Cystitis

Six E. coli strains as well as the type strain (ATCC 35469^T) of E. fergusonii, the closest E. coli-related species [31] (link), were selected for complete genome sequencing (Table 1). Among the E. coli strains, 2 were commensal: IAI1 (serogroup O8) was isolated from the faeces of a young healthy military conscript in the 1980s in France [23] (link) and ED1a (serogroup O81) was isolated in the 2000s from the faeces of a healthy man in France and belongs to a human-specific widespread commensal clone that is increasing in frequency [55] (link). Four E. coli strains were pathogenic. Enteroaggregative E. coli strain 55989 was originally isolated from the diarrheagenic stools of an HIV-positive adult suffering from persistent watery diarrhea in Central African Republic [86] (link). The enteroaggragative pathotype is recognized as an emerging cause of diarrhoea in children and adults worldwide [87] (link). Among the three extraintestinal pathogenic strains, IAI 39 (serotype O7:K1) was isolated from the urine of a patient with pyelonephritis in the 1980s in France [23] (link). UMN026 (serotype O17:K52:H18) was isolated from a woman with uncomplicated acute cystitis in 1999 in the USA (Minnesota) and is a representative of a recently emerged E. coli clonal group (“clonal group A”) that is now widely disseminated and a cause of drug-resistant urinary tract and other extraintestinal infections [88] (link). S88 (serotype O45:K1:H7) was isolated in 1999 from the cerebro-spinal fluid of a new born with late-onset neonatal meningitis in France and represents what is now considered a highly virulent emerging clone in France [89] (link). These strains were distributed in 3 of the 4 main E. coli phylogenetic groups: IAI1 and 55989 belong to group B1, UMN026 and IAI392 belong to each of the two major subgroups within group D, and ED1a and S88 belong to subgroups VIII and IX, respectively, within group B2 [42] (link). Few data are available on E. fergusonii strains. They have been isolated from humans and warm blood animals, sometimes in pathogenic (intestinal and extraintestinal) conditions [90] (link)–[92] . The main characteristics of the 14 strains (8 E. coli sensu strictu and 6 Shigella) with freely available genomes at the time of the study are presented in Table 1. These genomes were used for comparison purpose.

GPs participating in the JGPN register a diagnosis for each consultation using International Classification of Primary Care (ICPC) codes. As a first step, medical record data were collected from all consultations of adult males with ICPC codes U01 (dysuria/painful urination), U02 (urinary frequency/urgency), U70 (pyelonephritis/pyelitis), U71 (cystitis), and Y73 (prostatitis) for which a UTI-related antibiotic (nitrofurantoin, trimethoprim, ciprofloxacin, amoxicillin/clavulanic acid, trimethoprim/sulfamethoxazole, fosfomycin, or norfloxacin) was prescribed within 7 days before and after the start date. A UTI episode after a UTI consultation-free interval of 30 days was considered a new UTI episode. Second, UTI episodes were excluded if:

the patient had an indwelling urinary catheter;

no information on patient-reported symptoms and GP-assessed signs was available;

one UTI episode was associated with two prescriptions for different antibiotics on the same day;

the antibiotic prescription was for prophylactic or future use, and

the patient was anatomically female.

Next, medical records of all remaining UTI episodes were manually screened and scored for patient-reported symptoms and GP-assessed signs of a complicated UTI, namely fever reported by patients and/or assessed  by  GPs,  patient-reported  malaise  and/or cold shivers, and GP-assessed ‘clinically severely ill’, costovertebral angle tenderness, perineal pain, and/or signs of delirium.² A UTI was considered uncomplicated when all of these signs and symptoms were absent. If a sign or symptom was not recorded it was considered absent. Screening and scoring of 400 variables regarding signs and symptoms of complicated UTI was done in duplicate by two independent investigators with an agreement of 99% (kappa 0.94 [almost perfect agreement]).
The following data were extracted for each uncomplicated UTI episode: age, comorbidities (diabetes, cardiovascular, pulmonary, oncological, nephrological, urological, and neurological and immunocompromising conditions; see Supplementary Appendix S1), number and type of antibiotic prescriptions using Anatomical Therapeutic Chemical codes, and hospital referrals.
Treatment failure was defined as an antibiotic prescription for a different antibiotic >1 day after the initial prescription or an acute hospital referral to urology or internal medicine >1 day after the antibiotic prescription.
The Dutch UTI guideline recommends nitrofurantoin (first choice) or trimethoprim (second choice) for treatment of uncomplicated UTI in males.² GPs were therefore considered adherent to the guideline if they prescribed one of those antibiotics. If treatment failure had occurred in the year before the current UTI episode, ciprofloxacin, amoxicillin/clavulanic acid, and trimethoprim/sulfamethoxazole were considered adherent as well.
As data on UTI episodes from 2013 were not available, treatment failure in the year before 2014 could not be determined. The year 2014 was therefore excluded from the adherence analysis.