Right patient, right specimen

Best practice criteria exist to guide diagnosis and treatment of UTI, including IDSA and Loeb Minimum Criteria for UTI (see Supplemental Table 2). ^{Reference Ashraf, Gaur and Bushen3,Reference Nicolle, Gupta and Bradley5,Reference Gupta, Hooton and Naber13,Reference Loeb, Bentley and Bradley14} Despite the existence of these criteria, dogmas surrounding UTI often lead to inappropriate urine sampling. Cloudy or malodorous urine, falls, and confusion are commonly mistaken for symptoms of a UTI. ^{Reference Nicolle, Gupta and Bradley5}

Studies evaluating urine PCR did not utilize best practice criteria for UTI. A single-site, non-inferiority study performed in the outpatient setting of SUC techniques versus PCR included participants ≥ 60 years of age reporting non-specific symptoms such as urinary incontinence, cloudy or strong-smelling urine, and agitation. ^{Reference Wojno, Baunoch and Luke15} Another study examined all urinary specimens with bacteriuria of ≥ 10⁴ CFUs/mL without accounting for presence of clinical presentation. ^{Reference Cybulski, Schmidt and Grabiec16} Van der zee et al collected urine specimens from 211 patients in the hospital and outpatient settings with and without catheters in whom a UTI was suspected but not confirmed using validated definitions. ^{Reference Van Der Zee, Roorda, Bosman and Ossewaarde17}

Obtaining a quality urine specimen is also important, as contamination can occur when proper technique is not followed. ^{Reference Nicolle, Gupta and Bradley5} Although molecular urine diagnostics should theoretically assure appropriate urine sampling, one manufacturer advertises the option of swabbing the briefs of residents who are unable to give a clean-catch sample. ^{Reference Xu, Deebel, Casals, Dutta and Mirzazadeh7,18} This technique is not an acceptable practice, nor is it endorsed by societies’ guidelines. ^{Reference Ashraf, Gaur and Bushen3,Reference Nicolle, Gupta and Bradley5,Reference Gupta, Hooton and Naber13,Reference Loeb, Bentley and Bradley14}

Confusing results

Escherichia coli causes 75–95% of uncomplicated UTIs, with other common organisms including Klebsiella pneumoniae, Proteus mirabilis, and Staphylococcus saprophyticus. ^{Reference Gupta, Hooton and Naber13} Multiplex urine PCR panels assess samples for between 9 and 31 different bacteria and some Candida spp. (see Supplemental Table 1), arguably more targets than necessary. ^{Reference Lehmann, Hauser and Malinka9,Reference Wojno, Baunoch and Luke15,Reference Cybulski, Schmidt and Grabiec16} Users may be confused with urine PCR results which are reported as organisms per quantity of urine, that is, cells/mL or copies/µL. ^10–12 The standard colony-forming units (CFUs)/mL reported with SUC considers only those cells that can actively divide under specified conditions. There is a gap in data to provide guidance on the interpretation of quantitative urine PCR results reported as cells/mL or copies/µL per organism. ^{Reference Xu, Deebel, Casals, Dutta and Mirzazadeh7,Reference Lehmann, Hauser and Malinka9–12}

Urine PCR testing is more sensitive than SUC techniques, meaning that it is more likely to generate a positive result for organisms that do not represent the infectious agent. ^{Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6,Reference Xu, Deebel, Casals, Dutta and Mirzazadeh7,Reference Lehmann, Hauser and Malinka9,Reference Wojno, Baunoch and Luke15,Reference Van Der Zee, Roorda, Bosman and Ossewaarde17} It is only minimally superior for the detection of E. coli. ^{Reference Xu, Deebel, Casals, Dutta and Mirzazadeh7,Reference Wojno, Baunoch and Luke15,Reference Cybulski, Schmidt and Grabiec16} In a non-inferiority trial, 29% of PCR samples grew E. coli compared to 34% of the same samples utilizing SUC techniques. ^{Reference Wojno, Baunoch and Luke15} A small prospective study comparing the results of urine samples using both PCR and SUC techniques showed that 64% of PCR samples were positive for E. coli compared to 58% of SUC samples. ^{Reference Cybulski, Schmidt and Grabiec16} A single-site, non-inferiority study performed in the outpatient setting collected urine samples for SUC and PCR testing from participants ≥ 60 years of age. ^{Reference Wojno, Baunoch and Luke15} Of 582 patients sampled, 56% had a positive urine PCR result versus 37% with positive culture result; among the 217 with a positive urine culture, there was 90% agreement with urine PCR, suggesting that clinicians are not missing causative UTI organisms with culture alone. ^{Reference Wojno, Baunoch and Luke15} Many of the discordant results were organisms not commonly thought to be causative UTI organisms, such as Actinobaculum schaalii and coagulase-negative Staphylococci. ^{Reference Wojno, Baunoch and Luke15}

Lack of unbiased clinical outcomes

Molecular detection of urinary organisms lacks evidence showing improved patient outcomes. ^{Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6} A systematic review and meta-analysis included six studies comparing urine PCR to urine culture. ^{Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6} None assessed patient outcomes, specifically symptom response to antibiotics started due to urine PCR testing results. ^{Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6} Moreover, they concluded that the six studies carried high risk of bias due to manufacturer funding sources. ^{Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6} Since publication, one observational study claimed an association between the use of PCR plus antibiotic susceptibility testing and a reduction in emergency department utilization and hospital admissions; however, the study population was identified retrospectively by using diagnostic codes for infections of the genitourinary tract and dysuria. ^{Reference Daly19} Molecular testing in patients with poorly defined urinary symptoms lead to misdiagnoses and overtreatment, which can lead to harms associated with antibiotic overuse. ^{Reference Ashraf, Gaur and Bushen3,Reference Szlachta-McGinn, Douglass, Chung, Jackson, Nickel and Ackerman6,Reference Krzyzaniak, Forbes, Clark, Scott, Mar and Bakhit20–Reference Harding, Zhanel, Nicolle and Cheang23}