Sample and setting
An observational cross-sectional cohort study was conducted on the acute geriatric hospitalisation units of the University Hospitals Leuven in Belgium which have a collective capacity of 80 beds. Admission criteria for these wards are defined in the Belgian care programme for geriatric patients: the average age of patients should be ≥ 75 years and they should be in need of a geriatric multidisciplinary approach for one or more of the following reasons: a frailty profile; active polypathology; limited homeostasis; atypical clinical disease presentation; disturbed pharmacokinetics; risk of functional decline; risk of malnutrition; tendency to be inactive or bedridden, with an increased risk for institutionalisation and for dependency in activities of daily living; psychosocial problems .
All patients aged 75 years or older admitted with or without an indwelling urinary catheter between October 26 and December 18, 2015 were consecutively screened for inclusion within 72 h of admission. Patients with urinary catheters for any other reason than UR, patients with ascites, dialysis, uro- or nephrostomy, terminally ill patients, and postoperative patients were not eligible for inclusion. Patients who refused to participate, who were unable to understand and carry out the instructions for the PVR measurement, with a urinary catheter for chronic UR, who were discharged or died during the inclusion process were excluded. The study was approved by the Ethics Committee Research of the University Hospitals Leuven (mp08305). The study nurse (SV) obtained a written informed consent from all patients who agreed to participate in the study, or from a proxy when the patient was incapable to provide written informed consent.
Prior to the start of the study, information sessions were organised for the nurses from the acute geriatric hospitalisation units, in which the study nurse (SV) explained the aim of the study and the data collection procedure. In addition, information was given on the use of the portable ultrasound bladder scan and the execution of a bladder scan.
The nurses were asked to perform 3 consecutive post-void bladder scans as soon as possible after admission and at the latest within 72 h of admission as standard of care during the study period. The highest value of the 3 consecutive PVR measurements was recorded. The study nurse (SV) was present on Monday, Wednesday and Friday morning, performed informed consents and performed bladder scanning in patients where this had not yet been done. Before performing the bladder scan, the patients were asked to urinate. Whenever possible, patients had to walk to the toilet in the bathroom, with or without assistance, meanwhile their mobility was assessed.
Consecutively, the study nurse performed a bedside assessment. A mini-cog examination was performed and urinary symptoms, defaecation habits, and fall history were assessed by patient or caregiver interview using a structured questionnaire (SV). The demographic and remaining clinical data were collected from the patient’s electronic medical records (SV, KH, KF).
The PVR was measured using a portable ultrasound bladder scan (Verathon BladdersScan BVI 3000). Jalbani et al. found a high correlation (r2 = 0.97) between urinary catheterisation (gold standard) and bladder scanning and concluded that the use of the Verathon BladderScan BVI 3000 is as accurate as urinary catheterisation for measuring the PVR . As there are no official definitions or guidelines on what is considered a relevant PVR for UR, in this study we defined UR as a PVR of ≥ 150 ml. In addition, we also included a PVR ≥ 300 ml as secondary outcome. The cut-offs were chosen for the following reasons: the first bladder-filling sensation is felt at a volume of 150 ml. At a volume of 300 ml, the pressure receptors in the bladder wall will be activated, which creates a sense of fullness of the bladder and a need to urinate . Moreover, in clinical practice a PVR ≥ 300 ml is often considered as the threshold for insertion of a urinary catheter.
The following patient characteristics were recorded: age, gender and living situation. Patients living alone at home and together at home were considered as living at home and patients living in an assisted living facility or nursing home were considered as not living at home. Patients living in a convent were considered as not living at home when they were functionally or cognitively impaired.
Various comorbidities were retrospectively retrieved from the patient’s electronic medical records: neurological, gynaecological and urological history, a history of diabetes type 1 or 2, recurrent falls (more than 1 fall episode in the last 6 months), heart failure and chronic kidney disease (CKD). CKD was defined as a glomerular filtration rate (eGFR, calculated using the CKD-EPI formula) of < 60 ml/min/1.73 m2 . An age-adjusted definition (< 45 ml/min/1.73 m2) was also considered .
For each patient, the anticholinergic burden (ACB) score was calculated taking into account medication use in the last 24 h before the PVR measurement [28, 29]. Antidepressants with ACB score ≥ 2, as well as antipsychotics, opioids, and detrusor relaxants were examined separately for their associations with UR. Antiepileptics, histamine H1 receptor antagonists, anti-diarrhoea medications, decongestive drugs, spasmolytics, muscle relaxants, and antiparkinsonian agents were not considered individually. Low prevalence of these drugs precluded reliable statistical analysis.
Urinary incontinence was assessed by means of 3 questions: Do you have/does the patient have (1) an urge to urinate and involuntary loss of urine before reaching the toilet? (2) involuntary loss of urine when sneezing, coughing, lifting or moving something? (3) loss of small amounts or drops of urine without urge to urinate or loss of small drops of urine when changing position? Answering positive on these questions was classified as urge, stress or overflow incontinence, respectively. Dysuria was questioned as having a burning sensation or pain when urinating. Urinary frequency was questioned as urinating more often than usual. Voiding difficulty was questioned as straining to void. Subtotal voiding was questioned as the feeling of incomplete bladder emptying. Hypogastric pain was questioned as pain in the lower abdomen.
Constipation was defined as the absence of stools for 3 or more days before the PVR measurement, or if patients reported having had less than 3 stools in the last week with a need to strain and mostly hard or lumpy stools, or if the patient had faecal impaction . Faecal impaction was defined as a large mass of compacted faeces in the rectum or colon on clinical or radiographic examination, that could not be spontaneously evacuated .
The cognitive status of the patients was evaluated using the Mini-Cog . The Mini-Cog is a brief cognitive test that involves an assessment of an older person's ability to memorise three words, to draw a clock, and to recall the three words. Repeating only 1 or 2 of the 3 words and being unable to draw a clock correctly or not being able to repeat any of the words was considered as cognitive impairment .
Mobility was assessed by walking to the bathroom and going to the toilet. All patients were asked to urinate on the toilet prior to the PVR measurements. Patients were considered mobile if they could walk back and forth to the toilet and if they were able to stand up from the toilet independently. The use of mechanical aids was permitted. Needing assistance from another person or not being able to walk was defined as impaired mobility.
Laboratory evaluation and urinalysis
The patient's renal function was evaluated by serum creatinine level and eGFR (calculated using the CKD-EPI formula).
A urinary dipstick test was performed in all patients. In patients with positive leucocyte esterase or positive nitrites on the dipstick test, further urinalysis was performed. An automated leucocyte count of ≥ 10 leukocytes/mm3 and significant microscopic bacteriuria of > 100.000 CFU/ml for which antibiotics were started by the treating physician was considered as UTI in this study . If the patient was taking antibiotics at the time of the urine sample and cultures remained negative, the presence of UTI was determined based on the leucocyte count.
Reason for referral to the hospital
The reasons for referral to the hospital were divided into the following categories: confusion, falls or immobility, urinary symptoms, cardiopulmonary symptoms, digestive symptoms, fever or non-specific symptoms with biochemical inflammation (leucocytosis and elevated C-reactive protein) on admission, other. In patients with multiple reasons for admission, each reason was scored separately.
Descriptive and comparative statistics were used. Normally distributed continuous variables were reported as means with standard deviations (SD). Not normally distributed continuous variables were reported as medians with interquartile ranges (IQR). Categorical variables were reported as numbers and percentages. Baseline characteristics and clinical variables of patients with a PVR of < 150 ml versus ≥ 150 ml and a PVR of < 300 ml versus ≥ 300 ml were compared using unpaired t-tests for normally distributed continuous variables. Pearson Chi-squared tests or Fisher’s exact tests (if ≥ 1 cell had an expected count of less than 5) were used for dichotomous or nominal variables and Mann–Whitney U tests for not normally distributed continuous variables or ordinal variables. All tests were 2-tailed, assuming a 5% significance level. Multivariable logistic regression models were used to determine independent predictors for a PVR ≥ 150 ml and a PVR ≥ 300 ml. Significant variables in the univariable analysis were considered in a forward stepwise logistic regression procedure. P-values (Likelihood ratios), odds ratios (OR), and 95% confidence intervals (CI) are reported. The software package used was SPSS version 20 (SPSS Inc., Chicago, IL).