| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *University College Hospital, London, and Centre for Anaesthesia, University College London; Centre for Behavioural and Social Sciences in Medicine, University College London; Department of Surgery, University College London; Centre for Behavioural and Social Sciences in Medicine, University College London, and Department of Psychology, Thames Valley University, London; ||Institute of Child Health, University College London; and ¶Centre for Behavioural and Social Sciences in Medicine, University College London, UK.
Address correspondence and reprint requests to Gareth L. Ackland, PhD, FRCA, Centre for Anaesthesia, University College London, University College London Hospital, 235 Euston Road, London NW1 2BU, UK. Address e-mail to g.ackland{at}ucl.ac.uk.
Abstract
BACKGROUND: Postoperative cognitive dysfunction occurs in a proportion of patients after noncardiac surgery. Older patients are particularly vulnerable. We hypothesized that dehydration, a common perioperative problem in the elderly, may provoke cognitive dysfunction. We used a clinical scenario free of surgical/anesthetic intervention to determine whether dehydration caused by bowel preparation results in cognitive changes.
METHODS: Thirty-eight patients of an age associated with a significant incidence of postoperative cognitive dysfunction were recruited in a prospective observational study. A further control group of 14 patients undergoing sigmoidoscopy, who did not receive any bowel preparation, were matched for age, education, and gender.
RESULTS: Loss of total body weight (1.5 kg [95% CI: 0.9–2.2]; P < 0.001) occurred in patients undergoing bowel preparation (2.0 [95% CI: 1.3–2.6] percent total body weight), whereas sigmoidoscopy patients’ weight did not change (0.17 kg [95% CI: –0.2–0.6 kg]; P = 0.26). Total body water, derived from foot bioimpedance, indicated dehydration in the bowel preparation group only (mean impedance change 36 [Omega] [95% CI; 25–46], P < 0.001) with a calculated decrease of 2.6% in total body water (95% CI: 1.1–4.8; P < 0.001). Hematocrit increased after bowel preparation only (prebowel prep 0.41 [0.40–0.43] versus postbowel prep 0.43 [0.42–0.45]; P = 0.003). Despite this degree of dehydration, all cognitive tests were within 1 sd of the population mean of normal values. Repeated measures analysis of variance did not reveal significant changes for within group comparisons over time for motor speed (P = 0.51), executive function (P = 0.57), Trail Making Tests and recall (P = 0.88), other than a 3 s slowing in learning ability (Rey Auditory Verbal Learning Test; P = 0.04). Hydration status did not affect learning (P = 0.42), recall (P = 0.30) motor speed (P = 0.36), or executive function tests (P = 0.26).
CONCLUSION: Dehydration alone does not result in cognitive dysfunction.
Cognitive impairment is a significant problem in elderly patients during the postoperative period.1,2 The incidence of postoperative cognitive dysfunction (POCD), characterized by impairment of memory and concentration, can be as frequent as 44%–61% in elderly patients immediately after surgery,2,3 declining thereafter.1,4 The occurrence of POCD in the elderly may result in increased morbidity, delayed functional recovery, and prolonged hospitalization. Cognitive dysfunction may be triggered by several perioperative factors, including dehydration. Dehydration/hypovolemia is a potent physiological stimulus for the release of several neurohormones,5 many of which have profound effects on cognition.6–8 Other deleterious changes triggered by preoperative dehydration include intraoperative hypotension, hypoglycemia, and electrolyte imbalance. Dehydration, therefore, may play a role in POCD. A significant proportion of aged human subjects demonstrate impaired thirst, particularly after a period of relative fluid deprivation,9,10 which may be common in the preoperative period.11 In younger adults, rapid heat- and exercise-induced dehydration impair specific cognitive motor functions such as short-term memory, working memory, perceptive discrimination, and visual motor function.12–14 However, cognitive motor function is preserved after slowly progressive, more moderate dehydration (2.6% loss of body weight) over 24 h in healthy young subjects.15 The effects of dehydration in elderly subjects are unclear. We hypothesized that cognitive dysfunction is triggered by dehydration in patients at an age associated with vulnerability to POCD. By using the clinically relevant model of bowel preparation, the confounding factors of surgery and anesthesia were absent and, therefore, the contribution of dehydration alone could be assessed.
METHODS
After obtaining approval from the IRB at University College Hospital, London, and written informed consent, 38 patients scheduled for elective colonoscopy and 14 patients scheduled for elective sigmoidoscopy were recruited for this prospective, observational study. Eligible participants were approached if they met the following inclusion criteria1: aged 50 yr or more,2 no history or clinically evident cerebrovascular disease as reflected by new, transient, or fixed neurological deficits,3 no major visual or hearing impairments, or other sensory or motor impairments that prohibited them from completing the scheduled assessments,4 absence of acute or chronic psychosis, evident severe depression, severe learning disabilities and/or dementia,5 currently stable, defined as not being acutely ill or hospitalized at the time of the assessments, and6 able to undertake simple exercises in written and spoken English.
Sociodemographic characteristics including age, gender, ethnicity, education, and comorbidity were collected by means of a questionnaire. Under standardized conditions, all subjects were weighed in operating room garments and bioimpedance readings were recorded to estimate total body water. Single frequency 50 kHz leg–leg bioelectrical impedance analysis (TBF-300GS; Tanita, UK) was undertaken primarily for its ease of use, for the comparable results with conventional arm–leg analysis, and, most importantly, for its close correlation with dual radiograph absorptiometry and radio-labeled water which are the "gold standard" measurements of body water.16 Total body water was derived from the in-built analyzer algorithm, using the absolute value of foot bioimpedance, weight, gender, and height2/resistance. Subjects did not see any of the results. Only patients scheduled for colonoscopy underwent standardized Citramag bowel preparation. Weight, foot bioimpedance and cognitive function tests were repeated 72 h after the initial recording, immediately before the procedure. For both sigmoidoscopy and colonoscopy groups, venous blood was drawn via an 18 G needle for measurement of hematocrit before each weight and bioimpedance recording. Hemoglobin, hematocrit, and electrolyte levels were measured. Blood urea nitrogen was also collected in addition to plasma electrolytes for the colonoscopy patients.
Neuropsychological assessment was conducted by one experienced psychologist blinded to weight and bioimpedance data in a specially designated room. The neuropsychological assessment was designed to assess three cognitive domains: attention and executive functions (Trail Making Tests [TMT] A and B17) and memory/learning (Rey Auditory Verbal Learning Test [RAVLT]18). The TMT is a two-part measure of attention, visual scanning, motor speed, and planning ability. Form A (TMT-A) assesses motor speed and attention, requiring participants to connect 25 randomly arranged numbers in the proper order. Form B (TMT-B) is a parallel task to TMT-A but more complex, providing information on visual search, scanning, speed of processing, mental flexibility and executive functions. TMT-B requires participants to connect a series of numbers and letters in sequence (i.e., 1-A-2-B-3-C ... ... 13-L) as quickly as possible. Both forms of the TMT are timed (number of seconds) to completion with lower scores indicating better cognitive function.
RAVLT18 is a widely used auditory verbal memory task that assesses immediate memory as well as retrieval from verbal short-term memory storage. It consists of five presentations with a recall of a list of 15 words that are read to the participants by the examiner, one presentation of a second 15-word list and a sixth recall trial of the original word list. The score used in this study was the total verbal recall from trials 1–5 (RAVLT-T) as it enabled comparison to be made with normative samples. Alternate forms of the tests were used to minimize practice effects. These tests were selected because they have been widely used to assess performance after surgery and are sufficiently sensitive to detect minor brain injury,19 Furthermore, these tests assess cognitive functions that are affected by surgery (memory and attention and concentration) and benefit from the availability of normative population data.19 The cognitive tests and weight/body water measurements were repeated 3 ± 1 days later, on the day of the procedure.
To judge the patients’ performance on neuropsychological tests relative to normative performance,20 individuals’ performances on each of the neuropsychological tests were compared with a normative sample. An individual's neuropsychological performance was considered impaired on a particular test if it was >1 sd below the mean of the norms.20 This comparison was performed for all the neuropsychological tests. Comparisons between the absolute neuropsychological scores on each assessment were also made. Quality of life measures were also assessed using the SF-8 health survey questionnaire.21 Scores on the SF-8 were reversed (higher scores indicate better self-reported health-related quality of life). Since factors such as anxiety can affect patients ability to perform the neuropsychological tests described, anxiety was assessed using the short six-item version of the Spielberger State-Trait Anxiety Inventory.22 This questionnaire is scored on a four-point scale ranging from "not at all" (= 1) to "very much" (= 4). Higher scores indicate greater anxiety, the maximum being 24. In addition, patient self-report of cognitive functioning was assessed with the Subjective Cognition Scale23 which uses a three point scale, where higher scores indicate better perceived cognitive performance.
The primary outcome of this study was determined by the presence or absence of impaired neuropsychological performance for each individual subject, as defined by the test result in any test >1 sd below the mean of the norms.20 This comparison was performed for all the neuropsychological tests. Comparisons between the absolute neuropsychological scores on each assessment were also made. Sample sizes of 38 colonoscopy and 14 sigmoidoscopy patients were determined by using a power analysis based on the following assumptions:
= 0.05, β = 0.2.
Student's t-test was performed for normally distributed continuous variables, with the Wilcoxon's signed rank test for difference between median values used for nonparametric data. Repeated measures analysis of variance was used to examine whether there were within (time) group changes for each study subject as well as between (hydration status) group differences in neuropsychological outcomes. Categorical data were analyzed by 
2 or Fisher's exact test, as appropriate. All tests were two-sided with a value of P 
0.05 considered significant (NCSS 2004, UT). Data are presented as mean values (95% confidence intervals) or percentages of baseline values, unless stated otherwise.
RESULTS
Both sigmoidoscopy and colonoscopy groups were comparable with respect to demographic variables (Table 1), including gender, comorbidities, medications, and education. Only the colonoscopy patients sustained weight loss (mean 2.0 [95% CI: 1.3–2.6] % total body weight), with increased mean bioimpedance (36 [Omega] [95% CI: 25–46]; P < 0.0001), absolute calculated mean total body water loss of 1.2l (95% CI: 0.8–1.5; P < 0.0001) and calculated median decrease in % total body water 2.6 (95% CI: 1.1–4.8). No weight (0.17 kg [95% CI: –0.2–0.6 kg]; P = 0.26) or bioimpedance (–1 [Omega] [95% CI: –13–10]; P = 0.8) changes were observed in the sigmoidoscopy patients. Consistent with the weight and bioimpedance data, hematocrit and hemoglobin increased after bowel preparation only, with no changes occurring in the sigmoidoscopy group (Table 2).
|
|
SF-8 physical health survey (Table 3) scores revealed that patients awaiting colonoscopy were more anxious than those undergoing sigmoidoscopy (P < 0.05). All other self-reported quality of life measures were similar between the two groups.
|
Baseline neuropsychological test scores are shown in Table 4. Changes from baseline are depicted in Figure 1. No differences were noted either between treatments (P = 0.36) or within each group (P = 0.52). Neuropsychological performance relative to norms were all within 1 sd of the population mean, thereby indicating that cognitive functioning was not impaired in either sigmoidoscopy or colonoscopy groups. Within-group testing also revealed no changes other than a 3 s decline in the RAVLT, which was independent of hydration status. None of the dehydration markers (%weight change, absolute bioimpedance change and percent calculated total body water change) correlated with changes in cognitive performance (data not shown). There were no gender-related differences in any test of cognitive dysfunction, including the TMT-A test (P = 0.87). A significant negative association between increasing age and poorer baseline neuropsychological performance for TMT was found (r2 = 0.22; P = 0.005). A positive association between education and better baseline neuropsychological scores (apart from recall) was also found (r2 = 0.24; P = 0.003). Neither age nor educational status was associated with the relative improvement seen in TMT performance after bowel preparation.
|
|
DISCUSSION
The pathogenesis of POCD remains unclear.24,25 Even minor surgery in the elderly population can be associated with significant neuropsychological dysfunction when assessed soon after surgery despite advances in surgical and anesthetic techniques. This may contribute to postoperative morbidity in the elderly.26,27 Thus, the hypothesis that dehydration in elderly patients confers increased risk of cognitive impairment seems plausible, particularly in clinical scenarios where significant fluid deficits may occur.11 Furthermore, in certain clinical scenarios, such as preoperative bowel preparation for colonic surgery, dehydration is less likely to be adequately compensated for in the elderly population. We chose the model of bowel preparation/colonoscopy and sigmoidoscopy because it is a relevant clinical setting, with the attendant features that may affect cognitive function, such as anxiety preprocedure, but avoids other confounding factors including anesthesia and surgical trauma. Importantly, each patient acted as their own control. In addition, we compared the dehydrated colonscopy patients with patients who underwent sigmoidoscopy, where no dehydration occurs but a similar procedure is anticipated by the patients. The study was not designed to investigate cognitive effects after each type of procedure, which would have introduced the potential confounding factors of sedation, further stress and discomfort.
The clinical tools used to measure either cognitive dysfunction or POCD in previous work have not been standardized. The psychometric tests used in this study have been shown to be sensitive tests of cognitive dysfunction in several patient populations19 and were undertaken by an experienced psychology researcher. The lack of changes observed after bowel preparation-induced dehydration may be explained by the context in which dehydration occurred. Previous studies have identified cognitive dysfunction at similar levels of dehydration in elderly subjects,28 although several major confounding factors (including prolonged exercise and an uncontrolled external environment) make interpretation of these data difficult. The magnitude and period of time over which dehydration occurs are important factors in determining cognitive dysfunction. The critical level of acute water deficit causing a decrease in cognitive performance occurs at a level of 2% total body weight loss or more, at least in healthy young subjects.13,14 Adverse effects on cognition after rapid dehydration induced by heat or exercise are identical,13 suggesting that total body water deficit may be the key factor. Younger healthy subjects dehydrated by a similar percentage loss in total body weight to the colonoscopy patients in this study fail to exhibit any cognitive dysfunction.15 Younger subjects may exhibit cognitive compensating mechanisms for increased tiredness and reduced alertness during slowly progressive moderate (2%–3%) dehydration, through increasing subjective task-related effort.15 In this regard, the experimental observation that release of vasopressin and other stress hormones over the short term may improve cognitive function is important.29 Several other neurohormones that are released during dehydration exhibit striking effects on cognition. For example, hypercortisolemia impairs active learning, short-term memory and verbal memory.29 Within-group testing revealed that only RAVLT ability was impaired with respect to time, but this was not dependent on hydration status. However, given that the changes observed were still within the population norm, this is not clinically relevant. Other factors such as starvation-induced hypoglycemia, electrolyte imbalance (which does not occur in this model11), older age, and gender could also be important. Although middle-aged patients exhibited an incidence of POCD of 19% 1 week after surgery,25 increasing age is associated with an even higher incidence.24 We cannot discount that patients older than those recruited for this study may be more likely to exhibit cognitive dysfunction after a similar bowel preparation protocol. There is also a possibility that gender may play a role in differential performance on some cognitive tests of reaction time in younger subjects,15 but we did not find any gender-related differences.
In conclusion, dehydration induced by bowel preparation in patients of an age associated with POCD does not lead to cognitive dysfunction. Because this study was free of several confounding perioperative factors, including surgical trauma and anesthesia, these data suggest strongly that dehydration alone is unlikely to be a major factor in provoking cognitive dysfunction. However, it also remains possible that preoperative dehydration in combination with other perioperative factors and/or interventions could increase the risk of POCD.
Footnotes
Accepted for publication October 31, 2007.
Received from the Centre for Anaesthesia, University College London. Supported solely from institutional and/or departmental sources.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
