A prospective study of urinary retention and risk of death after proximal femoral fracture

Age and Ageing,  March, 1996  by Nigel K.G. Smith,  Mohammed K. Albazzaz

• E-mail
• Print
• Link

Introduction

Proximal femoral fracture is regarded as a complication of osteoporosis. The incidence of hip fracture rises exponentially with age and risk of fracture is related to measures of bone mass. Cognitive impairment, age and pre-fracture disability are well recognized risk factors for fatality in the first 6 months after a proximal femoral fracture [1, 2]. A prospective study of 518 patients found that life expectancy was related to a four-point scale of social dependence before the fracture [1]. A study involving 279 proximal femoral fractures identified 'residence in an institution' as a factor predicting fatality, in addition to impaired mental function. Pre-existing medical illnesses, such as diabetes, have also been found to influence survival following fracture [3].

Post-operative urinary retention is the inability to void with a full bladder during the post-operative period. It may be identified by symptoms ('not passing urine'), urethral catheterization or ultrasound. Studies of catheterization reveal that urinary retention occurs in over half of women following elective total hip replacement [3, 4]. Reports of randomized, controlled trials in patients undergoing total hip replacements recommend pert-operative catheterization to prevent retention [4, 5] but this recommendation has not been implemented in the operative fixation of hip fracture in the United Kingdom.

It is now possible to measure post-voiding residual volume by means of portable ultrasound, so that urinary retention can be identified using an easily repeatable, non-invasive and well tolerated method [6].

We prospectively studied 309 consecutive women aged 65 and over presenting with proximal femoral fracture to one trauma unit to estimate the incidence of pre- and post-operative urinary retention and identify their correlates.

Patients arid Methods

A daily register was kept of all women aged 65 and over admitted to the University Hospital, Nottingham with proximal femoral fractures between 13 December 1990 and 10 July 1991. The study was approved by the University Hospital Ethical Committee. The object of the study was explained to patients and they were invited to undergo an ultrasound examination, which provided additional information of value in their pert-operative care.

An abbreviated mental test score [7] was administered by the two observers and a medical history was taken, including past medical diagnoses, such as stroke, urinary symptoms, present medications, circumstances of fall, duration of lie, site of fracture and living accommodation. Pre-morbid mobility was assessed using modified functional ambulation categories [8] (1 = fully ambulant, including stairs, 2 = usually independent, not stairs, 3 = walking with supervision, 4 = walks with aids or careful supervision, 5 = bed- or chair-fast).

We recorded the operative procedures undertaken, the type of anaesthesia and use of analgesia. We made daily visits to the wards to record urinary symptoms and operative complications, until the patients were discharged from hospital.

A Bladderscan BVI 2000 portable ultrasound [Diagnostic Ultrasound, PO Box 2655, Kirkland, WA 98083] was used to take two measurements of post-voiding residual volume. Each patient was given the opportunity to void. Non-weight-bearing patients were transferred to a bedpan by two nurses. Two readings of post-voiding residual volume (PVRV) were made with the patient lying supine. The highest reading was used for analysis as recommended [6]. Two observers (N.K.G.S. and M.K.A.) recorded all the bladder volumes. The equipment and techniques of both observers were validated alongside catheterized volume in over 100 examinations of female patients aged 65-95. Inter- and intraobserver variation had been assessed and found to be negligible.

We took three readings of PVRV: pre-operative--within 24 hours of admission to hospital; post-operative--within 24 hours of operation; and recovery-5-7 days post-operatively. If the patient was catheterized at any time in hospital, we recorded the volume of urine drained in the first hour after catheterization. We defined urinary retention as a PVRV > 300ml.

After discharge, patients were traced using the hospital information system or by contacting their general practitioners.

Statistical analysis: We divided the data into four age groups and used a X[.sup. 2] test to search for age trends. Patients who had measurements of PVRV were compared with the `absent data' group using variance analysis including a non-parametric test, the Kruskal--Wallis H. We analysed the relationship between urinary retention and other variables using the Mantel--Haenszel procedure to estimate the corrected odds ratios for each factor across the four age groups. Kaplan--Meier survival curves were calculated and a step-wise Cox's regression was used to identify factors which influenced survival. We used the statistical package, SPSS-X3.

Results

Patient characteristics: Three hundred and nine women aged 65-103 (median age 83 years) were admitted over the 7-month period. Two-thirds (211) were admitted from home and three-quarters (239) fell indoors. For 27 women details of their falls were unobtainable and 20 patients walked after their fall, unaware of their fracture until between 1 and 4 days later. Of the remaining 262 patients 226 (80%) lay for less than 1 hour after falling.