.3% for medical prophylaxis. The most commonly treated infection was respiratory (27%). The 5 most frequently prescribed antibiotics were piperacillin/tazobactam (15%), cefazolin (12%), ceftriaxone (10%), vancomycin (8.4%), and ciprofloxacin (7%). Carbapenem use accounted for 5.2%. Of the targeted antibiotic therapies (n = 1481) 5.6% were for MRSA, 2.8% for ESBL, 0.74% for VRE, and 0.14% for carbapenemase-producing Enterobacteriaceae. Antibiotic indication and stop date were documented in 60% and 64% of charts, respectively. Guidelines were available for 77% of the therapies with reported compliance of 81%. SP was greater than 24 hours in 29% of instances.

Figure H05-1: — Global Point Prevalence Surevey Canada 2018, antimicrobial usage

Conclusions

This type of PPS of AMU is increasingly used in Canadian hospitals and identifies targets for interventions to improve AMU and provide benchmarks for hospitals to compare. Areas for improvement include indication documentation, antimicrobial stop dates, and prolonged SP. Overall the prevalence of treatment for multi-drug resistant organisms in Canada continues to be relatively low.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):25.

H06. Direct Vitek 2™ Susceptibility Testing (DST) on Enterobacteriacea from Blood Culture Samples

D Gregson ^1,², W Chan ^1,², S Buchan ¹, D Doyle ¹, J Pitout ^1,²

Objective

To compare DST testing of Enterobacteriaceae in patients with positive blood cultures (BC) with standard testing using Vitek 2^TM gram-negative cards.

Methods

BCs were performed using the BacT/Alert™ system. BCs positive for gram-negative bacilli on staining were subject to a 1.5 mL sample being lysed and washed for direct MALDI identification. Residual sample was processed for susceptibility testing as per the standard Vitek II protocol. Vitek II susceptibility results of plate-based cultures were considered the gold standard. Results were interpreted using CLSI guidelines. Reportable antibiotics for Escherichia, Klebsiella, and Proteus spp (EKP spp) in our laboratory are gentamicin (GNT), ceftriaxone (CTRX), piperacillin-tazobactam (PTZ), meropenem, ertapenem, ciprofloxacin (CIP) and cotrimoxazole (SXT). For Enterobacter, Citrobacter, Serratia and Morganella spp (EbCSM spp) ceftriaxone is replaced with cefepime and PTZ is not reported. Correlation was classified as categorical agreement (CA), minor errors (mE), major errors (ME) and very major errors (VME) according to CLSI guidelines.

Results

For 33 EbCSM isolates, there was categorical agreement in all results. Antibiotic resistance in this group was limited to 4 isolates resistant to SXT and 2 resistant to GNT. In the EKP group (n = 492), resistance to GNT/CIP/CTRX/PTZ/SXT occurred in 9/23/16/2/33% of isolates respectively. There was no carbapenem resistance. VME occurred only in the SXT results (5/161). One ME occurred for all antibiotics reported for a single isolate. For PTZ, mEs occurred in 29 (6%) cases. With the exception of PTZ and SXT, categorical agreement was >99% (95% CI: 98.4–100). Results from DST were available approximately 18 hours before standard test results.

Conclusion

With the exception of SXT results, DST correlates well with standard testing and can be used to appropriately direct therapy in patients with bacteremia due to these organisms.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):25–26.

I01. Surveillance for Babesia in Canadian Blood Donors using the Ultra-Sensitive Procleix® Babesia Assay (PBA) and secondary laboratory Testing Methods: July–October 2018

S Stramer ¹, MC Proctor ¹, L Tonnetti ¹, V Brès ², JM Linnen ², F Bernier ³, G Delage ³, Y Grégoire ³, J Labrie ³, M Bigham ⁴, SJ Drews ⁴, G Hawes ⁴, S O’Brien ⁴, V Scalia ⁴, M Fearon ⁴

Background

Babesia microti is a tick-borne intra-erythrocytic parasite, causing transfusion transmitted infections (TTI), which now threatens the donor blood supply in the northeastern/upper-midwestern U.S. A previous limited Canadian seroprevalence study (2013; n = 13,993) did not identify B. microti antibodies in Canadian donors Manitoba-eastward. The babesiosis and Lyme disease vector, Ixodes species, is spreading in Canada with an increase in Lyme disease cases reported. Here, we report a follow-up widescale surveillance study of Canadian blood donors using the highly sensitive Procleix® Babesia Assay (PBA) (LOD = 1–4 parasites/mL), an investigational nucleic acid test for detection of B. microti, B. divergens, B. duncani, and B. venatorum on the Panther® system.

Methods

Both randomized and selected EDTA-plasma retention tubes (n = 50,586) collected from Canadian blood donors (tick season; July-October 2018) were shipped to the American Red Cross (ARC) for screening by minipool lysates (MPLs; pools of 16) using the PBA. Reactive pools were deconvoluted and resultant donor lysates (IDLs) retested in duplicate. Reactive donations were tested by secondary methods specific for B. microti: ARC IgG immunofluorescence assay [IFA] and IMUGEN PCR (LOD = 66 parasites/mL). Another subset of randomly selected PBA-non-reactive samples (Manitoba-eastward, n = 14,710) were screened using ARC IFA and if positive, IMUGEN IFA and PCR.

Results

Of 50,586 donors, one (0.002%) was PBA-reactive with additional ARC IFA-positive (≥1:1024) and PCR-negative test results. Of PBA-non-reactive donors, 3/14,710 (0.02%) were positive by one secondary testing method (ARC IFA [range 1:128–1:512]), while 1/14,710 (0.007%) was positive by more than one secondary testing method (ARC IFA-positive [1:512], IMUGEN IFA-positive (≥1:128)]. None of the implicated donors had relevant travel history.

Conclusions

B. microti is gaining a foothold in the Canadian blood donor population. Future laboratory surveillance of Canadian blood donors using highly sensitive methods for Babesia (including species other than B. microti; current prevalence unknown) is warranted.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):26.

I02. Prevalence of Pertactin-Deficient Bordetella pertussis Isolates in Ontario, Canada, from 2009 to 2017

S Bolotin ^1,², A Marchand-Austin ¹, R Tsang ³, M Shuel ³, NS Crowcroft ^1,², K Schwartz ^1,², SL Hughes ¹, L Friedman ¹, K Cronin ⁴, J Ma ², G Van Domselaar ^3,⁵, M Graham ^3,⁵, S Tyler ³, FB Jamieson ^1,²

Background

Despite high vaccine coverage, a resurgence of Bordetella pertussis has been observed in recent years, particularly in countries that administer the acellular pertussis vaccine. In many jurisdictions, this resurgence has coincided with the emergence in 2011 of B. pertussis strains deficient in pertactin (PRN-N), a virulence factor that is a component of the acellular pertussis vaccine used in Canada and elsewhere. The objective of our study was to measure trends in the prevalence of pertactin-deficient strains in Ontario, from 2009–2017.

Methods

We characterized all available isolates from 2009–2017 using Western blot analysis performed at the National Microbiology Laboratory and Public Health Ontario laboratory. We performed epidemiological analyses to assess whether there were significant associations in PRN-N status by year, age-group or whole-cell versus acellular pertussis vaccine program-eligibility.

Results

Of the 413 isolates available for characterization, 34.6% (143/413) were PRN-N. These first emerged in 2011, reaching a maximum prevalence of 70.8% (34/48) in 2016, decreasing thereafter to 46.2% (30/65) in 2017 (χ² test for trend p = 0.0003). From 2009–2017, the <6-month age group had the highest PRN-N prevalence at 36/84 (42.9%) (χ² p = 0.29). There was no statistically significant difference in the proportion of PRN-N isolates from individuals eligible for priming doses of whole-cell versus acellular pertussis vaccine (p = 0.95).

Conclusions

PRN-N strains emerged in Ontario in 2011, coinciding with emergence in other jurisdictions globally; so far, PRN-N in Ontario remains lower in prevalence than is observed in some other jurisdictions. Although no statistical association was observed for PRN-N prevalence by vaccine program-eligibility, this may have been impacted by selection bias of available and archived B. pertussis isolates. Future studies should include case vaccination history to elucidate the potential association between PRN-N strains and whole-cell versus acellular vaccine priming.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):26–27.

I03. Clostridioides difficile Strain Divergence Over Time

DJ Speicher ^1,², K Luinstra ², J Maciejewski ², KK Tsang ¹, AG McArthur ¹, M Smieja ^2,³

Background

Clostridioides difficile infection (CDI) is a serious hospital-associated infection with severe outbreaks caused by the hypervirulent NAP1/MLST-1 strain. Whole genome sequencing has shown that most outbreak strains are clonal, whereas non-outbreaks display a wide diversity of strains. To examine strain diversity in clinical settings, a subset of C. difficile isolates from symptomatic CDI from an acute care hospital were compared with isolates from C. difficile colonized (CDC) asymptomatic patients from the same hospital.

Methods

A subset of PCR-positive stool samples from clinically confirmed CDI isolates from 2016 (13/110), 2017 (8/111), and 2018 (13/65), and CDC from 2017 (17/185) were cultured 3 times consecutively on CHROMagar™. C. difficile, sub-cultured on Columbia colistin-nalidixic acid (CNA) media, had DNA isolated, shotgun sequenced, and genome assembled for both MLST typing and genome-wide SNP phylogenetic analysis.

Results

Based on MLST profiles, the C. difficile types detected were diverse. Of the presumed binary toxin–positive/NAP1 strains (i.e., PCR tcdA/tcdB–positive) 7/12 (58%) were NAP1/MLST-1 and 3/12 (25%) were NAP7/MLST-11. NAP1/MLST-1 was not detected in any CDC isolate. NAP4/MLST-2,14 were detected in 2016 (n = 4), 2017 (n = 2), 2018 (n = 1), and in CDC isolates (n = 3). MLST-42 was dominant in CDC isolates (5/17; 29%) and decreased in prevalence in CDI isolates over time (2016 = 4; 2017 = 0; 2018 = 1).

Conclusion

C. difficile strains amongst both CDI and CDC individuals are highly divergent. While molecular assays are misclassifying 25% of “NAP1” strains, both NAP1 and NAP7 are hypervirulent. The number of MLST-42 CDC isolates is concerning as it has been reported to be the most common strain causing CDI among U.S. adults. This highlights the need for continued genomic surveillance of both CDI and CDC individuals. Genome-wide SNP phylogenetic analysis is currently being performed.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):27.

I04. A Multi-Jurisdictional Outbreak of Hepatitis A in Ontario, Canada, in 2017–2018: The Role of Genetic Sequencing in Outbreak Detection and Response

LM Macdonald ^1,², K Johnson ¹, C Lee ¹, B Warshawsky ^1,³, LW Goneau ^1,⁴, A Andonov ^5,⁶, H Sachdeva ^2,⁷

Objectives

In recent hepatitis A outbreaks in Europe and the United States, genetic sequencing has contributed to timely outbreak detection and response. We describe the epidemiology of an ongoing multi-jurisdictional outbreak of hepatitis A in Ontario, Canada, and discuss the role of hepatitis A virus (HAV) genetic sequencing.

Methods

Outbreak-confirmed cases were defined as hepatitis A cases with presence of anti-HAV IgM antibody occurring on or after June 1, 2017, in Ontario residents or visitors, with the same HAV genotype (1A) and genetic sequence (VRD_2016_521). Local health units reported epidemiological information in Ontario’s integrated Public Health Information System (iPHIS). Public Health Ontario (PHO) Laboratory referred positive specimens to the National Microbiology Laboratory for sequencing. PHO analyzed iPHIS data.

Results

From June 1, 2017 to December 31, 2018, 124 outbreak-confirmed cases of hepatitis A with the HAV genotype 1A, VRD_2016_521 sequence were reported in ten Ontario health units. This strain was circulating in recent HAV outbreaks in Europe and the U.K. among men who have sex with men (MSM). In Ontario, the median age of cases was 33 years, 59% were male, and 48% were hospitalized. Reported risk factors included illicit drug use (59%), being under-housed (20%), and MSM (12%); most did not travel. In the first affected health unit, sequencing and epidemiological data supported person-to-person transmission among MSM predominantly. However, different health units and groups were affected over time; responses (e.g., targeted promotion of hepatitis A vaccination) focused on populations most at risk locally. Overall, 35% of outbreak cases did not have any risk factor information (i.e., were associated with this outbreak based on genetic sequencing).

Conclusions

HAV genetic sequencing and public health surveillance enabled timely outbreak detection and response. Sequencing has helped understand the burden of this ongoing outbreak and inform local responses.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):27.

I05. Withdrawn

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):27–28.

I06. An Update of Cryptococcus Molecular Typing Results in British Columbia (2011–2018)

M Lee ¹, V Tang ¹, B Hon ¹, K Fernando ¹, N Chahil ¹, A Paccagnella ¹, L Hoang ^1,², M Morshed ^1,²

Objective

Cryptococcus gattii was previously associated with tropical and subtropical areas until it emerged in 1999 on Vancouver Island, British Columbia and later spread to the BC Mainland and into the Pacific northwestern U.S. In this study we are reporting on the trends of Cryptococcus in BC based on their Restriction Fragment Length Polymorphism (RFLP) patterns to compare the current circulating isolates with previous outbreak strains results.

Methods

C. gattii is reportable in BC, and, as with other Cryptococcus species, are referred to the laboratory for subtyping. Cryptococcus isolates (n = 251) were recovered using CGB agar from clinical and veterinarian samples between 2011 and 2018. The DNA of isolates was amplified through PCR targeting the URA5 gene. The typing was elucidated from the two sets of restriction enzymes on the amplicons.

Results

Approximately 19 C. gattii isolates on average are received annually, with 2012 being the peak year (n = 31), while only 8 isolates were received in 2018. Approximately 0.80% (2/251) of the CGB agar results is consistent with the RFLP typing result. This also has been observed in our past surveillance. Seven RFLP types have been identified in the study, 60.16% of them belong to G group (C. gattii) while 39.84% are N group (C. neoformans). The variant G (VG) group is subdivided into VGI-VI, of which VGIIa continues to be predominant in BC.

Conclusions

Until 2018, the numbers of Cryptococcus isolates were steady except for the peak year in 2012. However, this dropped dramatically in 2018. The VGIIa, the previous dominant outbreak type, continues to be the major RFLP type in BC (84.11% in the C. gattii group). All the animal isolates (from bird, cats, horses, dolphin, porpoises, and seals) are all C. gattii which belonged to VGI, VGIIa, and VGIIb type, strains endemic in BC.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):28.

J01. HCV Screening via Rapid Point of Care Testing in Patients on Opiate Substitution Therapy in Peel Region, Canada

A Misra ¹, D Wiens ², H Kassam ³, R Gill ⁴, SM Borgia ^5,⁶

Background

Chronic Hepatitis C virus (HCV) infection can lead to cirrhosis, liver failure, and liver cancer. Recent data has demonstrated that HCV-infected patients who are actively injecting drugs or on opiate substitution therapy (OST) have excellent outcomes with new direct-acting antiviral therapy. Barriers to HCV care for OST patients include a) invasive sampling, b) delayed result turnaround, c) lack of physician awareness, d) loss to follow-up. We sought to investigate trends in the standard-of-care blood testing (SOC) versus point-of-care (POC) with respect to diagnosis and linkage-to-care in patients on OST at 3 clinics in Peel Region, Ontario. We hypothesized that patients with a positive HCV-Ab obtained via the POC test will have increased linkage-to-care.

Methods

We used a commercially available POC rapid HCV-Ab test (OraQuick®) that has high specificity and sensitivity in detecting HCV-Ab. Patients were randomly assigned 1:1 to receive SOC or POC test.

Results

40 (100%) oral POC swabs tested negative for HCV-Ab. SOC blood results were received for 7/40 (18%) patients, all HCV-Ab negative. Of 80 patients enrolled, 32% were female; 67% were Canadian-born; and 19% were born in India. 62% of patients were unaware of their HCV status, whereas 38% were aware of a previous negative HCV-Ab result. 12/23 (52%) patients on OST enrolled from one clinic had no history of injection drug use.

Conclusions

The proportion of patients for whom HCV-Ab results were available was greater in POC than SOC. All patients for whom results have been received have tested negative and therefore no linkage opportunity was triggered. It was surprising that in a marginalized population with significant risk factors for HCV infection, no cases were identified; perhaps due to the small sample size, or lower incidence of injection drug use. The study demonstrated that POC testing uptake worked well as a rapid screening method.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):28–29.

J02. Rapid Starts to Stop Hepatitis C: Same Day Hepatitis C treatment Starts Enhancing Patient Engagement and Follow-Up in a Vulnerable, Treatment-Naive Population

S Greenan ¹, G Carruthers ^1,², L Barrett ^2,³

Background

Hepatitis C virus (HCV) elimination requires alternate care models for key populations. Beyond diagnosis, engaging people in HCV treatment that leads to treatment completion and cure is a large barrier to HCV elimination. Research in the HIV field has demonstrated better health care engagement with rapid, same day treatment starts. Our objective is to determine if rapid access to HCV treatment improves engagement in HCV and non-HCV health care.

Methods

Patients are identified and referred to PEI’s HCV elimination program through public health, community providers, or “bring-a-friend” strategies. Program staff facilitate baseline blood work, do preliminary drug-drug interaction checks, and book a first appointment within 1–2 weeks of blood work. Treatment-naïve patients without contraindications are offered glecaprevir/pibrentasvir to start treatment at the first visit. Medication adherence, side effects, SVR12, and attendance at opioid substitution therapy clinics are self-reported.

Results

Patients assessed between February and December 2018 were included; 143 patients were referred, and 103 (72%) were seen for initial visits. Of those who attended the first visit, 6 had contraindications to treatment (5 medication interactions and 1 pregnancy). Of the treatment-naive individuals eligible for treatment, 97 (100%) started treatment. 85 (87.6%) patients started treatment on their first visit, with 5 discontinuations (3 for non-HCV–related reasons, 2 due to undetected HCV viral load). One patient was lost to follow-up before SVR. 74 (76.2%) have completed treatment with 27 (27.8%) achieving SVR. Importantly, individuals with difficulty attending opioid substitution clinic appointments before HCV treatment had improved attendance at appointments out to the SVR timepoint. Attendance for other medical appointments was variably improved. No significant safety issues were noted.

Conclusions

Rapid treatment start is safe, and has a very high rate of successful HCV and non-HCV care engagement. Same day, first visit HCV treatment start should be explored as an HCV elimination tool.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):29–30.

J03. Burden and Clinical Impact of Non-Influenza Respiratory Viral (NIRV) Infections among Hospitalized Adults

N Lee ¹, S Mubareka ², L Zapernick ¹, C Bekking ³, M Labib ¹, D Waldner ¹, N Zelyas ⁴, S Smith ¹

Objectives

The burden and clinical significance of non-influenza respiratory viral (NIRV) infections among hospitalized patients are poorly understood. This study aimed to analyze their characteristics, severity, outcomes, and estimate impacts relative to influenza.

Methods

Multiplex PCR-based respiratory viruses surveillance data from two university-affiliated hospitals during the 2014/15–2017/18 seasons were analyzed. All adult (>17 years) hospitalized patients with acute respiratory illnesses tested positive for 1 (or more) of the 18 virus targets using Luminex RVP/RPP assays (FLU = influenza viruses; NIRV = RSV, parainfluenza viruses, rhinoviruses/enteroviruses, coronaviruses, human metapneumovirus and adenovirus) were included. Prospective data in infection control surveillance programs was extracted and electronic records were reviewed. Interim analysis results are reported.

Results

Among the 1364 infections analyzed, half (51.6%, 42.7%–56.5%) were caused by NIRVs. RSV (13.2%) and rhinoviruses/enteroviruses (13.0%) were the commonest. NIRV (n = 704) and FLU (n = 660) patients differed in their characteristics: age (62.4±20.1 versus 67.8±18.2 years, P<0.001), immunocompromised (36.4% versus 27.0%, P = 0.002), hospital-acquired (17.9% versus 13.5%, P = 0.032), diagnosed by LRT samples (9.0% versus 6.1%, P = 0.051). Clinical severity/outcomes were not significantly different: ICU admission (22.3% versus 19.9%), ICU length-of-stay (14.6±18.3 versus 14.8±21.7 days); assisted ventilation (24.4% versus 18.8%); lower respiratory and cardiovascular complications (60.6% versus 60.9%); probable bacterial superinfections (5.3% versus 3.2%); weighted ordinal outcome score (2.0±3.5 versus 1.9±3.2); 30-day mortality (9.0% versus 7.4%). Immunocompromised state independently predicted higher mortality (Cox-regression aHR 1.9; 95% CI: 1.2–2.9; P = 0.005).

Figure J03-1: — Kaplan-Meier curves showing probability of survival of hospitalized adults diagnosed with non-influenza respiratory virus infections (NIRVs) and influenza (FLU) using multiplex PCR assays

*Note.* Adjusted survival function (for age, gender, immunocompromised state showed insignificant differnece between (a) NIRV versus FLU (aHR 1.3; 95% CI: 0.8–2.0; P = 0.24), and (b) individual viruses (P=0.52)

Conclusions

NIRVs may cause severe illness and high mortality similar to influenza among the hospitalized adults. Burden of disease is substantial. The unmet need for antiviral therapy and vaccination against NIRVs in adults should be promptly addressed.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):30.

J04. Withdrawn

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):30.

J05. Decline in Yield of Acute HIV Infections Detected Using Pooled Nucleic Acid Testing Following Implementation of 4th Generation Screening

M Krajden ^1,², D Cook ¹, A Mak ², Y Chang ², A Yu ¹, P Levett ², A Jassem ²

Background

HIV RNA nucleic acid tests (NAT) have a shorter window period (i.e., the time from infection to laboratory-based HIV detection) compared with HIV screen tests which detect antibodies alone (1^st to 3^rd generation tests). With the introduction of 4^th generation screening tests, which detect both HIV antibodies and p24 antigen, many acute HIV infections (AHI) are detected earlier than by 3^rd generation screening, thus reducing the impact of pooled NAT testing.

Methods

We examined the AHI yield of pooled NAT for three time periods: 1) May 2013 to May 2015: 3^rd generation screening; 2) May 2015 to May 2017: 4^th generation; and 3) May 2017 to Aug 2018: 4^th generation. Screen test–negative samples from individuals at high risk of HIV infection were combined into pools of 24 and tested for HIV RNA. AHI was defined as screen test negative, and HIV RNA positive; early HIV was defined as screen test positive, immunoblot negative or indeterminate, and HIV RNA positive. We compared the AHI and early HIV case yields by time period.

Results

For the three periods, 598,269, 765,477, and 502,863 specimens respectively were screened, and 834, 969 and 869 pools were RNA NAT tested. Case yields were: 1) 30 AHI; 54 early HIV; 2) 7 AHI; 77 early HIV; and 3) 2 AHI; 29 early HIV. AHI case yields based on pools tested declined from 3.6% to 0.72%, and 0.23%, respectively (period 1 versus 2: p<0.0002).

Conclusions

AHI cases diagnosed solely by pooled NAT declined significantly after introduction of 4^th generation screening. The decline has continued into the most recent period, even as the number of tested pools increased over time. Pooled NAT adds considerably to program screening costs, and further studies are required to determine whether this approach to AHI detection remains cost-effective.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):30–31.

J06. Sequence-Based Typing of Enteroviruses from Clinical Specimens, 2017–2018

M Cheung ¹, R Chow ¹, K Gunadasa ¹, AN Jassem ^1,², M Krajden ^1,², PN Levett ¹

Objectives

Enteroviruses cause a wide range of human infections, ranging from mild respiratory tract infections, to severe central nervous system infections. Many enterovirus strains are associated with specific clinical syndromes. Surveillance of circulating enteroviruses within a population has important implications for public health. Historically, isolates of enteroviruses were characterized into 71 serotypes. More recently, sequence-based typing has led to the identification of many more genotypes than were previously recognized.

Methods

Over a 19-month period (June 2017 to December 2018), specimens positive by PCR for enterovirus/rhinovirus were sequenced. Subtyping of the specimens was performed by partial Sanger sequencing of either the 5’ untranslated region (5’-UTR) or the viral capsid protein (VP1). VP1 is better correlated with the serotype than is 5’-UTR, due to the capsid protein corresponding to the neutralization domains.

Results

110 specimens were sequenced (2017 n = 32, 2018 n = 78), 32 by 5’-UTR and 78 by VP1. Specimen types included CSF (25), superficial lesions (63), upper respiratory tract (15), stool (5) and blood (2). Sequences that yielded enough data for typing were obtained from 80 specimens; 30 specimens could not be subtyped. Genotypes identified included Coxsackie A2 (1), Coxsackie A6 (54), Coxsackie A9 (1), Coxsackie A16 (4), Coxsackie B5 (1), Coxsackie B9 (5), Echovirus 9 (1), Echovirus 30 (3), human Enterovirus 71 (4), Rhinovirus A (3) and Rhinovirus C (2).

Conclusions

Coxsackie A6 viruses were detected most frequently from skin lesions, with 90% of typeable skin lesions positive for Coxsackie A6 virus during this period. CSF samples yielded a variety of enterovirus genotypes. Sequence-based typing of enteroviruses was possible from 70% specimens, and yielded useful information about circulating enteroviruses.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):31.

K01. Optimization of a Next Generation Sequencing Assay for HBV Drug Resistance Testing

M Payne ^1,², G Ritchie ¹, L Merrick ¹, T Lawson ¹, C Lowe ^1,²

Background

Next-generation sequencing (NGS) for Hepatitis B virus (HBV) resistance testing is a highly sensitive method, able to detect low-level mutant subpopulations. Our clinical virology laboratory transitioned from the GS Junior System (Roche 454) to the MiSeq (Illumina), but identified that MiSeq sequencing of low diversity amplicon libraries containing multiple samples was challenging due to sample mis-assignment and low-quality reads. Our study investigates the validation of the MiSeq for HBV resistance testing and troubleshooting of sequencing errors to enable clinical reporting of low-level mutations.

Methods

We performed amplicon sequencing of the hepatitis B RT gene on the MiSeq Reagent Nano Kit v2. HBV ATCC® 45020D™ was utilized to determine error rates for base calling. Several modifications were made to improve sample read assignments and base calling accuracy: unique dual indexes for each patient sample, PhiX concentration was increased to 33%, cluster density was reduced from 800 to 400 K/mm², Q-score trimming (Q30), and primers with staggered 1–4 base pair barcodes. A total of 56 patient samples were tested on both the GS Junior and MiSeq.

Results

Initial MiSeq results using a dual index PCR method with the recommended PhiX concentration of 12%, resulted in unacceptable levels of error rates for codon calling of up to 7%. After the above modifications were made, the error rates were less than 0.2%. There was a high agreement rate for patient samples between the GS Junior and MiSeq, with regards to total drug resistance mutations and patient sample agreement, 74/79 (94%) and 51/56 (91%), respectively. HBV genotype results were concordant for 56/56 samples.

Conclusions

HBV resistance testing with the MiSeq required significant modifications to decrease sample mis-assignments and base calling errors. With these modifications, accuracy was improved, and mutations could be reported with confidence for subpopulation levels as low as 1%.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):31–32.

K02. Evaluation of a Target Hybridization-Based Next-Generation Sequencing Assay for Diagnosis of Acute Respiratory Tract Infections

SH Zhan ¹, SM Alamouti ¹, L Yip ², BS Kwok ¹, MA Qadir ¹, J Gelinas ¹, S Mubareka ²

Objective

The ONETest^TM (Fusion Genomics Corp., Canada) is a next-generation sequencing (NGS) assay utilizing target hybridization that identifies 48 viral and bacterial respiratory pathogens. In this retrospective pilot study, we evaluated the performance of the ONETest^TM to detect respiratory pathogens in patients diagnosed with acute respiratory tract infections.

Methods

We selected 36 archived nasopharyngeal swabs collected from patients who were admitted to the Sunnybrook Health Sciences Centre (Toronto, ON) in 2018. Mid-turbinate swabs from patients with respiratory symptoms were tested using the NxTAG Respiratory Pathogen Panel (RPP, Luminex Corp.) as per protocol. RNA extracted from a representative subset of the swabs (19 positive and 17 negative as per the RPP) was tested using the ONETest^TM. NGS library preparation and probe-based target hybridization were conducted following the ONETest^TM proprietary protocol; the target-enriched libraries were sequenced on an Illumina HiSeq; the NGS data were analyzed using the FusionCloud pipeline. The output includes pathogen identity, type/subtype, estimated pathogen load (normalized read count), target gene sequences, and mutations.

Results

The ONETest^TM showed a high agreement rate with the RPP; it detected either the pathogens found by the RPP (and/or pathogens missed by the RPP) or no pathogen (that is, negative by both assays) in 32 cases (~89%). In 8 cases (~22%), the ONETest^TM identified pathogens not detected by the RPP, and these consisted mainly of human coronaviruses. In one of the RPP-negative cases, the ONETest^TM detected two co-infecting pathogens, a human parainfluenza virus and a human metapneumovirus.

Conclusions

The ONETest^TM is an accurate technology for diagnosis of acute respiratory tract infections, with a favorable agreement rate with the RPP and a capacity to detect co-infections missed by the RPP. Furthermore, the ONETest^TM yields additional genomic and quantitative data not available using conventional methods.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):32.

K03. Laboratory Diagnosis of Vertical HIV Transmission: Test Ordering Practices for Infants in Alberta

NC Marshall ¹, CL Charlton ^1,²

Objectives

Human immunodeficiency virus (HIV) infection in infants is associated with severe disease and high mortality; therefore, accurate and timely diagnosis is crucial. The World Health Organization recommends virologic testing 0–2 days and 4–6 weeks after birth to diagnose in utero and perinatal transmission, respectively. No particular test type is endorsed in any pediatric guidelines, therefore clinical practice can differ in the use of nucleic acid tests (NATs) detecting viral genomic RNA or integrated proviral DNA (pvDNA). We evaluated virologic test selection and appropriateness of testing for infants across Alberta.

Methods

All HIV NATs ordered in Alberta between 2015–2018 were reviewed (n = 16,706). Tests for patients <1-week- and 1-month-old were considered appropriate for diagnosis of in utero and perinatal transmission, respectively.

Results

Of all HIV NATs performed in Alberta, 5% were for infants <18 months; of those, 68% were ordered at urban health centres. Of the infants tested, 58% received a NAT at either recommended diagnosis window (i.e., <1-week- or at 1-month-old). Overall, 60% of NATs were RNA and 40% were pvDNA; however, the two urban centres preferred different test types. Finally, we examined turnaround time from specimen collection to results: RNA and pvDNA tests took an average of 5 and 17 days, respectively, highlighting a practical difference between tests that reflects outsourcing the pvDNA test.

Conclusions

This work reveals differences in HIV NAT preferences for infants, ordering practices between regions, and turnaround times. Further research is needed to determine which test is most sensitive for infants, specifically, before optimizing testing to deliver the best possible care for HIV-exposed infants. Because delayed results impact timely treatment in infants with perinatally-acquired HIV, new approaches to reduce pvDNA turnaround time (e.g., offering tests in-house for patients <1-week-old) could improve management of this vulnerable population.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):32–33.

K04. Adherence to palivizumab dosing schedule—Manitoba Experience

K Hogue ¹, A Chiu ¹, J Embree ¹, S Monkman ¹, B Lavallee ¹

Background

Respiratory syncytial virus (RSV) infection is a major cause for infant hospitalization. Passive immunization with palivizumab (Synagis®) is given as 5 monthly doses over the RSV season to high-risk infants to reduce RSV-related hospitalizations. Adherence to dosing schedule optimizes its effectiveness.

Objectives

Assess whether location of palivizumab administration is associated with difference in adherence to administration schedule, specifically, whether administration outside of urban centers might impact adherence.

Methods

Patient age, sex, prematurity status, eligibility risk factor, birth weight and location/dose dates were collected from a pre-existing database of Manitoba RSV program (MBRSVP) participants from 2007–2018. Primary analysis included all children who received their first 3 doses at the same outpatient clinic comparing adherence (defined as receiving subsequent doses on schedule) based on location (urban versus non-urban). Our secondary analyses looked at children who received all 5 doses in the same location, those who received an initial dose in hospital followed by 3 doses at the same outpatient clinic, and finally, a subgroup analysis for adherence in urban versus rural versus northern patients from the primary analysis.

Results

Adherence was higher in children receiving palivizumab in urban versus rural locations (89% versus 53%, respectively). Using a chi-squared test we found that location was associated with adherence (X² = 158.68, p<2.2e-16). Using logistic regression analysis, controlling for location and other potential confounders, demonstrated a statistically significant relationship between location and adherence. Secondary analysis showed similar results. Subgroup analysis revealed differences in adherence of 89% versus 73% versus 40% for the urban, rural and northern groups respectively with X²= 224.66 and p<2.2e-16.

Conclusions

Manitoban patients who receive palivizumab doses in urban centers are more likely to be adherent to their dosing schedule. These data suggest that additional strategies and resources are required to improve adherence in rural and northern communities.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):33.

K05. Study of Carriage of Haemophilus influenzae Type A among Children

M Ulanova ¹, EB Nix ¹, RS Tsang ², B Tan ³, N Le Saux ⁴

Background

Over the last two decades, Haemophilus influenzae type a (Hia) has emerged as an important cause of invasive disease mainly affecting young Indigenous children. Hia carriage in the upper respiratory tract is both pre-requisite for invasive disease and reservoir for transmission. To identify target populations for immunization with a new Hia vaccine under development, we initiated a multicentre study of nasopharyngeal carriage among Canadian children.

Methods

With prior parental consent, we collected used nasopharyngeal anaesthetic tubes from healthy children <5 years of age who underwent routine dental surgery under general anaesthesia in a regional hospital of Northern Ontario (NO) and another dental clinic in Saskatoon (SA). In NO, all children were First Nations; in SA, children came from various ethnic groups. Detection of H. influenzae and serotype characterization were performed using PCR amplification of capsular polysaccharide synthesis genes. Multilocus sequence typing was done via amplification and sequencing of 7 housekeeping enzyme genes; assignment of sequence types was done through the H. influenzae MLST website.

Results

By January 2019, 295 nasopharyngeal specimens were collected and analyzed, 198 in NO, and 97 in SA. Hia was identified in 17 (8.5%) and 5 (5%), respectively. In SA, 4 out of 5 children with Hia carriage were First Nations.

Conclusions

The carriage rates of Hia in healthy children <5 years of age in NO and SA are comparable to H. influenzae type b (Hib) carriage among Alaska Native children in the pre-Hib vaccine era. To prevent invasive Hia disease it will be essential to decrease pathogen transmission. Pediatric conjugate Hia vaccines have the potential to decrease carriage of Hia and thus decrease transmission and disease among susceptible populations.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):33–34.

K06. Clinical Response and Outcomes in Patients with Recurrent Clostridium difficile Treated with Frozen-and-Thawed Fecal Microbiota Transplant

I Darwish ¹, H Peters ¹, R Mah ¹, C Lee ², T Steiner ¹

Objectives

Frozen-and-thawed fecal microbiota transplant (FMT) is as effective as fresh FMT for the treatment of recurrent or refractory Clostridium difficile infection (CDI). Frozen FMT provides several advantages over fresh FMT including a reduction in the frequency of donor recruitment and screening, and more convenient access. The objective of this study is to assess clinical response and long term outcomes in patients enrolled in an open-label trial of frozen-and-thawed FMT for recurrent CDI.

Methods

Eligible participants with recurrent CDI enrolled at our centre received frozen-and-thawed FMT from eligible donors as per our open-label trial protocol. A chart review and telephone survey was conducted for participants enrolled between June 2015 and December 2017. Patients with a minimum time of six months of reliable follow-up were included.

Results

Median time from first FMT to follow-up was 17 months. Ninety-three percent (53/57) of patients achieved cure with repeat FMT. Forty-nine percent (26/53) of patients who achieved cure received one FMT, 28% (15/53) of patients who achieved cure received two treatments, 13% (7/53) of patients who achieved cure received three treatments, and 9% (5/53) of patients who achieved cure received 4–6 treatments. Ninety-one percent (48/53) of patients who achieved cure had a sustained response with no recurrent CDI at follow-up. Of those with recurrent CDI, 60% (3/5) occurred following use of antibiotics for other conditions.

Conclusions

Frozen-and-thawed FMT is effective for the treatment of recurrent CDI and often requires repeat treatments to achieve clinical cure. With a median follow-up time of 17 months, 91% of patients with CDI had a durable cure. Patients with recurrent CDI were often provoked via antibiotic use.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):34–35.

L01. Verification of Three Multiplex Carbapenemase Nucleic Acid Amplification Tests (NAAT) Using Species-Diverse Carbapenem-Resistant Gram-Negative Bacilli

B Hazlett ¹, Y Sokolskyy ^1,², P Lo ¹, T Mazzulli ^1,³, A McGeer ^1,³, SM Poutanen ^1,³

Objectives

Rapid detection of carbapenemase-producing organisms (CPO) is important. Data suggest direct-from-specimen NAAT is more sensitive than culture. We evaluated three multiplex carbapenemase NAAT targeting KPC/NDM/OXA48-like/VIM/IMP genes [BDMax Checkpoints CPO (BD), Allplex Entero-DR (Seegene), and Easyplex® SuperBug CRE Assay Version C (Amplex Diagnostics)] using 150 well-characterized (phenotypic and PCR/sequencing) gram-negative bacilli (GNB).

Methods

150 GNB including 122 CPO (116 targeted-CPO: 88 KPC/8 NDM/5 OXA/5 NDM+OXA/7 VIM/3 IMP; 6 non-targeted CPO: 1 GES/4 SME/1 NMC), and 28 non-CPO were tested comprising 145 Enterobacteriacaeae/5 non-Enterobactericeae GNB. Isolates were recovered from –80^oC under ertapenem-selective pressure. Limit of detection (LOD) was calculated in triplicate using four QC strains following manufacturer direct-from-specimen-protocols using 10E4/10E5/10E6/10E7/10E8 cfu/L concentrations in ESwab transport medium with results from Xpert® Carba-R (Cepheid) as reference. Colony counts confirmed concentrations and average LOD was calculated. Accuracy was determined using ≥10-fold-higher concentrations than calculated LODs. Discrepancies were repeated.

Results

LOD (cfu/L) are shown (Figure L01-1). The most sensitive to least sensitive assay was Seegene, BD, Cepheid, then Amplex. All 34 non-targeted-CPO/non-CPO were negative by all assays. 1 KPC/1 NDM/3 KPC were initially missed by BD/Seegene/Amplex, respectively but were positive on repeat testing of fresh subcultures suggesting initial lost plasmids. 2 IMP were reproducibly missed by BD and Amplex. Final CPO-detection sensitivities/specificities were 100% for all non-IMP targets; respective 95% CI were: KPC 95.0–100/87.6–100; NDM 73.4–100/95.9–100; OXA48-like 68.0–100/96.0–100; VIM 59.6–100/96.1–100. Sensitivities/specificities for IMP were 100%(38.2–100)/100%(96.2–100)(Seegene) and 33%(5.6–79.8)/100%(96.2–100)(BD&Amplex).

Conclusions

All three assays were highly-accurate (100% sensitive/specific) for detection for KPC, NDM, OXA48-like and VIM CPO. IMP was more challenging for BD and Amplex. LOD was variable but not substantially different between assays. These results along with workflow, turn-around-time, footprint, interfaceability, cost, and laboratory needs can be used to determine suitability for different laboratories.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):35.

L02. Evaluation of Commercial Screening Agars for the Detection of Carbapenemase-Producing Enterobacteriaceae

Cl Vermeiren ^1,², EJ Sheldrake ¹, K Katz ^1,^2,^3,⁴

Objective

Carbapenemase-producing Enterobacteriaceae (CPE) are emerging around the world, including Canada, and are associated with case fatality rates as high as 50%. Gastrointestional carriage of CPE may serve as the reservoir for cross-contamination in the health care setting, thus active surveillance is important for effective containment and outbreak prevention. In this study we evaluate commercially available screening agars for the detection of CPE using a panel of CPE, non-CPE-, carbapenem-resistant, and carbapenem-susceptible strains.

Methods

A panel of strains was assembled including clinical strains from our laboratory and highly characterized strains selected from the CDC-FDA Antibiotic Resistance Isolate Bank as indicated in Table L02-1.

Table L02-1:

Carbapenemase types in the test panel

No. of Strains	Carbapenemase Enzyme Type
33	KPC
34	NDM
11	OXA-48/OXA-48 like
10	VIM
5	IMP
79	None

Open in a new tab

Two commercial agars for detection of CPE were compared including CHROMID^TM Carba Smart (CCS) and Colorex™ SuperCARBA (CSC). Two different inocula were used: 10⁵CFU (high) and 10²CFU (low). All media were incubated in accordance with the manufacturer’s recommendations. Strain viability was confirmed by concurrently planting dilutions to a non-selective blood agar plate and carbapenemase production was confirmed present or absent from all strains by PCR.

Results

Sensitivities for the high inocula were 89% and 93% while the low inocula was 40% and 75% for CCS and CSC, respectively. The specificities were 75% and 59% for CCS and CSC, respectively. Both plates failed to recover four strains including Klebsiella pneumoniae IMP-4, K. oxytoca KPC-3, P. mirabilis KPC-6, and Proteus mirabilis KPC-2. Interestingly, CCS failed to recover two OXA-48 like strains, including K. pneumoniae OXA-232 and K. pneumoniae OXA-181, even at the high inocula.

Conclusions

CSC displayed better sensitivity, at high inocula and particularly at low inocula. Although CCS had higher specificity, overall, CSC had superior performance.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):35–36.

L03. Activity of Ceftobiprole Against Canadian Bacterial Pathogens from the CANWARD Study

PR Lagacé-Wiens ^1,², HJ Adam ^1,², JA Karlowsky ^1,², MR Baxter ², AJ Walkty ^1,², KA Nichol ¹, GG Zhanel ²

Objectives

Ceftobiprole is a recently re-released 5^th generation cephalosporin, currently available on the Canadian and European markets. It demonstrates in vitro activity against Staphylococcus aureus (methicillin-susceptible [MSSA] and methicillin-resistant [MRSA] isolates), Streptococcus pneumoniae, ESBL-negative Enterobacteriaceae, and Pseudomonas aeruginosa. It also has improved activity against AmpC-positive Enterobacteriaceae relative to ceftriaxone and ceftazidime. Related to its broad spectrum of activity, ceftobiprole may offer a monotherapeutic option for the treatment of complicated skin and soft tissue infections and community-acquired and nosocomial pneumonia. The purpose of this study was to evaluate the in vitro activity of ceftobiprole against a contemporary collection of isolates from the CANWARD study.

Methods

Isolates were collected from the ongoing CANWARD study between 2008–2010 and 2015–2017. Antimicrobial susceptibility testing was performed using broth microdilution panels following CLSI recommendations (M07, 11^th edition). Minimum inhibitory concentrations were interpreted using Health Canada breakpoints (ceftobiprole) or CLSI breakpoints (ceftazidime comparator). Where no breakpoints were available for ceftobiprole, the pharmacokinetic/pharmacodynamic breakpoint of 4mg/L was used.

Results

Results are summarized in Table L03-1.

Table L03-1:

Results

Organism	Ceftobiprole	Ceftobiprole	Ceftazidime	Ceftazidime
Organism	MIC50/90	% susceptible	MIC50/90	% susceptible
Escherichia coli (n = 5059)	≤0.06/0.12	93.1%	≤0.25/1	93.8%
ESBL-positive (n = 341)	32/>32	4.1%	16/>32	32.8%
Klebsiella pneumoniae (n = 1603)	≤0.06/0.12	94.6%	≤0.25/1	95.1%
ESBL-positive (n = 80)	>32/>32	11.2%	>32/>32	22.5%
Enterobacter cloacae (n = 670)	≤0.06.2	88.8%	≤0.5/>32	80.1%
Klebsiella oxytoca/Raoultella spp (n = 432)	0.25/4	88.7%	≤0.5/≤0.5	98.8%
Serratia marcescens (n = 390)	0.06/0.25	98.5%	≤0.5/1	99.2%
Proteus mirabilis (n = 384)	≤0.06/≤0.06	99.2%	≤0.25/≤0.05	98.4%
Klebsiella aerogenes (n = 155)	≤0.06/≤0.06	97.4%	≤0.5/>32	74.8%
Citrobacter freundii (n = 101)	≤0.06/1	93.1%	≤0.5/32	82.2%
Morganella morganii (n = 98)	≤0.06/≤0.06	99.0%	≤0.5/8	88.8%
Pseudomonas aeruginosa (n = 2281)	4/16	64.7%	4/32	81.2%
Stenotrophomonas maltophilia (n = 421)	>32/>32	0.2%	>32/>32	24.0%
Acinetobacter baumannii (n = 110)	0.5/2	90.9%	4/32	78.2%
Staphylococcus aureus, MSSA (n = 4196)	0.5/0.5	100%	16/32	N/A
Staphylococcus aureus, MRSA (n = 1098)	1/2	100%	>32/>32	N/A
Staphylococcus epidermidis (n = 608)	0.5/1	99.7%	16/>32	N/A
Streptococcus pneumoniae (n = 1315)	≤0.03/0.06	99.8%	N/A	N/A
Enterococcus faecalis (n = 750)	0.5/1	99.9%	>32/>32	N/A
Enterococcus faecium (n = 279)	>64/>64	7.9%	>32/>32	N/A

Open in a new tab

Conclusion

Ceftobiprole is active in vitro against S. aureus (MRSA and MSSA) and Enterobacteriaceae. It is more active in vitro than ceftazidime against Acinetobacter baumannii and species with chromosomal AmpC beta-lactamases while it is less active against Klebsiella oxytoca/Raoultella sp, suggesting that its chromosomal β-lactamase (OXY) is more active against ceftopiprole than ceftazidime. While ceftobiprole appears to have anti-Pseudomonas activity, no Pseudomonas-specific breakpoints exist and PK/PD breakpoints were derived using only one dosing recommendation (500mg IV q8h). Further studies in ceftobiprole dose, interval or infusion time may reveal that a higher species-specific breakpoint for Pseudomonas could be considered, concurrent with alternative dosing recommendations. Activity is poor against Enterococcus faecium and Stenotrophomonas maltophilia.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):36–37.

L04. Antimicrobial Susceptibility Testing of Bacteria Directly from Positive Blood Culture Bottles Using Serum Separator Tubes

T Grund ¹, C Barth ¹, J Minion ¹, K Malejczyk ¹

Objective

Early targeted antimicrobial treatment can effectively reduce the mortality rate caused by bloodstream infections (BSI) and also enhance antimicrobial stewardship efforts in reducing the use of broad spectrum antibiotics. In addition to the Vitek MS® direct from blood culture bacterial identification currently practiced at our centre, the ability to perform direct susceptibility testing using the serum separator tubes (SST) has the potential to significantly reduce current turnaround time (TAT) for susceptibility results from positive blood cultures.

Method

Positive monomicrobial BACT/ALERT® blood culture bottles received in the microbiology lab between October 3 and December 31, 2018 were included. Ten milliliters of broth were aspirated from a positive blood culture bottle and injected into two 5 mL SST. The SSTs were centrifuged at 3,900 RPM for 5 minutes. The supernatant was discarded and a small amount of the remaining pellet containing the bacteria was used to make a suspension equivalent to 0.5 McFarland in 0.45% saline. Vitek 2® Susceptibility results from the pellet were compared with those performed routinely on isolates after 18-hour incubation on agar plates.

Results

100 positive blood cultures including common gram-negative and gram-positive organisms were evaluated. All reportable antibiotics had 100% essential and categorical agreement. There was one very major error for TMP-SMX for Staphylococcus aureus and 70% essential agreement but 100% categorical agreement on Nitrofurantoin for Klebsiella pneumoniae. Overall essential and categorical agreements were above the acceptable value of 90% with major and minor errors below acceptable value of 10%.

Conclusions

Based on the results of this pilot, direct from blood culture bacterial susceptibility testing using the SST provides accurate results for all of the reportable antimicrobials and significantly decreases TAT.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):37.

L05. Transient Carriage of Extended-Spectrum Beta-Lactamase (ESBL) Producing Klebsiella Pneumoniae among Healthy Recal Microbiota Transplant Stool Donors

S Hota ^1,², B Hazlett ³, J Fruitman ⁴, M Kissoon ³, SM Poutanen ^2,^3,⁴

Background

MTOP is a fecal microbiota transplantation program. Only healthy stool donors with no personal/family history of chronic illness are accepted. Negative microbiology screens including stool antimicrobial-resistant organism (ARO) cultures are required. Donors are routinely monitored and excluded if their health changes or they travel. Repeat screening is completed prior to release of donations. One donor was unexpectedly positive on repeat screening for an ESBL-producing Klebsiella pneumoniae. The purpose of this study was to determine the rate of ARO carriage among healthy stool donors.

Methods

All active MTOP stool donors between March 2017 and August 2018 were included. Stool aliquots stored at –80°C from each donation were thawed and tested for ARO [methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), ESBL, and carbapenemase-producing organisms (CPO)] by planting onto Oxoid Denim Blue agar, Oxoid Brilliance VRE agar, and Oxoid MacConkey/cefpodixime agar, respectively following clinical laboratory operating procedures.

Results

Four donors actively donated during the study period. The duration of time each participated varied from 3 months (Donors 1 and 2) to 13–14 months (Donor 3 and 4, respectively). Eighty donations were provided (3 each from Donors 1 and 2; 33 and 41 donations from Donors 3 and 4, respectively). All donors passed initial ARO screening and were well with no antimicrobial use nor travel history. Of the 80 donations, all were negative for MRSA, VRE, and CPO carriage but 3 (3.75%) were positive for ESBL Klebsiella pneumoniae [2 (6.1%) from Donor 3 and 1 (2.4%) from Donor 4]. The two positive results from Donor 3 were separated by 10 months.

Conclusion

Transient carriage of ESBLs in healthy pre-screened donors without illness, antimicrobial exposure, nor travel history suggests local transmission, possibly through food/water sources. Programs should consider screening all donations for ARO prior to acceptance into donor stool programs.

Off J Assoc Med Microbiol Infect Dis Can. 2019 Jul 11;4(Suppl):37–38.

L06. Odds of Extended-Spectrum Beta- Lactamase-Producing Enterobacteriaceae (ESBL-PE) Carriage Acquisition with Exposure to Systemic Antimicrobials During Travel: Systematic Review and Meta-Analysis

TC Wuerz ¹, K Atkins ²

Background

Recent cohort studies have identified international travel as an important risk factor for colonization with extended-spectrum beta-lactamase-producing Enterobacteriaceae (ESBL-PE), a multidrug resistant organism of public health concern. Antimicrobial use during travel likely amplifies this risk, but to what extent, and whether this risk varies by antimicrobial class, has not been well studied. A systematic review was conducted to estimate these associations.

Methods

Eligible studies were prospective cohort studies which reported on both receipt of systemic antimicrobials during travel versus none, as well as ESBL-PE isolated from stool or rectum acquired during travel. We carried out electronic searches in electronic databases. Studies were selected for full text review and included if eligible. We carried out a random effects meta-analysis.

Results

After removing duplicates, we reviewed 3,430 citations from electronic databases. Fifteen studies met inclusion criteria. The study population included mainly female travellers from high-income countries recruited primarily from travel and vaccination clinics. Asia and Africa were the most common regions travelled to. A median 10% of study participants reported systemic antimicrobial usage. We observed a combined odds ratio (OR) for ESBL-PE acquisition during travel of 2.37 among those who used antimicrobials compared with those who did not (95% CI: 1.69–3.33); there was substantial heterogeneity between studies. Fluoroquinolones were associated with the highest combined OR of ESBL-PE acquisition, at 5.55, compared with no antimicrobial use (95% CI: 2.68–11.5).

Conclusions

The odds of acquiring ESBL-PE during travel are increased substantially with exposure to antimicrobials, especially fluoroquinolones during travel. Further studies should be directed towards identifying mechanisms whereby antimicrobials affect an increased risk of ESBL-PE acquisition to identify potential protective factors. Public health efforts are warranted to decrease inappropriate antimicrobial usage during travel, including antimicrobial use for prevention or treatment of mild-to-moderate traveller’s diarrhea.

PERMALINK

Oral Presentations

A01. Epidemiology of Clostridioides difficile Infection in Canada: A Six-Year Review of Provincial Surveillance Data

Y Xia

M Tunis

C Frenette

K Katz

K Amaratunga

S Rhodenizer-Rose

A House

C Quach

Background

Methods

Results

Conclusion

A02. Detection of Pneumocystis jirovecii with qPCR to Differentiate between Colonization and Pneumonia

A Hadzic

P Jayaratne

D Leto

M Smieja

D Yamamura

Background and Objectives

Methods

Results

Conclusions

A03. Novel Metabolomics Approach for the Diagnosis of Respiratory Viruses Directly from Nasopharyngeal Specimens

CA Hogan

AT Le

TM Cowan

BA Pinsky

Objectives

Methods

Results

Conclusions

A04. Applying a Next-Generation Sequencing Pipeline in a Clinical Microbiology Laboratory for Surveillance and Outbreak Investigations: Defining the Genomic Epidemiology of Klebsiella aerogenes and Identifying Determinants of Carbapenem Resistance and Virulence

A Malek

K McGlynn

S Taffner

L Fine

B Tesini

J Wang

H Mostafa

S Petry

A Perkins

P Graman

D Hardy

N Pecora

Background

Methods

Results

Conclusion

A05. Investigation of the Presence and Impact of Environmentally-Adapted Escherichia coli in Private Well Water in Southeastern Ontario

E Tropea

M Kelly

S Brown

A Majury

Background

Methods

Results

Conclusions

B01. Withdrawn

B02. Evaluation of Five Commercial Nucleic Acid Tests for Bacterial Gastroenteritis in Alberta, Canada

BM Berenger

L Chui

C Ferrato

T Lloyd

DR Pillai

Objective

Methods

Table B02-1:

Results

Conclusions

B03. Comparison of Prevalence of Pathogenic Protozoa in Stool Samples Using Molecular Testing Compared to Traditional Microscopy

H Almohri

D Lungrin

D James

D Leto

V Tran

Background

Objective