A Real-world Data Analysis of Intermittent Catheterization, Showing the Impact of Prelubricated Versus Hydrophilic Catheter Use on the Occurrence of Symptoms Suggestive of Urinary Tract Infections

Take Home Message In a real-world data analysis of routine clinical practice, coated intermittent urinary catheters (IUCs) showed a good safety profile. Overall, the results for prelubricated (PRELUBE) and hydrophilic (HYDRO) IUCs were similar. However, when considering patients who used the same type of catheter throughout the study, PRELUBE IUCs were associated with a significantly lower mean number of symptoms suggestive of urinary tract infections. When choosing a coated IUC, physicians and patients should not rule out PRELUBE IUCs for safety reasons alone.

infection in people who were using various types of intermittent urinary catheters. When the same type of prelubricated catheter was used throughout the study period, the incidence of these symptoms was lower than for hydrophilic catheters.

Introduction
Urinary retention is an important health issue [1][2][3]. According to European and US guidelines, the first-line treatment option for chronic urinary retention is intermittent self-or heterocatheterization with single-use intermittent urinary catheters (IUCs) [4,5]. Known risk factors for urinary tract infections (UTIs) in IUC users include low catheterization frequency, bladder overdistension, female sex, inadequate fluid intake, poor technique, and absence of catheter lubrication [6,7]. In the literature, the number of suspected or proven UTIs among the users of IUCs varies greatly from one clinical study to another (from 0.13 to over 0.68 per month) [8,9].
Coated catheters (prelubricated or hydrophilic) appear to be the best option for safe, long-term intermittent catheterization [10][11][12][13]. However, literature data on the relative safety of the various types of coated IUCs are scarce. Systematic reviews and best practice reports have highlighted the lack of robust evidence for the superiority of one type of catheter or technique over another [13]. Furthermore, UTIs have been defined in different ways: clinical signs or symptoms, antibiotic prescription, positive urine culture with pyuria, etc. [8,14]. Rognoni and Tarricone's [8] meta-analysis of six randomized clinical trials (RCTs) found that the relative risk of a UTI (defined as clinical symptoms and/or culture results) was 0.84 (0.75-0.94) for hydrophilic catheters, relative to nonhydrophilic catheters.
The lack of ''gold standard'' RCT data means that it is difficult to provide guidance to clinicians or users on techniques and devices. We therefore analyzed real-world data (RWD) on routine IUC use by patients. Real-world analyses take account of patient characteristics, lifestyle factors, treatment compliance, regimens, comorbidities, and quality of life [15][16][17].
By analyzing a longitudinal patient database (LPD) compiled by 1950 general practitioners (GPs) in the UK (414 of whom prescribed IUCs), we compare prelubricated IUCs (primarily the Actreen product range from B.Braun Medical SAS, Saint-Cloud, France) and hydrophilic IUCs (primarily the SpeediCath product range from Coloplast A/S, Humlebaek, Denmark) with regard to the incidence and frequency of symptoms suggestive of UTIs (ssUTIs).

Study design and database
We performed a retrospective, longitudinal, RWD analysis with propensity score matching (PSM

Procedures
The analysis comprised four periods: a 12-mo preindex period, a 28-mo inclusion period (comprising the index date), a main (12-mo) follow-up period, and an extended (4-mo) follow-up period (Fig. 1).

Inclusion and exclusion criteria
The inclusion criteria were age 18 yr or over at the index date, at least one prescription of a prelubricated or hydrophilic catheter during the inclusion period, a preindex period of at least 1 yr, and data available for the main and extended follow-up periods. Incident and prevalent patients were, respectively, defined as those without and with a catheter prescription during the preindex periods. The exclusion criteria were as follows: (1) the prescription of two or more different types of catheters at the index date, and (2) the absence of a visit to the GP (ie, no follow-up data) during the follow-up period and the extended follow-up period.

Assessment of ssUTIs
The patients' event dates and diagnoses were mapped to the UK national thesaurus and then the International Classification of Diseases, 10th revision (ICD-10) [19], and prescription information was mapped to the Anatomical Therapeutic Chemical (ATC) classification [20].
Given the lack of urine culture data in the study database, ssUTIs were defined as either (1) the prescription of an antibiotic whose sole indication is a UTI (based on the European guidelines [4]) or (2) the prescription of a nonspecific antibiotic (ATC code) with a diagnosis (ICD-10 code) related to a UTI (Supplementary Table 1). If two ssUTIs were detected in the same patient within a 4-wk period, those were considered to be a single event.

Outcomes
The primary outcome measure was the occurrence of at least one ssUTI during the main follow-up period. Once patients with ssUTIs had been identified, the mean number of ssUTIs per affected patient was also cal-

Study groups before PSM
The retrospective data covered the period from June 1, 2015, to January 31, 2020. After applying the inclusion and exclusion criteria, data for 5296 patients were included: 458 (8.6%) with a first prescription of a prelubricated catheter at the index date (the ''PRELUBE'' group) and 4838 (91.4%) with a first prescription of a hydrophilic catheter (the ''HYDRO'' group; Table 1 and Fig. 2). Just under half of the patients had not used an IUC before the index date and were therefore incident patients. Of the prevalent patients, 70% had been using IUCs for at least a year.

Study groups after PSM
The following variables were included in a multivariable logistic regression model of the propensity score: age (p < 0.001), sex (p < 0.001), region of residence (p < 0.001), frequency of switching (p < 0.001), the main indication for IUCs (p = 0.014), and the number of ssUTIs experienced during the preindex period (p = 0.08; Table 2). Although the intergroup difference in incident versus prevalent status was not significant (p = 0.8), this important safety variable was also included in the PSM. Each of the 458 prelubricated catheter users was matched with a hydrophilic catheter user (Table 1), giving 916 users in all (females: n = 541, 59.1%; mean ± SD age: 58 ± 17). Figure 3 shows the propensity score distribution in the PRELUBE and HYDRO groups before and after PSM. After PSM, there were no statistically significant differences between the PRELUBE and HYDRO groups with regard to the score's variables (Table 1 and Fig. 3). Henceforth, all results quoted for the HYDRO group refer to the group after PSM.

3.3.
Primary outcome: incidence of ssUTIs The proportions of patients having experienced at least one ssUTI during the 12-mo follow-up period were similar in the PRELUBE versus HYDRO group (36.9% vs 41.5%; p = 0.155; Table 3). However, when considering the ''continuous use'' subpopulation, the proportion of patients with at least one ssUTI was significantly lower in the PRELUBE group than in the HYDRO group (44.6% and 55.0%, respectively; p = 0.015).
In both groups, the proportion of women with at least one ssUTI was slightly but nonsignificantly higher than Fig. 1 -Study timeline. The study comprised four periods: a 12-mo preindex period, a 28-mo inclusion period (comprising the index date), a main (12-mo) follow-up period, and an extended (4-mo) follow-up period. Among incident patients, the mean time to the first ssUTI during the main follow-up period was longer (albeit not significantly) in the PRELUBE group than in the HYDRO group (100.1 and 84.3 d, respectively; p = 0.299).

3.4.
Primary outcome: mean number of ssUTIs during exposure When considering the number of ssUTIs per affected patient, the overall mean number during the exposure period was significantly lower in the PRELUBE group than in the HYDRO group (0.9 vs 1.3, considering all the 458 patients in each matched group; p = 0.006). This difference in the mean number was also significant when considering (1) only patients with at least one ssUTI (n = 169 and n = 190 in the PRELUBE and HYDRO groups, respectively; 2.   (Table 3).

Discussion
In order to study users of prelubricated versus hydrophilic IUCs, we used PSM to form comparable PRELUBE and HYDRO groups (predominantly middle-aged/older adults, with female predominance and a third classified as overweight or obese). It is noteworthy that during the exposure period, the mean daily consumption was around three catheters (2.8-3.1) in both groups; this is lower than recommended in the guidelines, even though the LPD's GP panel is reportedly representative of UK practices. There are several possible explanations for this difference with the guidelines. First, the number of catheters prescribed is not restricted. Second, prescribing practice varies from one region to another, but most local payers follow national guidelines [25]. Third, some patients are still able to partly void their bladder through miction and so may not change their IUC five times a day [26]. Hence, the low consumption is likely to reflect routine clinical practice, rather than limited prescription by GPs. Importantly, these factors would probably affect PRELUBE and HYDRO users to the same extent. After PSM, the PRELUBE versus HYDRO difference in the proportion of patients with at least one ssUTI was not significant (36.9% vs 41.5%; p = 0.155). This finding constitutes robust evidence of similar levels of safety in prelubricated and hydrophilic IUCs. However, when comparing patients in the matched PRELUBE and HYDRO groups who used the same type of catheter throughout the exposure period, the mean number of ssUTIs per 12 mo was significantly lower in the PRELUBE group (0.9) than in the HYDRO group (1.3). The use of prelubricated catheters was still linked to a lower occurrence of ssUTIs when considering vulnerable patient populations (women and prevalent patients).
The present study of an LPD had several strengths. First, it constituted the first analysis of RWD on the clinical safety of prelubricated and hydrophilic IUCs. RWD and RCT data are considered mutually complementary, bearing in mind the different facets of both types of evidence, and realizing that RWD are based on the sample taken but nevertheless provide strong clinical applicability of RCT data [17]. Although RCTs will always be the gold standard for evaluating efficacy and safety, it is rarely possible to recruit a study population of several thousand community-dwelling patients over a short period of time. In contrast, RWD on effectiveness and safety can be gathered quickly once the extraction protocol has been configured. Patient populations, treatment patterns, follow-up monitoring, and comparator interventions are necessarily more heterogeneous in real-world studies than in RCTs [17]. The LPD analyzed in the present study reflects real-world IUC use and can provide alternative markers of safety. It is noteworthy that the data in the LPD were collected in an unbiased, noninterventional way and so reflect routine clinical practice in the GPs' surgeries. The data were entered during routine patient care and submitted on a regular basis to the coordinating center for cleaning and deidentification. This method of data collection enabled an a posteriori analysis of the patient's entire prescription and care history. Second, the study looked at pragmatic primary care management in the community by a representative panel of UK GPs and patients. Last, the proportions of users of prelubricated versus hydrophilic IUCs observed in the database reflect the UK sales data in 2020 [25].
The study also had a number of limitations, many of which were linked to its retrospective design and the use of prescription data (rather than clinical data). First, we did not have access to data on urine cultures or other reliable methods of attesting to a UTI, and could not confirm the presence of a bacteriologically confirmed UTI on the basis of the patient's LPD data. For example, our selected ICD codes would probably not have reliably identified cases of asymptomatic bacteriuria and so might have led to under-reporting of UTIs. However, antibiotics are often inappropriately prescribed when asymptomatic bacteriuria is suggested, leading to an overestimation of the frequency of UTIs [27]. Hence, we chose to refer to ''symptoms suggestive of a UTI'' on the basis of symptom-based diagnostic codes and antibiotic prescription data. Despite its disadvantages, this approach corresponds to real-life clinical practice because the diagnosis in primary care is not always confirmed by a gold standard urine culture. Furthermore, we expect that any bias in the diagnosis and treatment of asymptomatic bacteriuria would affect users of prelubricated and hydrophilic IUCs to the same extent. It has also been reported that ICD-10 diagnosis codes constitute a valid method for studying UTIs in primary care settings [28]. Second, anonymization of the patient data prevented us from querying the GPs' records; hence, there was a risk of under-reporting. Third, a selection bias was possible because data on patients managed by specialist physicians and/or in hospital were not collected. Accordingly, we included an extended follow-up period to ensure that the patient was still seeing his/her GP, and that his/her final status (''discontinuation'', ''stop'', ''switch'', or ''continuous'') could be determined. Fourth, the use of PSM always represents a tradeoff between bias reduction on one hand and precision on the other [23,24]. Fifth, it is possible that the administration of certain medications (eg, analgesics) masked some episodes. However, it is unlikely that this type of medication use would differ significantly between matched PRELUBE and HYDRO users. Although a poor selfcatheterization technique per se is well known to be associated with a greater frequency of UTIs, the two types of UTIs have very similar gross characteristics and selfcatheterization techniques, and so a difference between the PRELUBE and HYDRO groups is unlikely. Last, the study was limited to an LPD from the UK; it would be interesting to perform similar studies in countries or regions with different healthcare, social care, and reimbursement systems. Despite these limitations, we consider that an RWD analysis was perfectly relevant for measuring product effectiveness and safety in routine clinical practice. A patient and his/her GP are likely to consider many criteria when selecting a suitable coated IUC: design, compactness, ease of use, comfort, brand awareness, cost, and reimbursement; a holistic approach should be adopted. Cost is unlikely to have been a confounding factor in the present study. In the UK, IUCs are fully reimbursed for a large proportion of patients. Even when a patient is not exempted, the reimbursement level and thus the cost are similar for the PRE-LUBE and HYDRO IUCs. Likewise, GPs in the UK do not have financial reasons for preferring PRELUBE IUCs to HYDRO IUCs or vice versa.
Lastly, our results raise a number of important questions. What would be the results of similar studies in groups of patients with bacteriologically confirmed UTIs? Moreover, what are the characteristics of particular subtypes of prelubricated and hydrophilic catheters?

Conclusions
In an RWD analysis of routine clinical practice, coated IUCs showed a good safety profile. Overall, the results for PRE-LUBE and HYDRO IUCs were similar; hence, when choosing a coated IUC, physicians and patients should not rule out PRELUBE IUCs for safety reasons. Indeed, when considering patients who used the same type of catheter throughout the study, PRELUBE IUCs were even associated with a significantly lower mean number of ssUTIs.
Author contributions: Emmanuel Chartier Kastler had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.  Funding/Support and role of the sponsor: The study was funded by B.
Braun Medical (Saint-Cloud, France). The funding party did not influence the interpretation or reporting of the present research.