We recruited 33 consecutive patients (14 males and 19 females; mean age, 68.3±12.8 years; 26 with cerebral infarction and seven with cerebral hemorrhage; Table 1) among patients admitted to the Department of Rehabilitation Medicine of Asan Medical Center, according to the following criteria: 1) first stroke; 2) first bladder scan to measure post-void residual urine volume (PVR) within 30 days of onset; 3) PVR >100 mL, despite use of the maximum alpha-blocker dose (0.4 mg tamsulosin) and cholinergic agent (100 mg bethanechol) for 7 days; 4) no history of urinary symptoms before stroke; 5) no abnormal findings on a prostate exam; and 6) no urinary tract infection before admission and no serious medical complications during hospitalization, such as pneumonia or cardiac problems. This study protocol was approved by the Institutional Review Board of the participating university hospital.

Indwelling catheters were removed from patients with urinary retention. PVR was measured in each patient within 24 hours of their transfer or admission. A bladder scan was used to measure PVR, which was checked more than three times each day. The accuracy of measuring PVR using a portable bladder scan has been demonstrated [10,11]. PVR was measured within 30 minutes after urination using the battery-powered, portable BladderScan BVI 3000 (Verathon Inc., Bothell, WA, USA). The scanner was placed on the suprapubic area and held stationary during scanning. Bladder diameter and urine volume were calculated from the scan data. After scanning, a B-mode image and an indication of bladder urine volume (in mL) appeared on the screen. All clinical procedures were performed by well-trained nurses. Urinary catheterization was performed if PVR was ≥400 mL. If a patient was unable to void, we measured urine volume every 2-4 hours until it was ≥400 mL, at which time catheterization was performed. The alpha-blocker was administered after 2 days of measuring PVR, and the alpha-blocker dose was increased if PVR was >200 mL. A cholinergic drug was administered when PVR remained >200 mL, despite administration of the maximum alpha-blocker dose for 2 days.

Voiding diaries were documented, including frequency of urination and voided volume. Residual urine was measured by BladderScan three times per day until discharge. Urinary retention was defined as PVR ≥100 mL. If PVR was always <100 mL, the drug dose was reduced in the order of the cholinergic agent and then the alpha-blocker. Drug-resistant urinary retention was defined as PVR ≥100 mL, despite maximum doses of the alpha-blocker and cholinergic agent for 7 days.

The outcomes of drug-resistant urinary retention at the early stage of stroke were divided into three groups according to PVR at discharge, i.e., <100 mL, ≥100 mL but <400 mL, and ≥400 mL or incapable of voiding.

An urodynamic study (UDS) was performed once during admission in patients who still had urinary retention after 7 days of drug administration to analyze the voiding dysfunction pattern. Drugs for managing urinary retention were discontinued the day before the UDS. The UDS was performed as described in the International Continence Society guidelines. Filling cystometry was performed with the patient in the supine position on an urodynamic table. The bladder was filled with normal saline at a medium (10-100 mL/min) filling rate. Bladder filling was terminated when maximum bladder capacity was reached, and a rectal catheter was inserted to measure intra-abdominal pressure. Vesical, abdominal, and detrusor pressures were monitored simultaneously during the filling and voiding phases. The drugs were restarted after the UDS. The mean duration from onset to the UDS was 24.6±9.4 days.

The UDS findings were defined as areflexic, if there was no contraction of the detrusor, and hyper-reflexic, if there were involuntary detrusor contractions during the filling phase.

Clinical data included the presence of diabetes mellitus, urinary tract infection after transfer or admission, a communication disorder, functional status, walking ability, cognitive function, and kidney function. A patient's communication problem was assessed by a speech therapist, and the presence of aphasia or a cognitive communication disorder was determined. Patient functional status was assessed using the Korean version of the Modified Barthel Index (K-MBI), as the validity and reliability of the K-MBI has been demonstrated [12]. Walking ability was assessed using the Functional Ambulation Category (FAC), which is based on the level of assistance required during a 15-m walk. The six categories are: 0, non-ambulatory; 1, need for continuous support from one person; 2, need for intermittent support from one person; 3, requires verbal supervision only; 4, help required on stairs and uneven surfaces; and 5, able to walk independently anywhere. The reliability and validity of the FAC have been well established [13]. Cognitive function was assessed using the Korean version of Mini-Mental State Examination (K-MMSE). The reliability and validity of the K-MMSE have been established [14]. Kidney function was assessed using estimated glomerular filtration rate (eGFR) from the Modification of Diet in Renal Disease formula [15]. If eGFR was <60 mL/min/1.73 m², the patient was diagnosed with impaired kidney function.

Data were analyzed using SPSS ver. 18.0 (SPSS Inc., Chicago, IL, USA). We divided the patients into two groups to evaluate differences in drug-resistant urinary retention according to brain lesion location, i.e., patients who had a supratentorial lesion (supratentorial group) and those with an infratentorial lesion (infratentorial group). Demographic data (i.e., sex, age, type of stroke, and lesion site) were compared between the patient groups using the Kruskal-Wallis test. The relationships between continuous variables (i.e., days from onset to transfer, total hospital days, K-MBI, FAC, and K-MMSE scores) and outcome were analyzed using Spearman correlation analysis, and the relationships between the categorical variables (i.e., diabetes mellitus, communication disorder, upper urinary tract infection, and kidney function) and outcome were analyzed using a linear-by-linear association analysis. Lesion type was analyzed after excluding patients with mixed lesions; thus, leaving only 30 patients.