difficulties in collecting specimens which are suitable for
analysis from incontinent patients. The prevalence of incontinence
in the acute phase of stroke is high and the use of
incontinence pads as a first line intervention for incontinence
limits the use of the refractometer USG and Ucol measurements
(Mentes et al. 2006). Furthermore, the use of absorbent
gelling materials in disposable pads can influence the
optical/chemical properties of the urine and thus the accuracy
of the USG measurements (Kirkpatrick et al. 1997).
We found that there was also no clear increase in USG,
regardless of measurement method used, which preceded any
increase in U:C ratio as the best indicator of dehydration
routinely available. There was also poor agreement between
different measures of USG and Ucol. The test strips consistently
under estimated USG and the colour chart over
estimated dehydration compared with the refractometer.
Therefore, our study findings do not support the accuracy
of the urine test strip USG and Ucol measurements as useful
markers of hydration status in acute stroke patients.
Previous study findings have found inter-rater reliability of
USG urine test strips and Ucol range from poor to average
and the urine test strips have been found not to be reliable
when compared with refractometry (Winkens et al. 1992, de
Buys Roessingh et al. 2001, Stuempfle & Drury 2003). Our
study findings also support those of Fletcher et al. (1999)
who suggested that when high levels of dehydration are
indicated by blood indices this may not be reflected in USG
and Ucol as accurately as during mild dehydration. We also
found that measurements of fluid intake and output were
incomplete probably because, they are time consuming for
nursing staff to complete (Watkins et al. 1997).
An interesting finding was that there was good agreement
between nurses’ opinion of the patient’s hydration status
(with all dehydrated patients having a dry mouth) and the
refractometer USG readings and blood indicators of hydration
status. A recent study by Vivanti et al. (2010) found that
oral symptoms such as tongue furrows and dry mucous
membranes were most indicative of dehydration in older
adults. However, other clinical features of dehydration such
as poor skin turgor and low venous pressure are considered
unreliable and insensitive to mild to moderate dehydration
and are often difficult to elicit in elderly patients (Vivanti
et al. 2008). Acute changes in body weight after imposing
fluid restrictions or exercise has also been described as a good
indicator of hydration status, but may be affected by bowel
movements as well as food and fluid and would be difficult
and unethical to measure in sick immobile stroke patients
(Vivanti et al. 2010). Other methods to determine hydration
status including skin-fold thickness, axillary sweating and
orthostatic tolerance tests all pose problems for immobile
patients and have poor inter- and intra-reliability (Vivanti
et al. 2008). Therefore, we suggest that further studies should
aim to determine whether combining a practical subjective
assessment such as the nurses’ opinion with an objective
measure such as the refractometer device provide a reliable
measure of hydration status.
Limitations of this study include: the small numbers of
patients in relation to urine samples so we are unable to
assume that the readings taken from the same person on
different days can be treated as independent in terms of
agreement between test methods. However, repeat analyses
of the first 20 urine tests on admission suggest similar
patterns of agreement between the refractometer, test strips
and urine colour chart. Moreover, the lack of daily measurements
of U:C ratio or osmolality taken at a similar time to the
USG readings and insufficient numbers of patients made it