Treatment-Related Cystitis During Intravesical Chemotherapy: Incidence, Risk Factors, And Nursing-Pathway Associations

Summary

Treatment-related cystitis is common during intravesical chemotherapy and may compromise treatment completion. This retrospective single-center cohort quantified cystitis incidence, identified associated clinical and procedural factors, and examined whether a preventive nursing pathway was associated with lower cystitis burden and fewer treatment interruptions.

Abstract

Intravesical chemotherapy is routinely used after transurethral resection of bladder tumors for non-muscle-invasive bladder cancer, but treatment-related cystitis (TRC) may disrupt planned care. This retrospective, single-center, observational cohort study quantified TRC incidence, identified associated clinical and procedural factors, and examined the association between a preventive nursing pathway and TRC-related outcomes. We analyzed 120 consecutive patients receiving intravesical chemotherapy and compared patients managed under standard care (n = 60) with those managed under a preventive nursing pathway (n = 60) according to routine clinical pathway assignment. TRC was assessed within 7 days after each documented instillation, and interruption-free completion was evaluated during the documented intravesical treatment course. TRC events were captured from electronic medical records using predefined symptom documentation and follow-up records. Risk factors were evaluated using multivariable regression, and time to event analysis was used to assess treatment interruption attributable to TRC. TRC occurred in 45.8% of patients. A shorter interval from transurethral resection to first instillation (OR = 0.83 per 7-day increase), multiple catheterization attempts (OR = 1.41), and higher baseline symptom burden were associated with increased TRC risk. After adjustment, management under the preventive nursing pathway was associated with lower odds of TRC (aOR = 0.54, 95% CI: 0.29–0.98; P = 0.043). As this study was non-randomized and based on routine care pathway assignment, these findings should be interpreted as adjusted associations rather than evidence of a causal intervention effect. The findings suggest that TRC is a substantial real-world burden during intravesical chemotherapy and that structured nursing documentation, baseline assessment, catheterization quality control, symptom monitoring, and early escalation may support risk-stratified follow-up. As no direct mucosal or biomarker measurements were obtained, urothelial barrier injury is discussed as a plausible mechanism rather than a demonstrated effect.

Introduction

Non-muscle-invasive bladder cancer (NMIBC) is a heterogeneous disease with a high recurrence rate, and postoperative management therefore focuses on reducing recurrence while limiting progression to muscle-invasive disease^1,2. Transurethral resection of bladder tumors (TURBT) followed by intravesical treatment remains a central strategy in postoperative management. Intravesical management includes chemotherapy agents, such as gemcitabine, mitomycin C, and pirarubicin, as well as immunotherapeutic approaches such as Bacillus Calmette-Guérin (BCG)^3,4. The present study focuses specifically on patients receiving intravesical chemotherapy; BCG is mentioned only as part of the broader intravesical treatment context and is not classified as chemotherapy.

Local exposure to intravesical agents may provoke lower urinary tract symptoms, including urinary frequency, urgency, dysuria, suprapubic discomfort, nocturia, and hematuria. In this study, these post-instillation irritative events are referred to as treatment-related cystitis (TRC) when they occurred within the predefined observation window and were documented in the electronic medical record. Previous literature has proposed that intravesical exposure may contribute to urothelial irritation through mucosal inflammation and possible disruption of bladder barrier function, including the glycosaminoglycan layer⁵. As this study did not directly measure mucosal biomarkers, urothelial barrier injury is treated as a plausible explanatory mechanism rather than a directly observed endpoint. TRC remains clinically relevant because it can affect symptom burden, patient tolerance, and continuity of planned instillation schedules⁶.

TRC also creates practical challenges in routine care. Irritative symptoms after instillation may reduce quality of life and increase treatment-related worry before subsequent instillations^7,8. Chemical or non-infectious cystitis can also resemble bacterial urinary tract infection, making clinical differentiation difficult when symptom records, urinalysis, and urine culture results are incomplete or inconsistent. This overlap may contribute to empirical antibiotic use in some settings, although antibiotic-related conclusions require cautious interpretation unless infection evidence is clearly documented⁹. These issues make it important to describe TRC using operational definitions that distinguish symptom burden, infection-related events, medication use, and treatment interruption.

Randomized trials and comparative studies have provided important evidence on the oncological efficacy and toxicity profile of intravesical therapies, but trial populations and tightly controlled protocols may not fully reflect the variation encountered in routine clinical practice. Real-world electronic medical record data can help characterize how TRC is documented and managed in everyday care, particularly among patients with comorbidities, baseline urinary symptoms, recent postoperative recovery, or procedure-related difficulties^10,11. However, because the present study was retrospective, single-center, and modest in sample size, its findings should be understood as context-specific observational evidence rather than broadly generalizable causal evidence.

Accordingly, this study used retrospectively extracted clinical and nursing documentation from a single center to estimate TRC incidence, describe timing and severity patterns, identify clinical and procedural factors associated with TRC, and examine whether management under a preventive nursing pathway was associated with lower TRC burden and better treatment continuity. The study was designed to generate practical, auditable evidence for nursing-process optimization rather than to prove a randomized intervention effect. By clarifying high-risk points such as early post-TURBT instillation, repeated catheterization attempts, baseline symptom burden, and infection-risk documentation, the analysis aims to support structured symptom monitoring, escalation, and treatment-continuity management during intravesical chemotherapy¹².

Protocol

The study was approved by the Ethics Committee of the Second Affiliated Hospital of Wenzhou Medical University (approval number: B2025461X; approval date: January 10, 2025). The requirement for individual informed consent was waived because the study used retrospectively collected routine-care records and involved no direct patient contact or change to clinical management. All patient identifiers, including names, medical record numbers, telephone numbers, and addresses, were removed before analysis. The de-identified dataset was stored in password-protected files accessible only to authorized study personnel.

Study design and data sources

This was a retrospective, single-center, observational cohort study of patients receiving intravesical chemotherapy for non-muscle-invasive bladder cancer. The study quantified treatment-related cystitis (TRC), identified clinical and procedural factors associated with TRC, and examined whether management under a preventive nursing pathway was associated with TRC-related outcomes. Because pathway assignment was not randomized, all pathway-related findings were interpreted as adjusted associations rather than causal intervention effects.

Patients were screened from January 1, 2023, to December 31, 2024. The index date was the first eligible intravesical chemotherapy instillation during the study period. TRC was assessed within 7 days after each documented instillation. Interruption-free completion was evaluated from the index instillation until planned completion, TRC-related delay or discontinuation, discontinuation for non-TRC reasons, loss to follow-up, or the administrative end of follow-up on January 31, 2025. Data were extracted from January 15, 2025, to February 28, 2025.

Data sources included electronic medical records, chemotherapy orders, instillation execution logs, nursing assessment records, catheterization records, education notes, follow-up records, laboratory reports, urinalysis results, and urine culture results. The conceptual data-extraction framework linking intravesical chemotherapy exposure, procedural factors, TRC phenotypes, nursing-process nodes, and downstream care outcomes is shown in Figure 1.

Participant eligibility and cohort construction

Eligible patients were adults aged 18 years or older who received at least one intravesical chemotherapy instillation for non-muscle-invasive bladder cancer between January 1, 2023, and December 31, 2024. Disease risk categories were extracted from postoperative pathology reports and the treating urologist's documented risk assessment in the electronic medical record. In this study, low-risk disease was defined as solitary, low-grade Ta tumors without carcinoma in situ or other high-risk features. Intermediate-risk disease refers to NMIBC that did not meet low-risk or high-risk criteria, including recurrent, multifocal, or larger low-grade Ta tumors, or other cases documented by the treating urologist as intermediate risk. Selected high-risk disease referred to patients with high-risk NMIBC features, such as high-grade Ta/T1 disease or carcinoma in situ, for whom intravesical chemotherapy, rather than BCG, was recorded as the planned intravesical treatment because of clinical judgment, patient suitability, BCG unavailability, intolerance, contraindication, or local treatment planning. Patients receiving Bacillus Calmette-Guérin alone or other non-chemotherapy intravesical immunotherapy regimens were excluded.

Patients were required to have traceable chemotherapy orders, instillation execution records, nursing records on the instillation date, and symptom or follow-up documentation within 7 days after the index instillation. Each patient was included once. If a patient had more than one eligible treatment course, the first course during the screening period was used as the index course, and repeated encounters were linked under the same patient identifier.

Patients were excluded if they had active urinary tract infection, fever, or physician-documented infection requiring postponement of instillation on the treatment day; had systemic antibiotics for active urinary tract infection within 7 days before the index instillation; received BCG-only or other non-chemotherapy intravesical regimens; lacked records needed to classify exposure, outcome, or group assignment; or had no follow-up documentation within the 7-day outcome window.

Eligible patients were classified into the standard-care group or preventive nursing-pathway group according to the routine clinical pathway recorded in the electronic medical record. The allocation mechanism was a before-and-after implementation process. Patients treated before implementation of the structured preventive nursing pathway were assigned to the standard-care group, and patients treated after implementation were assigned to the preventive nursing pathway group. Baseline and procedural variables were compared and adjusted for to reduce selection and temporal confounding. The screening and group-allocation process is shown in Figure 2.

Definition of intravesical chemotherapy exposure and procedural factors

Intravesical chemotherapy exposure was extracted from chemotherapy orders and instillation execution records. Variables included drug type, dose, preparation volume, administration route, planned retention time, actual retention time, treatment phase, and cumulative number of instillations. For each instillation, the administration record was reviewed for the location of administration, catheterization method, drug preparation volume, route of delivery, infusion procedure, planned dwell time, actual dwell time, voiding instruction, and whether the patient completed the planned retention period. Intravesical chemotherapy was administered in the urology treatment room, outpatient procedure area, or inpatient urology ward according to the patient's treatment setting. After bladder drainage through a sterile urinary catheter, the prescribed chemotherapy solution was instilled into the bladder through the catheter by slow manual injection or gravity-assisted instillation according to routine departmental practice. The catheter was then removed or clamped according to the recorded method, and patients were instructed to retain the solution for the planned dwell time unless severe discomfort, hematuria, fever, or other clinically concerning symptoms occurred. Eligible chemotherapy agents were gemcitabine, pirarubicin, mitomycin C, and epirubicin. Bacillus Calmette-Guérin was not classified as chemotherapy.

For patients who underwent transurethral resection of bladder tumors, the interval from surgery to first instillation was calculated in days. This interval was analyzed as a continuous variable per 7-day increase and categorized as ≤7 days, 8–14 days, and >14 days for descriptive and time-to-event analyses.

Catheterization attempts were counted for each instillation. One successful catheter insertion was coded as one attempt. Withdrawal followed by reinsertion or any repeated documented insertion attempt was counted as an additional attempt. Difficult catheterization was defined as repeated attempts, resistance during insertion, need for senior nurse or physician assistance, catheter-size change due to insertion difficulty, or documented interruption because of discomfort. Catheter-related visible bleeding was defined as gross hematuria, blood-stained urine, or visible blood on the catheter or drainage material during or immediately after catheterization.

Catheterization method, catheter size, intermittent or indwelling catheter use, bladder drainage or irrigation, retention time, procedural discomfort, catheterization attempts, difficult catheterization, and catheter-related bleeding were extracted when documented. When exact administration speed was not numerically recorded, it was coded as undocumented rather than assumed. For reproducibility, the extracted administration variables included whether instillation was documented as slow manual injection, gravity-assisted instillation, or unspecified administration; whether the bladder was drained before drug delivery; whether the catheter was removed or clamped after instillation; and whether early voiding occurred before completion of the planned dwell time.

Standard-care pathway

Standard care was the routine nursing pathway used before the implementation of the structured preventive nursing pathway. It included physician-order verification, patient identity confirmation, routine inquiry about fever or severe urinary symptoms, aseptic catheterization, administration of the prescribed intravesical chemotherapy agent, instruction on retention time, and general advice to report severe discomfort, fever, gross hematuria, or other concerning symptoms.

Nursing documentation under standard care was mainly narrative and task-based rather than checklist-based. The nursing record usually documents order verification, patient identification, completion of catheterization, drug administration, general tolerance during the procedure, and brief post-instillation instructions when recorded. Documentation of baseline urinary symptoms, catheterization difficulty, number of catheterization attempts, visible bleeding, retention-time completion, and post-instillation symptom follow-up was not required in a standardized format. These items were extracted only when they appeared in narrative nursing notes, treatment execution records, outpatient records, or follow-up documentation.

Baseline symptom assessment in standard care was usually recorded in narrative nursing notes. Urine testing was performed according to the physician order or routine department practice. Patient education usually included retention-time instructions, hydration advice, and warning symptoms, but checklist-based education was not required. Follow-up was performed through usual outpatient or nursing practice, and structured 24–72 h follow-up was not mandatory. Physician escalation was based on nurse judgment, patient-initiated contact, fever, gross hematuria, severe dysuria, or treatment intolerance.

Preventive nursing bundle and implementation fidelity

The preventive nursing pathway consisted of five modules. It was defined as "preventive" because the pathway was implemented before, during, and shortly after each intravesical chemotherapy instillation to identify baseline urinary risk, reduce avoidable catheterization-related irritation, standardize patient education, detect early post-instillation symptoms, and trigger clinician review before symptoms resulted in treatment delay or interruption¹³. The pathway was not intended to replace physician-directed treatment or to prove a causal intervention effect. Rather, it translated intravesical-therapy safety recommendations and postoperative instillation nursing principles into auditable nursing steps¹⁴. Unlike standard care, which relied mainly on narrative documentation, patient-initiated contact, and nurse judgment, the preventive pathway required structured checklist completion, predefined escalation criteria, fixed-window follow-up, and documented clinician-nurse communication.

Before implementation, urology nurses received department-level training on the pathway. Training covered pre-instillation risk screening, urine-status review, catheterization-quality documentation, symptom recognition, patient-education content, 24–72 h follow-up recording, and physician-escalation triggers. The training was delivered by the urology nursing lead together with a supervising urologist during routine departmental training sessions. Nurses were instructed to complete the checklist at the instillation level, and checklist completion was reviewed during early implementation to improve consistency of documentation.

Module 1: Pre-instillation risk assessment and infection screening. Before each instillation, the responsible urology nurse assessed baseline urinary symptoms, fever, dysuria, gross hematuria, urinary tract infection history within 12 months, diabetes, immunosuppression, recent antibiotic exposure, TURBT-to-instillation interval, and urinalysis results when available. The assessment was recorded using a structured pre-instillation checklist rather than only narrative nursing notes. The checklist included body temperature status, current urinary frequency, urgency, dysuria, nocturia or suprapubic discomfort, visible hematuria, recent antibiotic exposure, history of urinary tract infection, diabetes or immunosuppression, recent catheterization difficulty, and available urine-test findings. Fever, suspected infection, severe baseline urinary symptoms, gross hematuria, or abnormal urine findings triggered physician review before instillation¹⁵. This module differed from standard care because standard care included routine inquiry about fever or severe symptoms and urine testing when ordered, whereas the preventive pathway required documented risk screening before each instillation.

Module 2: Standardized catheterization and instillation. Nurses performed hand hygiene, aseptic preparation, drug and dose verification, catheter-size selection, gentle catheter insertion, documentation of catheterization attempts, documentation of visible bleeding or marked discomfort, and escalation after difficult catheterization. The catheterization procedure included confirmation of patient identity and chemotherapy order, privacy protection, sterile field preparation, hand hygiene, urethral disinfection, selection of an appropriate catheter size according to sex, anatomy, and prior catheterization difficulty, slow catheter insertion without force, complete bladder drainage before drug delivery, and documentation of whether catheter insertion was smooth or difficult. The instillation record captured catheter type and size when documented, the number of catheterization attempts, resistance during insertion, patient discomfort, visible bleeding, bladder drainage before instillation, route of drug delivery through the catheter, and whether the catheter was removed or clamped after instillation. If two attempts were unsuccessful, marked resistance occurred, or visible bleeding was observed, repeated manipulation was stopped, and senior nurse or physician review was required¹⁶. This module differed from standard care because routine care mainly documented completion of catheterization and drug administration, whereas the preventive pathway required structured recording of catheterization-quality indicators and a predefined stop-and-review rule.

Module 3: Post-instillation education. Immediately after instillation, patients received instructions on planned retention time, voiding precautions, hydration when clinically appropriate, expected mild irritative symptoms, warning symptoms, contact procedures, and avoidance of self-initiated antibiotic use without clinical evaluation. The education checklist included the prescribed dwell time, what to do if the patient could not retain the drug for the planned period, voiding precautions after the retention period, hydration advice when not contraindicated, expected mild urinary frequency or burning, and warning symptoms requiring clinical contact. Warning symptoms included fever, gross hematuria, severe dysuria, severe bladder pain or spasm, persistent or worsening symptoms beyond 48–72 h, inability to void, suspected infection, or symptoms affecting willingness to continue treatment. Patients were also informed how to contact the outpatient unit, ward nurse, or treating clinical team and were advised not to start antibiotics without clinical assessment or urine testing when possible¹⁷. This module differed from standard care because education under standard care was usually general and narrative, whereas the preventive pathway required checklist-based documentation of retention-time instruction, expected symptoms, warning symptoms, contact route, and antibiotic-use advice.

Module 4: Structured 24–72 h follow-up. Follow-up was completed through outpatient review, telephone contact, or electronic documentation within 24–72 h after instillation. The follow-up checklist recorded urinary frequency, urgency, dysuria, nocturia, suprapubic discomfort or bladder spasm, hematuria, fever, medication use, and whether symptoms affected daily activity or willingness to continue treatment. The follow-up format was standardized. It recorded symptom presence, approximate onset time, symptom severity, persistence or worsening, fever status, hematuria status, early voiding before completion of the planned dwell time, symptom-relief medication use, antibiotic use, unplanned outpatient or emergency contact, and whether the next instillation required physician review or possible delay. Follow-up was completed by the responsible urology nurse or a trained urology follow-up nurse. Abnormal findings were communicated to the treating urologist or on-duty urology clinician according to predefined escalation criteria¹⁸. This module differed from standard care because standard care did not require a fixed 24–72 h follow-up window or a structured symptom checklist after every instillation.

Module 5: Tiered management and clinician-nurse collaboration. Physician review was required for fever, gross hematuria, suspected infection, severe dysuria or bladder pain, persistent or worsening symptoms beyond 48–72 h, grade ≥2 TRC, positive urine culture, empirical antibiotic initiation, emergency visit, or planned delay or interruption of the next instillation. The review process followed a structured nurse-to-clinician escalation format. The nurse summarized the instillation date, chemotherapy agent, planned and actual dwell time when documented, early voiding, catheterization difficulty or visible bleeding, current symptoms, fever status, available urinalysis or culture results, medication use, and whether treatment delay, urine testing, symptom-relief treatment, antibiotic assessment, or emergency evaluation was needed. One responsible nurse initiated the review, and the treating urologist or on-duty urology clinician made the clinical decision. When the first nurse encountered difficult catheterization or unclear symptom severity, a senior nurse could be asked to reassess before physician escalation¹⁹. This module differed from standard care because escalation in standard care depended mainly on nurse judgment or patient-initiated contact, whereas the preventive pathway used predefined review triggers and a structured information handover.

Implementation fidelity was scored at the instillation level using five checklist items: pre-instillation risk assessment, infection screening or urine-status review, standardized catheterization documentation, post-instillation education, and structured 24–72 h follow-up. Each item was scored as 1 if completed and documented and 0 if absent or undocumented. The total score ranged from 0 to 5. High fidelity was defined as 4–5 points, and low fidelity was defined as 0–3 points. Missing checklist items were scored as 0 unless completion was clearly documented elsewhere in the record. Fidelity scoring was based on nursing documentation and was completed before outcome modeling. Because the study was retrospective, fidelity was assessed from documented evidence only; undocumented completion was not assumed. The five fidelity items were selected because they represented the main operational differences between the preventive pathway and standard care: mandatory pre-instillation screening, documented urine-status review, structured catheterization-quality recording, checklist-based education, and fixed-window post-instillation follow-up. This scoring method was used to evaluate documentation-based implementation completeness rather than individual nurse performance²⁰.

Definition and grading of treatment-related cystitis

The primary outcome was TRC, defined as new-onset or clearly aggravated lower urinary tract irritative symptoms within 7 days after intravesical chemotherapy. Eligible symptoms included urinary frequency, urgency, dysuria, suprapubic discomfort or bladder spasm, increased nocturia, gross hematuria, or clinically meaningful irritative discomfort documented in medical, nursing, outpatient, telephone follow-up, or treatment-delay records.

TRC severity was graded using a CTCAE-style study-specific rule based on symptom burden, functional impact, and intervention requirement because no universally adopted grading system exists specifically for chemotherapy-related cystitis documented retrospectively from routine nursing and electronic medical records²¹. Grade 1 TRC was mild irritative symptoms without prescription medication, unplanned visit, treatment delay, or functional limitation. Functional limitation was defined as documentation that urinary symptoms interfered with sleep, usual daily activity, mobility outside the home or ward, ability to attend work or routine activities, willingness to continue instillation, or ability to complete the planned retention period. Grade 2 TRC was moderate symptoms requiring symptom-relief medication, additional clinical contact, urine testing, or documented nursing or physician management, without emergency care or treatment discontinuation. Grade 3 TRC was severe symptoms resulting in an emergency visit, unplanned invasive urological evaluation, instillation delay of at least 7 days, treatment interruption, discontinuation, or physician-documented inability to continue the planned schedule. Unplanned invasive urological evaluation referred to symptom-driven catheter reinsertion, bladder irrigation, cystoscopic evaluation, or other unplanned instrument-based urinary tract assessment performed because of severe hematuria, urinary retention, suspected clot retention, severe bladder pain, or persistent symptoms after instillation. Routine catheterization for the planned chemotherapy procedure was not counted as an invasive evaluation for TRC grading. Grade ≥2 TRC was defined as clinically significant TRC because it reflected symptoms requiring medication, additional clinical assessment, urine testing, nursing or physician management, treatment delay, interruption, or emergency-level care²².

Outcome adjudication was performed by two trained clinical data extractors using predefined rules. Both extractors were trained before data abstraction using the study codebook, which specified the 7-day outcome window, eligible TRC symptoms, severity-grade criteria, infection-related phenotype criteria, care-utilization variables, and rules for handling ambiguous or incomplete documentation. The first extractor screened medical, nursing, outpatient, telephone follow-up, laboratory, and treatment-delay records. A second extractor independently reviewed all grade ≥2 events, infection-related events, treatment-delay events, and ambiguous cases. Ambiguous cases were reviewed by a second investigator and resolved by consensus with a urology clinician. The adjudicating clinician was a urologist or urology clinician with experience in intravesical therapy and was not responsible for assigning patients to the standard-care or preventive nursing-pathway group. The adjudicator reviewed the extracted evidence, symptom timing, urine-test findings, medication records, treatment-delay notes, and clinician documentation to determine whether the event met the TRC definition and severity grade. Adjudicators were not involved in the original care-pathway assignment. Complete blinding to group assignment was not always possible because nursing records could indicate pathway elements. To reduce classification bias, the adjudication form required extractors to apply the same symptom window, severity thresholds, infection-related criteria, and treatment-delay rules to both groups. Disagreements were resolved by consensus, and no outcome was upgraded solely because the patient belonged to the preventive nursing-pathway group.

TRC-related care utilization included symptom-relief medication use, cystitis-relevant antibiotic use, unplanned outpatient or emergency visits, symptom-driven instillation delay, and treatment interruption or discontinuation attributed to cystitis. A care-utilization event was attributed to TRC only when the record linked the medication, visit, delay, interruption, or discontinuation to post-instillation urinary symptoms, suspected cystitis, hematuria, bladder pain, or infection-related evaluation within the defined outcome window.

Infection-related phenotype classification

TRC events were classified as infection-related or non-infectious. This phenotype classification was used to separate post-instillation irritative symptom events with supportive infection evidence from events more consistent with chemical or non-infectious bladder irritation²³. An event was classified as infection-related if at least one of the following was documented within 7 days after instillation: positive urine culture meeting the local laboratory threshold for clinically significant bacteriuria; physician-diagnosed urinary tract infection supported by urinalysis or culture evidence; fever with urinary symptoms and antibiotic treatment; or antibiotic therapy explicitly prescribed for suspected or confirmed urinary infection with supportive clinical documentation.

Urinalysis evidence included pyuria, positive nitrite, positive leukocyte esterase, or other laboratory indicators interpreted by the treating clinician as infection-supporting evidence. Empirical antibiotic use without culture confirmation was not automatically classified as infection-related unless the record documented a physician-assessed urinary tract infection or supportive urinalysis findings. Irritative symptoms in this classification referred to urinary frequency, urgency, dysuria, nocturia, suprapubic discomfort, bladder spasm, irritative bladder pain, or gross hematuria occurring after intravesical chemotherapy²⁴. Events with these irritative symptoms but no documented fever, positive urine culture, infection-supporting urinalysis, physician-diagnosed urinary tract infection, or infection-directed antibiotic rationale were classified as non-infectious or chemical TRC. Events with incomplete infection evidence were retained in the primary TRC analysis and examined in sensitivity analyses. This approach was used to avoid misclassifying all post-instillation urinary discomfort as urinary tract infection, while still preserving uncertain cases in the primary safety analysis.

Baseline symptom burden, infection risk, and covariates

Baseline lower urinary tract symptoms (LUTS) burden was assessed from pre-instillation clinical or nursing records. Because this retrospective study used routine electronic medical records rather than a prospectively administered validated symptom questionnaire, the LUTS score was reconstructed as a pragmatic documentation-based symptom score. It was designed to capture baseline urinary symptom burden before intravesical chemotherapy rather than to replace validated LUTS instruments²⁵. The LUTS score ranged from 0 to 12 and included four domains: urinary frequency, urgency, dysuria, and nocturia or suprapubic discomfort. Each domain was scored from 0 to 3: 0, absent; 1, mild; 2, moderate; and 3, severe or affecting daily activity. A score of 1 was assigned when the symptom was documented as present but mild, occasional, or not affecting usual activity; a score of 2 was assigned when the symptom was described as moderate, recurrent, requiring attention, or causing noticeable discomfort; and a score of 3 was assigned when the symptom was documented as severe, persistent, sleep-disrupting, activity-limiting, or affecting willingness to proceed with treatment. Total scores were categorized as mild, 0–3; moderate, 4–7; and severe, 8–12. Moderate-to-severe baseline symptom burden was defined as a score of 4 or higher. When symptom severity was not explicitly graded in the record, the lowest severity level supported by the documentation was assigned. If a symptom was not mentioned in the pre-instillation record, it was scored as absent only when the note clearly indicated no urinary symptoms; otherwise, missing symptom documentation was treated as unavailable rather than assumed absent. Because the score was reconstructed from routine documentation rather than a formally validated questionnaire, it was treated as a pragmatic clinical symptom score and used for risk adjustment and subgroup description rather than as a validated patient-reported outcome measure²⁶.

High baseline infection risk was defined as the presence of at least two of the following before the index instillation: urinary tract infection within 12 months, diabetes mellitus, immunosuppression or long-term corticosteroid or immunosuppressive use, baseline pyuria or positive urine infection marker, indwelling catheterization, or physician-documented elevated infection risk. These factors were selected because they are routinely available in clinical records and may indicate increased susceptibility to urinary infection, post-instillation urinary symptoms, or diagnostic uncertainty between infection-related and non-infectious cystitis²⁷. Baseline pyuria or a positive urine infection marker included documented pyuria, positive leukocyte esterase, positive nitrite, or other urine-test findings interpreted by the treating clinician as supporting possible infection. Physician-documented elevated infection risk was accepted only when the record explicitly described the patient as having increased infection risk, recurrent urinary infection risk, immunocompromised status, or a need for closer infection surveillance.

Covariates included age, sex, diabetes, immunosuppression, prior urinary tract infection within 12 months, prior pelvic radiotherapy, baseline LUTS score, baseline infection-risk status, non-muscle-invasive bladder cancer (NMIBC) risk category, intravesical agent, treatment phase, dose or volume, retention-time category, cumulative instillation number, TURBT-to-first-instillation interval, catheterization method, catheterization attempts, difficult catheterization, visible bleeding, documentation completeness, pre-instillation assessment, education delivery, structured 24–72 h follow-up, symptom scoring, and timely escalation. These covariates were grouped into five domains before modeling: demographic and baseline clinical factors, tumor and treatment-related factors, baseline urinary and infection-risk factors, procedural factors, and nursing-process factors. Demographic and baseline clinical factors included age, sex, diabetes, immunosuppression, prior urinary tract infection, and prior pelvic radiotherapy. Tumor and treatment-related factors included NMIBC risk category, intravesical agent, treatment phase, dose or volume, retention-time category, cumulative instillation number, and TURBT-to-first-instillation interval. Baseline urinary and infection-risk factors included the reconstructed LUTS score and baseline infection-risk status. Procedural factors included catheterization method, catheterization attempts, difficult catheterization, and visible bleeding. Nursing-process factors included documentation completeness, pre-instillation assessment, education delivery, structured 24–72 h follow-up, symptom scoring, and timely escalation. These variables were included because they could plausibly influence TRC occurrence, symptom detection, treatment tolerance, or pathway assignment in this non-randomized retrospective cohort²⁸.

Missing-data handling

Missingness was assessed for all baseline, exposure, procedural, nursing-process, and outcome variables. Data completeness was checked after linkage of electronic medical records, chemotherapy orders, instillation execution logs, nursing notes, catheterization records, laboratory reports, outpatient records, and follow-up documentation. Duplicate records were identified by patient identifier, instillation date, chemotherapy order, and execution record. When duplicate entries referred to the same instillation, the most complete record was retained, and conflicting information was resolved by checking the original chemotherapy order, nursing execution note, and follow-up record. The data-extraction workflow, including source databases, screening, duplicate handling, outcome adjudication, and missing-data handling, is summarized in Figure 1.

Patients missing group assignment, index instillation date, or primary TRC outcome were excluded. These variables were treated as essential because they were required to define pathway exposure, time origin, and the primary outcome. For included patients, covariate missingness was reported by variable. The extent and pattern of missingness were reviewed before modeling to distinguish unavailable documentation from documented absence of a finding. Complete-case analysis was used for the primary model when missingness was low. Low missingness was defined as limited covariate missingness that did not materially reduce the analysis sample or affect model stability. Missing-indicator analysis was used in sensitivity analysis when covariate missingness was clinically relevant. For clinically important covariates with incomplete documentation, a missing-indicator category was created in sensitivity analyses so that patients were not excluded solely because one adjustment variable was unavailable. No primary or secondary outcome values were imputed, and patients without sufficient documentation to classify the primary TRC outcome were excluded rather than assumed to be event-free²⁹.

For instillation-level records, undocumented checklist items were scored as not completed. Missing procedural documentation was not interpreted as absence of difficulty or bleeding unless the record explicitly stated absence. For example, catheterization difficulty, visible bleeding, early voiding, or retention-time completion was coded as absent only when the record clearly documented that the event did not occur. If the record was silent, the item was coded as undocumented for procedural-variable extraction. For fidelity scoring, however, undocumented checklist items were scored as 0 because the fidelity score was intended to measure documented implementation completeness rather than actual but unrecorded nursing behavior. This rule was used to avoid overestimating pathway fidelity in retrospective records. The same missing-data rules were applied to both the standard-care and preventive nursing-pathway groups to reduce differential classification caused by more detailed documentation in the pathway group.

Statistical analysis

Continuous variables were reported as mean ± standard deviation or median with interquartile range. Categorical variables were reported as counts and percentages. Between-group comparisons used t-tests or Wilcoxon rank-sum tests for continuous variables and chi-square or Fisher's exact tests for categorical variables.

Patient-level TRC incidence was defined as the proportion of patients with at least one TRC event within the 7-day post-instillation windows during the observed treatment course. Instillation-level TRC incidence was defined as TRC episodes per 100 documented instillations. Because repeated instillations were clustered within patients, instillation-level inferential analyses used generalized estimating equations with robust standard errors. Patient-level analyses were primary.

Unadjusted comparisons between the standard-care and preventive nursing-pathway groups were performed for any TRC, grade ≥2 TRC, recurrence, symptom-relief medication use, cystitis-relevant antibiotic use, unplanned visits, and symptom-driven delay or interruption. These comparisons were interpreted descriptively.

The primary multivariable model used patient-level TRC as the dependent variable. Because 55 patient-level TRC events were observed, the model was kept parsimonious. The core adjusted model included care pathway, TURBT-to-first-instillation interval, catheterization attempts, baseline LUTS burden, baseline infection-risk status, and prior urinary tract infection history. Additional variables were examined in sensitivity analyses. Model stability was assessed using events-per-variable considerations, coefficient stability, confidence interval width, and alternative model specifications. Firth's penalized logistic regression was used as a small-sample robustness check.

Propensity score analysis was used as a secondary confounding-control strategy for the non-randomized pathway comparison. The propensity model estimated the probability of preventive nursing-pathway management using age, sex, NMIBC risk category, intravesical agent, TURBT-to-first-instillation interval, baseline LUTS burden, baseline infection-risk status, prior urinary tract infection history, diabetes, and baseline or procedural catheterization indicators available before outcome occurrence. The primary propensity approach used stabilized inverse probability of treatment weighting. Covariate balance was assessed using standardized mean differences, with values below 0.10 considered acceptable. Propensity-score covariate adjustment and propensity-score matching were used as sensitivity analyses.

Time-to-event analyses were performed for time to first TRC and interruption-free completion. The time origin was the index instillation date. For time to first TRC, the event was the first documented TRC episode. Patients without TRC were censored at last documented follow-up, planned course completion, non-TRC discontinuation, loss to follow-up, or January 31, 2025, whichever came first. For interruption-free completion, the event was TRC-related delay, interruption, or discontinuation. Non-TRC discontinuation was treated as censoring and examined as a competing event in sensitivity analysis. Kaplan-Meier curves and log-rank tests were used descriptively. Cox models with robust variance estimation were used when adjusted hazard ratios were reported. The proportional hazards assumption was checked using graphical inspection and Schoenfeld residuals.

Subgroup analyses examined high baseline infection risk, early instillation within 7 days after TURBT, moderate-to-severe baseline LUTS burden, and difficult catheterization or multiple attempts. These analyses were prespecified but exploratory. Interaction terms were tested, and subgroup-specific estimates were not interpreted as differential pathway associations unless the interaction test supported effect modification.

Sensitivity analyses examined alternative TRC windows, grade ≥2 TRC, exclusion of infection-related events, non-infectious and infection-related phenotypes, alternative propensity-score strategies, Firth penalized regression, and alternative missing-data handling. The primary endpoint was patient-level any TRC within the prespecified observation framework. Secondary, subgroup, instillation-level, and time-to-event analyses were interpreted as exploratory and hypothesis-generating. All tests were two-sided, and P < 0.05 was used as the nominal significance threshold. Analyses were conducted using R version 4.3.2.

Results

Cohort characteristics and documented uptake of the preventive nursing pathway

A total of 120 patients receiving intravesical chemotherapy were included, with 60 patients in the standard-care group and 60 patients in the preventive nursing-pathway group. The cohort was predominantly male (81.7%) and middle-aged to older. Baseline demographic and clinical characteristics were generally comparable between groups (Table 1). The median interval from TURBT to first...

Discussion

This retrospective single-center cohort study found that treatment-related cystitis (TRC) was frequently documented among patients receiving intravesical chemotherapy. Nearly half of the cohort experienced at least one TRC episode, and more than one-fifth developed grade ≥2 TRC. These findings are consistent with previous reports showing that intravesical treatment can be accompanied by clinically relevant urinary toxicity and treatment-related discomfort³⁰. In the present cohort, TRC was no...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Disposable sterile urinary catheter, 12–16 Fr	Well Lead Medical Co., Ltd.	Manufacturer product information page: Wellead urology catheter / all silicone Foley catheter	Used for intravesical catheterization and bladder instillation procedures
Electronic medical record system	Second Affiliated Hospital of Wenzhou Medical University	Not applicable; institutional electronic medical record system	Source of demographic data, diagnosis, comorbidities, treatment records, clinical notes, and follow-up documentation
Epirubicin hydrochloride for injection	Pfizer or hospital-procured equivalent manufacturer	Manufacturer prescribing information or institutional pharmacy procurement record	Intravesical chemotherapy agent recorded in chemotherapy orders
Follow-up documentation records	Second Affiliated Hospital of Wenzhou Medical University	Not applicable; institutional outpatient, telephone, and nursing follow-up records	Source of 24–72 h follow-up completion, post-instillation symptoms, care escalation, and treatment-interruption documentation
Gemcitabine for injection	Jiangsu Hansoh Pharmaceutical Group Co., Ltd.	Manufacturer product information page: Gemcitabine for Injection	Intravesical chemotherapy agent recorded in chemotherapy orders
Intravesical chemotherapy execution logs	Second Affiliated Hospital of Wenzhou Medical University	Not applicable; institutional treatment execution record	Source of instillation date, execution status, retention time, catheterization-related documentation, and cumulative instillation count
Microsoft Excel	Microsoft Corporation	https://www.microsoft.com/microsoft-365/excel	Data checking, preliminary data cleaning, table preparation, and verification of extracted variables
Mitomycin C for injection	Hospital-procured equivalent manufacturer	Institutional pharmacy procurement record	Intravesical chemotherapy agent recorded in chemotherapy orders
Pirarubicin hydrochloride for injection	Shenzhen Main Luck Pharmaceuticals Inc. or hospital-procured equivalent manufacturer	Institutional pharmacy procurement record	Intravesical chemotherapy agent recorded in chemotherapy orders
R package: geepack	CRAN	https://cran.r-project.org/package=geepack	Generalized estimating equation analysis for instillation-level clustered outcomes
R package: ggplot2	CRAN	https://cran.r-project.org/package=ggplot2	Generation of distribution plots and publication-ready statistical figures
R package: ipw	CRAN	https://cran.r-project.org/package=ipw	Stabilized inverse probability of treatment weighting for propensity-adjusted analyses
R package: logistf	CRAN	https://cran.r-project.org/package=logistf	Firth penalized logistic regression for small-sample robustness analysis
R package: MatchIt	CRAN	https://cran.r-project.org/package=MatchIt	Propensity-score matching sensitivity analysis
R package: survival	CRAN	https://cran.r-project.org/package=survival	Kaplan–Meier analysis, Cox proportional hazards modeling, and time-to-event analysis
R package: survminer	CRAN	https://cran.r-project.org/package=survminer	Visualization of Kaplan–Meier curves and number-at-risk tables
R package: tableone	CRAN	https://cran.r-project.org/package=tableone	Baseline characteristic tables and standardized mean difference calculation
R software, version 4.3.2	R Foundation for Statistical Computing	https://www.r-project.org/	Statistical analysis, regression modeling, time-to-event analysis, and figure generation
Urinalysis reports	Clinical Laboratory, Second Affiliated Hospital of Wenzhou Medical University	Not applicable; institutional laboratory information system	Source of baseline urine WBC, RBC, nitrite, leukocyte esterase, and infection-risk indicators
Urine culture reports	Clinical Laboratory, Second Affiliated Hospital of Wenzhou Medical University	Not applicable; institutional laboratory information system	Source of urine culture results used for infection-related phenotype classification