Fluid Balance in Elderly Patients following Acute Stroke P. A. O'NEILL, I. DAVIES, K. J. FULLERTON, D. BENNETT
Introduction There has been increased interest in the last few years in water balance in the acute phase of patients suffering a first stroke. This was .stimulated by the studies of Gordon et al. [1] and Barer [2], where a high incidence of dysphagia was reported following a hemisphere stroke. Furthermore, Gordon et al. [1] considered that dysphagia was associated with an increased risk of developing dehydration, though the association in their study did not reach statistical significance. Barer [2] failed to show any clear relationship. Most clinicians are aware of the syndrome of inappropriate secretion of antidiuretic hormone (arginine vasopressin: AVP) and intracranial disease. There have been only limited studies, however, on changes in plasma sodium or osmolality following a stroke. These have mainly been confined to patients with subarachnoid haemorrhage and have not measured plasma AVP levels [3, 4]. There has been only one study of changes in AVP concentration following cerebral infarction or haemorrhage [5]. However, the stage of the stroke when the measurements were made, serial changes of the
hormone, or the relationship to fluid intake were not reported [5]. In addition, patients with subarachnoid haemorrhage were included [5]. We have conducted several studies of water homoeostasis in elderly subjects and have shown that a high plasma osmolality is a predictor of a reduced survival [6, 7]. In this study we have examined the effects of acute stroke and the presence of dysphagia on the variables associated with the control of water balance. The aim was to observe the response to the usual clinical management of a patient with a stroke.
Methods Patients: Consecutive patients with a diagnosis of acute stroke, admitted under the care of the geriatricians at the University Hospital of South Manchester, were eligible for inclusion. The criteria were that the onset of symptoms could be accurately determined, and that they had started less than 24 hours prior to the first blood sample being taken. Stroke was defined as acute onset of weakness including the arm, leg, or both, which subsequently persisted for more than 24 h. Informed written consent was obtained from either the patients or their next of kin Age and Ageing 1992;21:280-285
Downloaded from http://ageing.oxfordjournals.org/ at Nipissing University on October 16, 2014
Summary We studied the relationship between plasma osmolality, arginine vasopressin (AVP), and fluid input in patients during the acute phase of a first stroke. Fifteen consecutive patients were studied (median age 79) and their blood sampled on days 0, 1, 2, 3, 7 and 14. Plasma osmolality was related to fluid input over days 0-3 (p = 0.0013) and AVP over 14 days (p < 0.001). Patients with a poor outcome had higher AVP concentrations (p = 0.02). Those on intravenous fluids received a higher volume (p 10): II; Poor: remaining dependent in activities of daily living (Barthel score of ^ 1 0 ) , or dead. The outcome measures are directly derived from those used in a previous large study on prognostic factors in stroke [8].
281
P. A. O'NEILL ET AL.
28a 315 n Alive
I ] £
285 ^
o •• O
275 -
1 k- - \
Individual viluti 15 patltnti
30 -
/ml)
• 305 -
L
II
1 .
s 20 10 -
4
Dead
0 -
265 2
3
7
14
Time (days)
Figure 1. Mean plasma osmolalities and 95% confidence intervals over days 0-14 in those patients alive or dead at 6 months.
time (p = 0.20) (Figure 1); on day 0 the mean was 294.3 mOsm/kg (95% CI 283.2 to 303.4) and on day 14 it was 290.1 mOsm/kg (95% CI 277.0 to 298.3). Inter- and intra-individual variation was wide, the maximum plasma osmolality observed was 336 mOsm/kg and the minimum was 246 mOsm/kg. Serum sodium and packed cell volume: All patients had serum sodium and packed cell volume recorded at day 0. The mean sodium was 135.3 (3.8)mmol/l with a range of 129-144 mmol/1. The mean packed cell volume was 0.411 (0.049) with a range of 0.281-0.490. Fluid input: Data for fluid input were complete in all but one patient for days 0-7. No analysis was possible of urine output as five of the subjects were incontinent and not catheterized in the initial stages of their stroke. The mean fluid input on day 0 was 731 ml (95% CI 425-1031 ml) and on day 3 it was 1071 ml (95% CI 719-1363 ml). Arginine vasopressin: There was no significant change in plasma AVP concentrations with time (p = 0.14), but there was a wide variation in the values both between and within subjects (Figure 2). The mean plasma concentration on day 0 was 7.65 pg/ml (95% CI 5.41-10.70) and on day 14 it was 5.07 pg/ml (95% CI 3.27-7.62). Two distinct AVP responses could be discerned. When pOsm and AVP were considered for each subject individually, there were five
0
1 2
3
14 Days post stroke
Figure 2. Individual AVP concentrations for the patients following stroke. subjects with negative correlations (range — 0.13 to -0.78) and ten subjects with a positive correlation (range 0.15-0.84). However, only two of these reached statistical significance owing, in part, to the small number of sample points for each individual. In three subjects the plasma osmolality rose to above 300 mOsm/kg during days 0-3 (day 0: 298, 298, 309; day 3: 327, 309, 336 mOsm/kg). In these patients the AVP increased in parallel from 8.9, 15.7, 9.2 to a peak of 40.4, 22.8, 32.1 pg/ml. This indicated a normal response in the face of a negative fluid balance (minus 1770, 1450, 1100 ml recorded over days 0-3, not allowing for insensible losses). In contrast there were five patients in whom the plasma osmolality was less than 280 mOsm/ kg at some stage in the 14 days following their stroke and where AVP was measured (lowest value 278, 275, 278, 273, 246 mOsm/kg). In these subjects there was an inappropriate secretion of AVP with persisting concentrations of the hormone even at the lowest plasma osmolality (3.5, 5.1, 6.6, 12.8, 6.1 pg/ml) and the maximum AVP values recorded during this period were 4.4, 17.7, 12.5, 18.0, 6.9 pg/ml. Relationship of variables to outcome: There were no significant differences in plasma osmolality or fluid input between survivors and nonsurvivors or between good and poor outcome. However, in those patients with a good outcome significantly lower hormone concentrations were found (p = 0.02, see Figure 3).
Downloaded from http://ageing.oxfordjournals.org/ at Nipissing University on October 16, 2014
0
FLUID BALANCE IN STROKE 15 -1 Poor Outcom* Good Outcome
a. >
0.1). For the oral fluid group (n = 5) input was 688 ml (399-992) at day 0 and 757 ml (4651101) at day 7 (p > 0.1). There was a significant difference in total input between the two groups
Discussion 285 -I
Days following Stroke
Figure 4. Mean plasma osmolalities and 95% confidence intervals for those subjects allowed oral fluid following their stroke and those requiring intravenous fluid infusion for 7 days (p = 0.04).
Relationship of variables to plasma osmolality: There was a significant relationship between serum sodium and plasma osmolality, and packed cell volume and plasma osmolality, when averaged between days 0 and 1 (data not presented). Over days 0-3 the fluid input was the major predictor of plasma osmolality (p = 0.0013). There was also a relationship between plasma osmolality and plasma AVP over the 14 days (r = 0.38, p