- Email: [email protected]
The Effects of Dialysis on Brain Water and EEG in Stable Chronic Uremia
The Effects of Dialysis on Brain Water and EEG in Stable Chronic Uremia Carlo Basile, MD, Jack D.R. Miller, MB, Zoly J. Koles, PhD, Michael Grace, PhD, and Raymond A. Ulan, MD • Cerebral edema in uremic animals and humans, as well as an EEG deterioration in humans, has been reported after dialysis. Both are manifestations of the dialysis disequilibrium syndrome (DDS). This study was designed to analyze the changes induced by dialysis in the EEG pattern (spectral analysis), in the cerebral hydration, and ventricular size (computed tomography [CT] of the brain) In a group of 11 stable uremic patients. They volunteered for a randomized crossover study of 4 months each of standard hemodialysis (HD) and hypertonic hemodiafiltration (H HDF). H HDF is a dialysis technique that is shorter and more efficient than HD. An EEG recording, a CT scan of the brain, and blood biochemistry were performed before and after a HD (four hours, blood flow rate 250 mll min) and a H HDF run (three hours, blood flow rate 400 ml/min). Approximately 6 weeks of stabilization on each treatment were allowed before these studies. No difference was found in the density of seven specific brain structures (base and apical cuts), when comparing pre- v post-HD, pre- v post-H HDF, pre- HD v pre-H HDF, and post-HD v post-H HDF. Furthermore, no difference was evident either in the bicaudate diameter of the lateral ventricles or in the transverse diameter of the third ventricle. In addition, no significant in-between- and withintreatment difference was observed when analyzing the EEG% power (3-7/7-13 Hz) data. In conclusion, this study shows neither a postdialysis change In brain density and ventricular size nor a postdlalysls EEG deterioration in a group of stable uremic patients undergoing both a rapid and a standard dialysis treatment. Thus, DDS must be considered a peculiarity of rapid HD in patients affected by acute renal failure or at the start of their program of maintenance HD. Most of the uremic stable patients on maintenance HD are highly unlikely to develop DDS under any circumstance. © 1987 by the National Kidney Foundation, Inc. INDEX WORDS: Dialysis disequilibrium syndrome; computed tomography; EEG spectral analysis; hypertonic hemodiafiltration; standard hemodialysis.
T
HE DIALYSIS disequilibrium syndrome (DDS) describes a constellation of symptoms ranging from mild (eg , nausea , vomiting, headache, muscle cramps, and fatigue following dialysis) to severe (eg, medical emergencies due to seizures, arrhythmias , and even coma). 1.2 It appears that modern methods of dialysis have altered the clinical picture of DDS . Most of the seizures, comas, and deaths were reported prior to 1970, whereas the symptoms of DDS reported in the 1970s have generally been mild, consisting of From the Department of Medicine. the Division of Nephrology. the Department of Radiology. and the Department of Applied Sciences ill Medicin e, University of Alberta, Edmonton, Canada. Supported by Grant 7219 from the Special Services and Research Committee of University of Alberta Hospitals and from a grant from Hospal Canada Ltd. Dr Basile is a recipielll of the Len Grasley Memorial Research Fellowship granted by the Kidney Foundation of Canada, Alberta Branch. Address reprilll requests to Carlo Basile, MD, Via Battisti 192 , 74100 Taranto , Italy. © 1987 by the National Kidney Foundation, 11lc. 0272-6386/87/0906-0003$3.00/0 462
nausea, weakness, headache, fatigue, and muscle cramps.2 The pathogenesis of DDS is not fully understood. A rise in cerebrospinal fluid (CSF) pressure, intracellular acidosis of brain, and brain edema may all be contributory. 3 The CSF pressure in both uremic patients and experimental animals is normal. Following hemodialysis (HD) of uremic humans or laboratory animals with acute renal failure, the CSF pressure rises. The presence of elevated CSF pressure is thought by many to indicate cerebral edema. An increase in CSF pressure, however, may be secondary to either increased CSF volume, increased brain blood volume (cerebral vasodilation) , or increased brain water content (cerebral edema).4 In most of the experimental studies, however, brain edema was a consistent finding. 57 The computed tomography (CT) of the brain could be a simple and reliable technique for the investigations on DDS8; although the CT scan cannot differentiate between increases in brain blood volume and cerebral edema, the measurement of the brain density8 and the ventricular size9 is able
American Journal of Kidney Diseases, Vol IX, No 6 (June), 1987: pp 462-469
DIALYSIS DISEQUILIBRIUM SYNDROME
to detect even minimal and subclinical brain swelling . lo With such a technique, La Greca et al 8 observed a significant postdialysis decrease in brain density in a group of stable uremic patients on maintenance HD . 8 Deterioration of the EEG has been frequently seen in association with DDS, suggesting that the clinical manifestations of DDS are attributable to cerebral disfunction. 11·14 Furthermore, it has been demonstrated that rapid HD in uremic animals was able to develop DDS,15 whereas slow HD ,15 the use of glycerol or mannitol,7 of bicarbonate, 14 and of a high sodium concentration in the dialysate, 16 were able to prevent it. Moreover, the last three treatments7.14.16 have been shown to prevent the EEG deterioration during dialysis. Thus, we designed a study that, using the CT of the brain and the computerized EEG analysis before and after a dialysis run, had to answer the following questions: (1) Is any change occurring in brain density, ventricular size, and EEG recording when stable uremic patients undergo a routine HD run? (2) Is any worsening occurring in brain density, ventricular size, and EEG patterns when the same patients are dialyzed with a shorter and more efficient type of treatment? This design was included in a more comprehensive randomized crossover study (H HDF study) that had to compare the effects of hypertonic hemodiafiltration (H HDF) and HD on several pathophysiologic aspects of uremia, ie, cardiac function, plasma volume changes, and alterations of the peripheral and central nervous system. 17 H HDF is a dialysis technique we have developed recently that is shorter and more efficient than standard HD.18.19 Thus, the specific design of the present study (DDS study) was to compare the effects of a rapid and efficient dialysis (H HDF) and a standard HD on brain density, ventricular size, and EEG pattern in a group of stable uremic patients . MATERIALS AND METHODS
H HDF Study Protocol Thirteen stable uremic patients at the dialy sis unit at the University of Alberta Hospital volunteered for a crossover study, which included 4 months each of HD and H HDF in a random sequence. The study had been approved by the Ethical Review Committee and the Radioisotope and Radiation Committee a t the University of Alberta Hospitals. The criteria for admission to the study were (1) maintenance HD for at least 6 months, (2) absence of systemic diseases , and
463 (3) no clinical evidence of coronary artery disease and heart failure. The HD prescription was that believed " adequate" by the attending nephrologist at that time for each of the patients: the blood flow rate was - 250 mLimin; the sodium and the acetate concentration in the bath were - 135 and - 35 mmollL, respectively ; one patient underwent bicarbonate dialysis . Hollow fiber dialyzers (all but one in cuprophane and the other in cellulose acetate), with surface areas ranging from 1.1 to 1.4 m 2 (mean surface area 1.23 m2 ) , were used. H HDF consists of a short-time low-volume (up to a maximum of 6 L of postdilution replacing solution containing sodium at a concentration either of 180 or 220 mmol/L hemofiltration session simultaneous with acetate HD (Na + concentration in the dialysate -135 mmoIlL). The sodium modeling of H HDF has been described elsewhere. 2o . 2I The equipment consists of a dialysate delivery module in which a periodically renewed fresh dialysate circulates in a determined constant volume. Any quantity of liquid removed from this closed-circuit fixed volume will be automatically compensated by an equal quantity of fluid extracted from the patient (Monitral; Hospal, Canada).22 The hypertonic reinjection solution is infused in a postdilution site with the aid of a roller pump (BSM 22; Hospal , Montreal), whose infusion rate can be manually regulated according to the requirements of fluid balance . The dialyzer used was the Biospal 3000 S (polyacrylonitrile membrane S, 1.2 m 2 ; Hospal) . Patients maintained their usual dialysis schedule (three times a week) throughout the study. The treatment time , however, was reduced in H HDF by approximately 25 % to 30 % for each of them, as it has been previously demonstrated that this is the percentage reduction in treatment time that can be safely done , compared to HD.19
DDS Study Protocol Patients. The DDS study protocol was completed b y II of the 13 patients who volunteered for the H HDF study. Of the dropouts, one patient was transplanted soon after commencing the study and the second dropped out during his HO control period for personal reasons . The 11 patients (ten men , one woman; mean age 37.5 ± 2.8 SEM years; mean body weight 74.7 ± 7.1 kg) were o nvarious medications that were not changed throughout the study (among the drugs involved we re a {3-blocker in one patient, a vasodilator for hypertension in another, but none were taking digitalis) . Dialysis procedures. Patients underwent the computerized EEG recording and the CT of the brain just before and immediately after a midweek dialysis run , always in the same sequence . Each patient underwent' two assessments, once during the 4 months of the H HOF treatment and once during the 4 months of the HO treatment. A period of stabilization on each procedure was allowed, and studies were conducted 44 ± 6 days after the start of H HOF and 39 ± 6 days after the start of HO. The characteristics of the two dialysis treatments are as shown in Table I. Three points are to be emphasized : first , a different blood flow rate and treatment time was maintained for the HO and H HOF runs, as is routinely done for our HO and H HOF patients; second, we designed it to pair-match the body weight loss; and third, all the experimental runs were performed with the same dialysis machine (Monitral, Hospal) .
464
BASILE ET AL
Table 1. Factor Body weight loss (kg) Duration of treatment (min) Weight loss rate (g/min) Blood flow rate (mUmin) Dialysate flow rate (mUmin) Dialyzer membrane and surface Dialysis solution components (mmoI/L) Sodium Potassium Calcium Magnesium Chloride Glucose Bicarbonate Acetate Acetic Acid Reinfusion solution components (mmoI/L) Sodium Chloride Bicarbonate
Factors Relating to Dialysis Treatments Standard HD (N
=
11)
2.3 ± 0.3240 ± 09.6 + 1.4-250 -500 Cuprophane or cellulose acetate (1.1-1.4 m2)
pt
H HDF(N = 11)
NS <.001 <.01
2.5 ± 0.4180 ± 014.1 ± 2.2-400 -500 Polyacrylonitrile (1.2 m2) 135.0 1.0 2.0 0.75 111.5 11.1
135.0 1.0 1.55 0.75 105.6 35.0 (1 patient) 35.0 2.0 (1 patient)
30
180 or 220 100 or 120 80 or 100
-Means ± SEM. tStudent's paired t test Spectral EEG analysis. Eight channels of EEG data were recorded. These channels were bipolar in configuration and were derived from the following locations on the International 10-20 System 23 : F 7 -T 3, T,-T 3, P,-T" F,-P3' and homologous locations over the right hemisphere. The right earlobe served as ground. EEG data were recorded on a model 8-16 C Grass electroencephalograph (USA). The EEG recording session consisted of a 2-minute period during which the patient performed a mental arithmetic exercise with eyes closed. Recordings were processed off-line using VAX 111750 computer. Processing of EEG consisted of a traditional spectral analysis. The recording from each channel was divided into segments consisting of 128 consecutive samples. Each segment was modified with a Hanning data window and was Fourier transformed. Coefficients of each transformed segment were combined as appropriate to form estimates of the power in the 3 to 7 Hz (tJ) and 7 to 13 Hz (0') bands. Segments were obtained from the recording in a half overlapped fashion with each new segment utilizing the 64 unmodified samples from the previous segment. All transformed segments were reduced to two frequency bands, as described, and those corresponding to each channel were averaged by band to form a measure of average power for the channel for each recording period. These measures of average power were then normalized by expressing the power in each band as a fraction of the total power in the two bands for each channel. To quantify the relative amounts of slow-wave-related activity, we calculated the amount of spectral power between 3 and 7 Hz, and divided it by the amount of power between 7 and 13 Hz. This yields a frequency fraction that increases as slow-wave activity in the EEG increases, and decreases as slow waves disappear. CT of the brain. A scout film was obtained in each patient. With the gantry vertical and the patient's chin slightly flexed, two cuts were obtained: (I) the basal cut at the level of the third ventricle, showing the frontal horns, the thalami, and internal
capsule on each side, in addition to sylvian fissures; and (2) a cut approximately 30 mm above the first, showing the centrum semiovale on each side with good visualization of cortex and sulci. At this time, skin markings were made on the head for alignment of the laser beam mounted on the CT gantry so that when the scan was repeated after dialysis, exactly the same cuts at the same angle were obtained. The measurements were all performed on the Independent Console of a GE9800 scanner (USA). For brain density measurements, an area of 0.12 cm2 was employed, and the average CT numbers (Hounsfield Units) of the area were obtained at the following sites: (1) base cut: frontal cortex left side, left internal capsule posterior limb, thalamus, CSF in the third ventricle superior cut (Fig lA); and (2) apical cut (through the level of centrum semiovale): left frontal parietal cortex grey matter, right occipital parietal cortex grey matter, left centrum semiovale (Fig 1B). Exactly the same area was measured for each set of scans, using an XY axis localization given by the machine. It should be borne in mind that the density of each area is inversely proportional to its water content (eg, bone = 1000, CSF = 5 to 10 Hounsfield Units). The following ventricular measurements were also made on the base cuts for all available scans: bicaudate diameter of the frontal horns measuring the maximum distance between the vertical walls at the level of the caudate nuclei, and the widest portion of the third ventricle (Fig 1C). In addition, ten healthy subjects with a similar mean age (37.2 ± 5.3 years) underwent the same CT studies and served as controls. It must be stressed also that those of us who analyzed the EEG (ZJK) and the CT (JDRM) results were completely unaware of what type of dialysis treatment was under study. Biochemistry. Blood was drawn for biochemical and gas analysis before and after a dialysis run. BUN, plasma
465
DIALYSIS DISEQUILIBRIUM SYNDROME
creatinine, calcium, protein, albumin, glucose, sodium, potassium, and chloride were determined by means of routine automated methods. Plasma magnesium was measured using the atomic absorption spectrophotometry (model 303; PerkinElmer, USA) , and plasma osmolality was measured using the freezing-point depression. Blood gases were measured using a Corning 178 gas-analyzer (USA). Statistics. All data are expressed as means ± SEM. Statistical analysis was done using paired and unpaired Student's t tests, when appropriate. Analyses were performed using a basic statistical program on the Amdahl 370 computer.
RESULTS
Before dialysis, there was no significant difference in any of the biochemical data between HD and H HDF. After dialysis, a significant difference in plasma sodium, magnesium, osmolality, arterial PC0 2 , and bicarbonate was noted between the two treatments (Table 2). Brain density was not affected by dialysis, therefore the CT numbers of the seven specific brain structures (base and apical cuts) did not show any change when comparing the pre- with the postdialysis data, either of HD or of H HDF. Furthermore, no significant difference could be noted when comparing the CT numbers at the start and at the end of both treatments (Table 3). Similarly, no significant difference could be demonstrated between the two treatments or within either treatment regarding the transverse diameter of the third ventricle and the bicaudate diameter of the lateral ventricles (Table 3). The group of control healthy subjects showed a density of the same brain structures that was generally higher than that of the uremic population. The difference reached a statistical significance in many of these structures (Table 4). Finally, no significant in-between- and withintreatment (with the exception of one channel in HD) difference was observed when analyzing the EEG% power (3-7/7-13 Hz) data (Table 5). DISCUSSION
Fig 1. (A) Basal cut of a CT scan showing the site of brain density sampling: 1, left frontal cortex; 2, left internal capsule; 3, left thalamus. (B) Apical cut showing the site of density sampling: 1, left parietal cortex; 2, right parietal cortex; 3, left centrum semiovale. (C) Basal cut showing the site of ventricular size measurement: 1, bicaudate diameter of the lateral ventricles; 2, transverse diameter of the third ventricle.
This study was undertaken to evaluate the acute and chronic effects of H HD F and HD on brain hydration and EEG patterns in 11 stable uremic patients. They were studied approximately 6 weeks after they had started on H HDF or HD to allow for stabilization with the assigned method. This was designed to investigate if a difference in cerebral hydration and EEG could be demonstrated due to a chronic effect of H HDF, as compared with HD. Studies were also carried out after dialysis to obtain a profile of acute intradialysis
466
BASILE ET AL Table 2.
Biochemical Data in 11 Patients Before and After a Run of H HDF and a Run of HD Pre
Post
H HDF
BUN (mmoI/L) Plasma creatinine (,.,moI/L) Plasma calcium (mmoI/L) Plasma protein (giL) Plasma albumin (giL) Plasma glucose (mmoI/L) Plasma sodium (mmoI/L) Plasma potassium (mmoI/L) Plasma chloride (mmoI/L) Plasma magnesium (mmoI/L) Plasma osmolality (mmol/kg H2 O) Arterial P0 2 (mm Hg) Arterial PC0 2 (mm Hg) Arterial [H'I3] (nmoI/L) Arterial bicarbonate (mmoI/L)
26.0 1268 2.30 66.9 40.1 4.9 139.3 4.9 103.9 1.13 309.2 97.4 33.7 37.1 20.2
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
HD
2.0 75 0.06 0.8 0.5 0.3 0.4 0.3 1.6 0.04 2.4 3.4 1.5 3.2 0.6
26.8 1154 2.35 68.7 41.9 5.0 138.3 4.5 105.7 1.19 304.7 95.9 32.5 40.9 19.2
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
P*
H HDF
1.6 82 0.12 1.3 1.1 0.3 1.0 0.3 1.6 0.05 1.3 2.8 1.2 0.7 0.8
10.7 584 2.89 76.4 48.0 6.1 139.4 3.0 101.3 1.0 293.4 90.8 35.0 35.1 23.8
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
1.1 36 0.05 2.5 1.6 0.4 0.9 0.1 0.9 0.02 2.3 4.3 1.7 0.7 0.9
HD
< .05
<.001 <.01 < .02 <.02
10.6 555 2.77 76.1 47.8 5.5 135.5 3.1 100.3 1.12 286.9 90.2 29.5 35.6 19.9
± ± ± ± ± ± ± ± ± ± ± ±
1.2 47 0.07 1.6 1.3 0.3 1.0 0.2 0.9 0.02 1.5 4.1 ± 1.4 ± 0.6 ± 1.0
All values are expressed as means ± SEM. 'Student's paired t test.
changes. To the best of our knowledge, this is the first work that combines CT of the brain and EEG spectral analysis in the study of DDS. H HDF runs were conducted in a shorter (three v four hours) and more efficient way, compared with HD; in fact, BUN and plasma creatinine were not significantly different at the end of the sessions, even if the H HDF treatment time had been reduced by 25 %. Thus, one would predict for H HDF a worse outcome in terms of increased cerebral hydration and EEG, according to the literature,5.6.15 unless the possible protective effect of Table 3.
the hypertonic reinfusion solution is taken into account. 16 Surprisingly, no difference could be observed in the brain density, ventricular size, and EEG% power (3-717-13 Hz) when comparing the predialysis with the postdialysis data, in both H HDF and HD. Furthermore, no difference could be observed when comparing the two treatments, in both the predialysis and the postdialysis setting. Our CT findings are completely different from those previously reported in HD patients 8 : La Greca et al observed a significant decrease in brain
Brain Density and Ventricular Size in 11 Patients Before and After a Run of H HDF and HD H HDF Pre
Brain density (Hounsfield units) Left frontal cortex Left internal capsule Left thalamus CSF in the third ventricle Right occipital parietal cortex (grey matter) Left frontal parietal cortex (grey matter) Left centrum semiovale Ventricular size (cm) Transverse diameter of the third ventricle Bicaudate diameter of the lateral ventricles
37.3 30.8 33.7 7.6
± ± ± ±
HD Post
1.2 1.4 0.9 1.0
37.9 30.8 33.3 5.4
± ± ± ±
1.4 1.3 0.9 1.4
Pre
37.3 31.0 33.3 7.7
± ± ± ±
Post
1.3 1.0 0.7 1.0
37.2 29.4 33.0 6.6
± ± ± ±
1.6 1.0 0.4 1.0
45.2 ± 1.7
44.5 ± 1.8
44.7 ± 1.9
46.5 ± 2.8
43.9 ± 1.8 32.0 ± 1.0
44.8 ± 1.9 30.4 ± 0.9
42.5 ± 1.8 30.0 ± 0.9
42.9 ± 1.6 29.8 ± 1.3
0.40 ± 0.05
0.36 ± 0.06
0.42 ± 0.06
0.39 ± 0.73
1.36 ± 0.10
1.31 ± 0.12
1.39 ± 0.09
1.37 ± 0.12
All values are expressed as mean ± SEM. No statistical difference was found when comparing pre- v post-H HDF, pre- v post-HD, pre-H HDF v pre-HD, post-H HDF v post-HD.
467
DIALYSIS DISEQUILIBRIUM SYNDROME Table 4.
Comparison of Brain Density and Ventricular Size Between 10 Normal Adults (N) and 11 Patients Before a Run of Either H HDF or HD p'
H HDF
Brain density (Hounsfield units) Left frontal cortex Left internal capsule Left thalamus CSF in the third ventricle Right occipital partietal cortex (grey matter) Left frontal parietal cortex (grey matter) Left centrum semiovale Ventricular size (cm) Transverse diameter of the third ventricle Bicaudate diameter of the lateral ventricles
37.3 30.8 33.7 7.6
± 1.2 ± 1.4
p'
N
<.001
± 0.9
± 1.0
43.8 30.5 35.9 9.0
± ± ± ±
1.0 0.8 0.9 0.7
HD
<.001 <.05
48.2 ± 1.2
42.5 ± 1.7 43.9 ± 1.8 32.0 ± 1.0
± ± ± ±
1.3 1.0 0.7 1.0
44.7 ± 1.9
48.6 ± 0.9 34.2 ± 0.4
<.05
37.3 31.0 33.3 7.7
<.01 <.001
42.5 ± 1.8 30.0 ± 0.9
0.40 ± 0.05
0.42 ± 0.05
0.42 ± 0.06
1.36 ± 0.10
1.35 ± 0.17
1.39 ± 0.09
All values are expressed as means ± SEM . • Unpaired Student's t test.
Dialysis patients, however, seem to suffer a chronic subclinical brain swelling. Cerebral edema has not been demonstrated in the brain of patients or laboratory animals with chronic renal failure, either by biochemical or histologic criteria. 2 This preliminary observation, however, seems to be worthy of further investigation. Our patients showed important EEG abnormalities when compared to normal subjects, as already reported, 17.2426 but did not show any EEG deterioration when undergoing either the rapid or the standard dialysis treatment. It must be pointed out that our evaluation of the EEG did not take into account possible effects on visually evoked cortical potentials,24 discriminant analysis,25 or photic driving.23 These parameters may transiently dete-
density during dialysis; furthermore, they claimed that HD induced a cerebral water gain from a predialysis "dry" brain state to a postdialysis normohydrated brain situation. On the basis of this report, it is very hard to explain why DDS symptoms appear at the moment of the correction to normohydration. To the contrary, we could not observe any postdialysis decrease in brain density, either in HD or in H HDF; furthermore, our patients showed a generalized decreased brain density when compared to normal subjects. The reason for this discrepancy with the report of La Greca et al 8 is not clear. Summarizing our CT data, neither a rapid dialysis technique (H HDF) nor the standard HD are able to provoke an acute change in brain density and ventricular size. Table 5.
EEG, % Power (3-717-13 Hz) in 11 Patients Before and After a Run of H HDF and HD H HDF
HD Post
Pre
Pre
Post
Left hemisphere FTTJ Ts-TJ PJ-Ts FTP J
1.48 0.67 0.75 0.99
± ± ± ±
0.26 0.14 0.16 0.20
1.27 0.76 0.77 1.00
± ± ± ±
0.21 0.14 0.14 0.22
1.49 0.84 0.91 1.02
± ± ± ±
0.27 • 0.23 0.25 0.29
1.17 1.00 0.98 1.13
± ± ± ±
0.17 0.24 0.22 0.26
Right hemisphere Fa-T 4 T6-T 4 P4 -T 6 Fa-P 4
1.05 0.51 0.61 0.79
± ± ± ±
0.13 0.11 0.16 0.16
0.99 0.63 0.66 0.85
± ± ± ±
0.17 0.11 0.14 0.18
1.54 0.59 0.73 1.00
± ± ± ±
0.46 0.19 0.27 0.31
1.36 0.73 0.82 0.96
± ± ± ±
0.24 0.17 0.24 0.21
All values are expressed as means ± SEM. No statistical difference was observed when comparing pre- v post-H HDF, pre-H HDF v pre-HD, post-H HDF v post-HD. 'Only one significant difference (P < .02) was found when comparing pre- v post-HD; paired Student's t test.
468
BASILE ET AL
riorate during dialysis. Lewis et al,27 however, have been able to show that evoked potential latencies tend to improve in the first and in the following hours after dialysis. Our EEG data are in agreement with those of Kiley et al in uremic patients 28 and those of Arieff et al, 7 who were unable to observe any significant alteration in the EEG spectral analysis in uremic dogs during and after rapid HD. To the contrary, many studies have shown an EEG deterioration during dialysis.1I14.16 As it has already been pointed out, 7 though, not all the patients in these reports suffered an EEG deterioration during dialysis12.13.16; furthermore, the assessment of these reports shows that in no instance was a systematic examination of the EEG performed, and no standardized criteria were used for evaluating or quantitating deterioration of the EEG. In conclusion, this study shows neither a post-
dialysis change in brain density and ventricular size nor a postdialysis EEG deterioration in a group of stable uremic patients undergoing both a rapid and a standard dialysis treatment. Thus, DDS must be considered a peculiarity of rapid HD in patients, particularly children, affected by acute renal failure or at the start of their program of mai!ltenance HD. 3 Most of the uremic stable patients on maintenance HD are highly unlikely to develop DDS under any circumstances. 29 Furthermore, the diagnosis of DDS has become recently a "wastebasket" for a number of disorders that can occur in patients with renal failure. 30 ACKNOWLEDGMENT The authors wish to thank the personnel of the Dialysis Unit at the University of Alberta Hospitals for their assistance and cooperation, C. Ness for her aid in preparing the manuscript, and S. Breitkreuz and 1. Buicliu for their skillful technical assistance.
REFERENCES I. Wakim KG: The pathophysiology of the dialysis disequilibrium syndrome. Mayo Clin Proc 44:406-429, 1969 2. Arieff AI: Neurological complications of uremia, in Brenner BM, Rector FC (eds): The Kidney. Philadelphia, Saunders, 1981, pp 2306-2343 3. Arieff AI: Dialysis disequilibrium syndrome, in Massry SG, Glassock RJ (eds): Textbook of Nephrology. Baltimore, Williams & Wilkins, 1983, pp 24-26 4. Rosomoff HL, Zugibe FT: Distribution of intracranial contents in experimental edema. Arch Neurol 9:26-34, 1963 5. Pappius HM, Oh JH, Dossetor 18: The effects of rapid hemodialysis on brain tissues and cerebrospinal fluid of dogs. Can J Physiol Pharmacol 45: 129-147, 1967 6. Arieff AI, Guisado R, Massry SG, et al: Central nervous system pH in uremia and the effects of hemodialysis. J Clin Invest 58:306-311, 1976 7. Arieff AI, Lazarowitz VC, Guisado R: Experimental dialysis disequilibrium syndrome: Prevention with glycerol. Kidney Int 14:270-278, 1978 8. La Greca G, Biasioli S, Chiaramonte S, et al: Studies on brain density in hemodialysis and peritoneal dialysis. Nephron 31: 146-150, 1982 9. Krane EJ, Rockoff MA, Wallman JK, et al: Subclinical brain swelling in children during treatment of diabetic ketoacidosis. N Engl J Med 312:1147-1151, 1985 10. Villafana T: Physics and instrumentation, in Lee SH, Rao KCVG (eds): Cranial Computed Tomography. New York, McGraw-Hill, 1983, pp 1-46 II. Kennedy AC, Linton AL, Luke RG, et al: Electroencephalographic changes during hemodialysis. Lancet 1:408411, 1963 12. Hampers CL, Doak PB, Callaghan MN, et al: The electroencephalogram and spinal fluid during hemodialysis. Arch Intern Med 118:340-346, 1966 13. Jacob JC, Gloor P, Elwan OH, et al: Electroencephalo-
graphic changes in chronic renal failure. Neurology 15:419429, 1965 14. Hampl H, Klopp HW, Michels N, et al: Electroencephalogram investigations of the disequilibrium syndrome during bicarbonate and acetate dialysis. Proc Eur Dial Transplant Assoc 19:351-359, 1982 15. Arieff AI, Massry SG, Barrientos A, et al: Brain water and electrolyte metabolism in uremia: Effects of slow and rapid hemodialysis. Kidney Int 4:177-187, 1973 16. Port FK, Johnson WJ, Klass DW: Prevention of dialysis disequilibrium syndrome by use of high sodium concentration in the dialysate. Kidney Int 3:327-333, 1973 17. Basile C, Bornstein RA, Elleker MG, et al: The chronic efficacy and safety of hypertonic hemodiafiltration: A randomized cross-over study (submitted) 18. Basile C, Di Maggio A, Curino E, et al: Pharmacokinetics of netilmicin in hypertonic hemodiafiltration and standard hemodialysis. Clin Nephrol 24:305-309, 1985 19. Basile C, Di Maggio A, Manente G, et al: Solute kinetics in hypertonic hemodiafiltration and standard hemodialysis. Am J Kidney Dis 7:483-489, 1986 20. Basile C, Di Maggio A , Longo S, et al: Sodium balance in hypertonic hemodiafiltration. Blood Purif 2:70-75, 1984 21. Basile C, Di Maggio A, Scatizzi A: Long-term experience with hypertonic hemodiafiltration. Proc Eur Dial Transplant Assoc 22:303-307, 1985 22. Kirkwood RG, O'Brien M, Butruille Y, et al: Initial clinical evaluation of the Hospal artificial kidney system. Trans Am Soc Artif Intern Organs 24:509-515, 1978 23. International Federation of Societies for Electroencephalography and Clinical Neurophysiology: The ten-twenty electrode system. Am J EEG Technoll:13-19, 1961 24. Teschan PE, Ginn HE, Bourne JR, et al: Quantitative indices of clinical uremia. Kidney Int 15:676-697, 1979 25. Teschan PE, Bourne JR, Reed RB, et al: Electrophys-
DIALYSIS DISEQUILIBRIUM SYNDROME
iological and neurobehavioral responses to therapy: the National Cooperative Dialysis Study. Kidney Int 23:S58-S65 , 1983 (suppl 13) 26. Bourne JR, Teschan PE: Computer methods, uremic encephalopathy and adequacy of dialysis. Kidney Int 24:496-506, 1983 27. Lewis EG, O'Neill WM, Dustman RE, et al: Temporal effects of hemodialysis on measures of neural efficiency. Kidney Int 17:357-363, 1980
469 28. Kiley JE, Woodruff MW, Pratt KL: Evaluation of encephalopathy by EEG frequency analysis in chronic dialysis patients. Clin Nephrol 5:245-250, 1976 29. Arieff AI, Massry SG: Dialysis disequilibrium syndrome, in Massry SG, Sellers AL (eds): Clinical Aspects of Uremia and Dialysis. Springfield, IL, Thomas, 1976, pp 34-52 30. Mahoney CA, Arieff AI: Uremic encephalopathies: Clinical, biochemical, and experimental features. Am J Kidney Dis 2:324-336, 1982
T
HE DIALYSIS disequilibrium syndrome (DDS) describes a constellation of symptoms ranging from mild (eg , nausea , vomiting, headache, muscle cramps, and fatigue following dialysis) to severe (eg, medical emergencies due to seizures, arrhythmias , and even coma). 1.2 It appears that modern methods of dialysis have altered the clinical picture of DDS . Most of the seizures, comas, and deaths were reported prior to 1970, whereas the symptoms of DDS reported in the 1970s have generally been mild, consisting of From the Department of Medicine. the Division of Nephrology. the Department of Radiology. and the Department of Applied Sciences ill Medicin e, University of Alberta, Edmonton, Canada. Supported by Grant 7219 from the Special Services and Research Committee of University of Alberta Hospitals and from a grant from Hospal Canada Ltd. Dr Basile is a recipielll of the Len Grasley Memorial Research Fellowship granted by the Kidney Foundation of Canada, Alberta Branch. Address reprilll requests to Carlo Basile, MD, Via Battisti 192 , 74100 Taranto , Italy. © 1987 by the National Kidney Foundation, 11lc. 0272-6386/87/0906-0003$3.00/0 462
nausea, weakness, headache, fatigue, and muscle cramps.2 The pathogenesis of DDS is not fully understood. A rise in cerebrospinal fluid (CSF) pressure, intracellular acidosis of brain, and brain edema may all be contributory. 3 The CSF pressure in both uremic patients and experimental animals is normal. Following hemodialysis (HD) of uremic humans or laboratory animals with acute renal failure, the CSF pressure rises. The presence of elevated CSF pressure is thought by many to indicate cerebral edema. An increase in CSF pressure, however, may be secondary to either increased CSF volume, increased brain blood volume (cerebral vasodilation) , or increased brain water content (cerebral edema).4 In most of the experimental studies, however, brain edema was a consistent finding. 57 The computed tomography (CT) of the brain could be a simple and reliable technique for the investigations on DDS8; although the CT scan cannot differentiate between increases in brain blood volume and cerebral edema, the measurement of the brain density8 and the ventricular size9 is able
American Journal of Kidney Diseases, Vol IX, No 6 (June), 1987: pp 462-469
DIALYSIS DISEQUILIBRIUM SYNDROME
to detect even minimal and subclinical brain swelling . lo With such a technique, La Greca et al 8 observed a significant postdialysis decrease in brain density in a group of stable uremic patients on maintenance HD . 8 Deterioration of the EEG has been frequently seen in association with DDS, suggesting that the clinical manifestations of DDS are attributable to cerebral disfunction. 11·14 Furthermore, it has been demonstrated that rapid HD in uremic animals was able to develop DDS,15 whereas slow HD ,15 the use of glycerol or mannitol,7 of bicarbonate, 14 and of a high sodium concentration in the dialysate, 16 were able to prevent it. Moreover, the last three treatments7.14.16 have been shown to prevent the EEG deterioration during dialysis. Thus, we designed a study that, using the CT of the brain and the computerized EEG analysis before and after a dialysis run, had to answer the following questions: (1) Is any change occurring in brain density, ventricular size, and EEG recording when stable uremic patients undergo a routine HD run? (2) Is any worsening occurring in brain density, ventricular size, and EEG patterns when the same patients are dialyzed with a shorter and more efficient type of treatment? This design was included in a more comprehensive randomized crossover study (H HDF study) that had to compare the effects of hypertonic hemodiafiltration (H HDF) and HD on several pathophysiologic aspects of uremia, ie, cardiac function, plasma volume changes, and alterations of the peripheral and central nervous system. 17 H HDF is a dialysis technique we have developed recently that is shorter and more efficient than standard HD.18.19 Thus, the specific design of the present study (DDS study) was to compare the effects of a rapid and efficient dialysis (H HDF) and a standard HD on brain density, ventricular size, and EEG pattern in a group of stable uremic patients . MATERIALS AND METHODS
H HDF Study Protocol Thirteen stable uremic patients at the dialy sis unit at the University of Alberta Hospital volunteered for a crossover study, which included 4 months each of HD and H HDF in a random sequence. The study had been approved by the Ethical Review Committee and the Radioisotope and Radiation Committee a t the University of Alberta Hospitals. The criteria for admission to the study were (1) maintenance HD for at least 6 months, (2) absence of systemic diseases , and
463 (3) no clinical evidence of coronary artery disease and heart failure. The HD prescription was that believed " adequate" by the attending nephrologist at that time for each of the patients: the blood flow rate was - 250 mLimin; the sodium and the acetate concentration in the bath were - 135 and - 35 mmollL, respectively ; one patient underwent bicarbonate dialysis . Hollow fiber dialyzers (all but one in cuprophane and the other in cellulose acetate), with surface areas ranging from 1.1 to 1.4 m 2 (mean surface area 1.23 m2 ) , were used. H HDF consists of a short-time low-volume (up to a maximum of 6 L of postdilution replacing solution containing sodium at a concentration either of 180 or 220 mmol/L hemofiltration session simultaneous with acetate HD (Na + concentration in the dialysate -135 mmoIlL). The sodium modeling of H HDF has been described elsewhere. 2o . 2I The equipment consists of a dialysate delivery module in which a periodically renewed fresh dialysate circulates in a determined constant volume. Any quantity of liquid removed from this closed-circuit fixed volume will be automatically compensated by an equal quantity of fluid extracted from the patient (Monitral; Hospal, Canada).22 The hypertonic reinjection solution is infused in a postdilution site with the aid of a roller pump (BSM 22; Hospal , Montreal), whose infusion rate can be manually regulated according to the requirements of fluid balance . The dialyzer used was the Biospal 3000 S (polyacrylonitrile membrane S, 1.2 m 2 ; Hospal) . Patients maintained their usual dialysis schedule (three times a week) throughout the study. The treatment time , however, was reduced in H HDF by approximately 25 % to 30 % for each of them, as it has been previously demonstrated that this is the percentage reduction in treatment time that can be safely done , compared to HD.19
DDS Study Protocol Patients. The DDS study protocol was completed b y II of the 13 patients who volunteered for the H HDF study. Of the dropouts, one patient was transplanted soon after commencing the study and the second dropped out during his HO control period for personal reasons . The 11 patients (ten men , one woman; mean age 37.5 ± 2.8 SEM years; mean body weight 74.7 ± 7.1 kg) were o nvarious medications that were not changed throughout the study (among the drugs involved we re a {3-blocker in one patient, a vasodilator for hypertension in another, but none were taking digitalis) . Dialysis procedures. Patients underwent the computerized EEG recording and the CT of the brain just before and immediately after a midweek dialysis run , always in the same sequence . Each patient underwent' two assessments, once during the 4 months of the H HOF treatment and once during the 4 months of the HO treatment. A period of stabilization on each procedure was allowed, and studies were conducted 44 ± 6 days after the start of H HOF and 39 ± 6 days after the start of HO. The characteristics of the two dialysis treatments are as shown in Table I. Three points are to be emphasized : first , a different blood flow rate and treatment time was maintained for the HO and H HOF runs, as is routinely done for our HO and H HOF patients; second, we designed it to pair-match the body weight loss; and third, all the experimental runs were performed with the same dialysis machine (Monitral, Hospal) .
464
BASILE ET AL
Table 1. Factor Body weight loss (kg) Duration of treatment (min) Weight loss rate (g/min) Blood flow rate (mUmin) Dialysate flow rate (mUmin) Dialyzer membrane and surface Dialysis solution components (mmoI/L) Sodium Potassium Calcium Magnesium Chloride Glucose Bicarbonate Acetate Acetic Acid Reinfusion solution components (mmoI/L) Sodium Chloride Bicarbonate
Factors Relating to Dialysis Treatments Standard HD (N
=
11)
2.3 ± 0.3240 ± 09.6 + 1.4-250 -500 Cuprophane or cellulose acetate (1.1-1.4 m2)
pt
H HDF(N = 11)
NS <.001 <.01
2.5 ± 0.4180 ± 014.1 ± 2.2-400 -500 Polyacrylonitrile (1.2 m2) 135.0 1.0 2.0 0.75 111.5 11.1
135.0 1.0 1.55 0.75 105.6 35.0 (1 patient) 35.0 2.0 (1 patient)
30
180 or 220 100 or 120 80 or 100
-Means ± SEM. tStudent's paired t test Spectral EEG analysis. Eight channels of EEG data were recorded. These channels were bipolar in configuration and were derived from the following locations on the International 10-20 System 23 : F 7 -T 3, T,-T 3, P,-T" F,-P3' and homologous locations over the right hemisphere. The right earlobe served as ground. EEG data were recorded on a model 8-16 C Grass electroencephalograph (USA). The EEG recording session consisted of a 2-minute period during which the patient performed a mental arithmetic exercise with eyes closed. Recordings were processed off-line using VAX 111750 computer. Processing of EEG consisted of a traditional spectral analysis. The recording from each channel was divided into segments consisting of 128 consecutive samples. Each segment was modified with a Hanning data window and was Fourier transformed. Coefficients of each transformed segment were combined as appropriate to form estimates of the power in the 3 to 7 Hz (tJ) and 7 to 13 Hz (0') bands. Segments were obtained from the recording in a half overlapped fashion with each new segment utilizing the 64 unmodified samples from the previous segment. All transformed segments were reduced to two frequency bands, as described, and those corresponding to each channel were averaged by band to form a measure of average power for the channel for each recording period. These measures of average power were then normalized by expressing the power in each band as a fraction of the total power in the two bands for each channel. To quantify the relative amounts of slow-wave-related activity, we calculated the amount of spectral power between 3 and 7 Hz, and divided it by the amount of power between 7 and 13 Hz. This yields a frequency fraction that increases as slow-wave activity in the EEG increases, and decreases as slow waves disappear. CT of the brain. A scout film was obtained in each patient. With the gantry vertical and the patient's chin slightly flexed, two cuts were obtained: (I) the basal cut at the level of the third ventricle, showing the frontal horns, the thalami, and internal
capsule on each side, in addition to sylvian fissures; and (2) a cut approximately 30 mm above the first, showing the centrum semiovale on each side with good visualization of cortex and sulci. At this time, skin markings were made on the head for alignment of the laser beam mounted on the CT gantry so that when the scan was repeated after dialysis, exactly the same cuts at the same angle were obtained. The measurements were all performed on the Independent Console of a GE9800 scanner (USA). For brain density measurements, an area of 0.12 cm2 was employed, and the average CT numbers (Hounsfield Units) of the area were obtained at the following sites: (1) base cut: frontal cortex left side, left internal capsule posterior limb, thalamus, CSF in the third ventricle superior cut (Fig lA); and (2) apical cut (through the level of centrum semiovale): left frontal parietal cortex grey matter, right occipital parietal cortex grey matter, left centrum semiovale (Fig 1B). Exactly the same area was measured for each set of scans, using an XY axis localization given by the machine. It should be borne in mind that the density of each area is inversely proportional to its water content (eg, bone = 1000, CSF = 5 to 10 Hounsfield Units). The following ventricular measurements were also made on the base cuts for all available scans: bicaudate diameter of the frontal horns measuring the maximum distance between the vertical walls at the level of the caudate nuclei, and the widest portion of the third ventricle (Fig 1C). In addition, ten healthy subjects with a similar mean age (37.2 ± 5.3 years) underwent the same CT studies and served as controls. It must be stressed also that those of us who analyzed the EEG (ZJK) and the CT (JDRM) results were completely unaware of what type of dialysis treatment was under study. Biochemistry. Blood was drawn for biochemical and gas analysis before and after a dialysis run. BUN, plasma
465
DIALYSIS DISEQUILIBRIUM SYNDROME
creatinine, calcium, protein, albumin, glucose, sodium, potassium, and chloride were determined by means of routine automated methods. Plasma magnesium was measured using the atomic absorption spectrophotometry (model 303; PerkinElmer, USA) , and plasma osmolality was measured using the freezing-point depression. Blood gases were measured using a Corning 178 gas-analyzer (USA). Statistics. All data are expressed as means ± SEM. Statistical analysis was done using paired and unpaired Student's t tests, when appropriate. Analyses were performed using a basic statistical program on the Amdahl 370 computer.
RESULTS
Before dialysis, there was no significant difference in any of the biochemical data between HD and H HDF. After dialysis, a significant difference in plasma sodium, magnesium, osmolality, arterial PC0 2 , and bicarbonate was noted between the two treatments (Table 2). Brain density was not affected by dialysis, therefore the CT numbers of the seven specific brain structures (base and apical cuts) did not show any change when comparing the pre- with the postdialysis data, either of HD or of H HDF. Furthermore, no significant difference could be noted when comparing the CT numbers at the start and at the end of both treatments (Table 3). Similarly, no significant difference could be demonstrated between the two treatments or within either treatment regarding the transverse diameter of the third ventricle and the bicaudate diameter of the lateral ventricles (Table 3). The group of control healthy subjects showed a density of the same brain structures that was generally higher than that of the uremic population. The difference reached a statistical significance in many of these structures (Table 4). Finally, no significant in-between- and withintreatment (with the exception of one channel in HD) difference was observed when analyzing the EEG% power (3-7/7-13 Hz) data (Table 5). DISCUSSION
Fig 1. (A) Basal cut of a CT scan showing the site of brain density sampling: 1, left frontal cortex; 2, left internal capsule; 3, left thalamus. (B) Apical cut showing the site of density sampling: 1, left parietal cortex; 2, right parietal cortex; 3, left centrum semiovale. (C) Basal cut showing the site of ventricular size measurement: 1, bicaudate diameter of the lateral ventricles; 2, transverse diameter of the third ventricle.
This study was undertaken to evaluate the acute and chronic effects of H HD F and HD on brain hydration and EEG patterns in 11 stable uremic patients. They were studied approximately 6 weeks after they had started on H HDF or HD to allow for stabilization with the assigned method. This was designed to investigate if a difference in cerebral hydration and EEG could be demonstrated due to a chronic effect of H HDF, as compared with HD. Studies were also carried out after dialysis to obtain a profile of acute intradialysis
466
BASILE ET AL Table 2.
Biochemical Data in 11 Patients Before and After a Run of H HDF and a Run of HD Pre
Post
H HDF
BUN (mmoI/L) Plasma creatinine (,.,moI/L) Plasma calcium (mmoI/L) Plasma protein (giL) Plasma albumin (giL) Plasma glucose (mmoI/L) Plasma sodium (mmoI/L) Plasma potassium (mmoI/L) Plasma chloride (mmoI/L) Plasma magnesium (mmoI/L) Plasma osmolality (mmol/kg H2 O) Arterial P0 2 (mm Hg) Arterial PC0 2 (mm Hg) Arterial [H'I3] (nmoI/L) Arterial bicarbonate (mmoI/L)
26.0 1268 2.30 66.9 40.1 4.9 139.3 4.9 103.9 1.13 309.2 97.4 33.7 37.1 20.2
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
HD
2.0 75 0.06 0.8 0.5 0.3 0.4 0.3 1.6 0.04 2.4 3.4 1.5 3.2 0.6
26.8 1154 2.35 68.7 41.9 5.0 138.3 4.5 105.7 1.19 304.7 95.9 32.5 40.9 19.2
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
P*
H HDF
1.6 82 0.12 1.3 1.1 0.3 1.0 0.3 1.6 0.05 1.3 2.8 1.2 0.7 0.8
10.7 584 2.89 76.4 48.0 6.1 139.4 3.0 101.3 1.0 293.4 90.8 35.0 35.1 23.8
± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
1.1 36 0.05 2.5 1.6 0.4 0.9 0.1 0.9 0.02 2.3 4.3 1.7 0.7 0.9
HD
< .05
<.001 <.01 < .02 <.02
10.6 555 2.77 76.1 47.8 5.5 135.5 3.1 100.3 1.12 286.9 90.2 29.5 35.6 19.9
± ± ± ± ± ± ± ± ± ± ± ±
1.2 47 0.07 1.6 1.3 0.3 1.0 0.2 0.9 0.02 1.5 4.1 ± 1.4 ± 0.6 ± 1.0
All values are expressed as means ± SEM. 'Student's paired t test.
changes. To the best of our knowledge, this is the first work that combines CT of the brain and EEG spectral analysis in the study of DDS. H HDF runs were conducted in a shorter (three v four hours) and more efficient way, compared with HD; in fact, BUN and plasma creatinine were not significantly different at the end of the sessions, even if the H HDF treatment time had been reduced by 25 %. Thus, one would predict for H HDF a worse outcome in terms of increased cerebral hydration and EEG, according to the literature,5.6.15 unless the possible protective effect of Table 3.
the hypertonic reinfusion solution is taken into account. 16 Surprisingly, no difference could be observed in the brain density, ventricular size, and EEG% power (3-717-13 Hz) when comparing the predialysis with the postdialysis data, in both H HDF and HD. Furthermore, no difference could be observed when comparing the two treatments, in both the predialysis and the postdialysis setting. Our CT findings are completely different from those previously reported in HD patients 8 : La Greca et al observed a significant decrease in brain
Brain Density and Ventricular Size in 11 Patients Before and After a Run of H HDF and HD H HDF Pre
Brain density (Hounsfield units) Left frontal cortex Left internal capsule Left thalamus CSF in the third ventricle Right occipital parietal cortex (grey matter) Left frontal parietal cortex (grey matter) Left centrum semiovale Ventricular size (cm) Transverse diameter of the third ventricle Bicaudate diameter of the lateral ventricles
37.3 30.8 33.7 7.6
± ± ± ±
HD Post
1.2 1.4 0.9 1.0
37.9 30.8 33.3 5.4
± ± ± ±
1.4 1.3 0.9 1.4
Pre
37.3 31.0 33.3 7.7
± ± ± ±
Post
1.3 1.0 0.7 1.0
37.2 29.4 33.0 6.6
± ± ± ±
1.6 1.0 0.4 1.0
45.2 ± 1.7
44.5 ± 1.8
44.7 ± 1.9
46.5 ± 2.8
43.9 ± 1.8 32.0 ± 1.0
44.8 ± 1.9 30.4 ± 0.9
42.5 ± 1.8 30.0 ± 0.9
42.9 ± 1.6 29.8 ± 1.3
0.40 ± 0.05
0.36 ± 0.06
0.42 ± 0.06
0.39 ± 0.73
1.36 ± 0.10
1.31 ± 0.12
1.39 ± 0.09
1.37 ± 0.12
All values are expressed as mean ± SEM. No statistical difference was found when comparing pre- v post-H HDF, pre- v post-HD, pre-H HDF v pre-HD, post-H HDF v post-HD.
467
DIALYSIS DISEQUILIBRIUM SYNDROME Table 4.
Comparison of Brain Density and Ventricular Size Between 10 Normal Adults (N) and 11 Patients Before a Run of Either H HDF or HD p'
H HDF
Brain density (Hounsfield units) Left frontal cortex Left internal capsule Left thalamus CSF in the third ventricle Right occipital partietal cortex (grey matter) Left frontal parietal cortex (grey matter) Left centrum semiovale Ventricular size (cm) Transverse diameter of the third ventricle Bicaudate diameter of the lateral ventricles
37.3 30.8 33.7 7.6
± 1.2 ± 1.4
p'
N
<.001
± 0.9
± 1.0
43.8 30.5 35.9 9.0
± ± ± ±
1.0 0.8 0.9 0.7
HD
<.001 <.05
48.2 ± 1.2
42.5 ± 1.7 43.9 ± 1.8 32.0 ± 1.0
± ± ± ±
1.3 1.0 0.7 1.0
44.7 ± 1.9
48.6 ± 0.9 34.2 ± 0.4
<.05
37.3 31.0 33.3 7.7
<.01 <.001
42.5 ± 1.8 30.0 ± 0.9
0.40 ± 0.05
0.42 ± 0.05
0.42 ± 0.06
1.36 ± 0.10
1.35 ± 0.17
1.39 ± 0.09
All values are expressed as means ± SEM . • Unpaired Student's t test.
Dialysis patients, however, seem to suffer a chronic subclinical brain swelling. Cerebral edema has not been demonstrated in the brain of patients or laboratory animals with chronic renal failure, either by biochemical or histologic criteria. 2 This preliminary observation, however, seems to be worthy of further investigation. Our patients showed important EEG abnormalities when compared to normal subjects, as already reported, 17.2426 but did not show any EEG deterioration when undergoing either the rapid or the standard dialysis treatment. It must be pointed out that our evaluation of the EEG did not take into account possible effects on visually evoked cortical potentials,24 discriminant analysis,25 or photic driving.23 These parameters may transiently dete-
density during dialysis; furthermore, they claimed that HD induced a cerebral water gain from a predialysis "dry" brain state to a postdialysis normohydrated brain situation. On the basis of this report, it is very hard to explain why DDS symptoms appear at the moment of the correction to normohydration. To the contrary, we could not observe any postdialysis decrease in brain density, either in HD or in H HDF; furthermore, our patients showed a generalized decreased brain density when compared to normal subjects. The reason for this discrepancy with the report of La Greca et al 8 is not clear. Summarizing our CT data, neither a rapid dialysis technique (H HDF) nor the standard HD are able to provoke an acute change in brain density and ventricular size. Table 5.
EEG, % Power (3-717-13 Hz) in 11 Patients Before and After a Run of H HDF and HD H HDF
HD Post
Pre
Pre
Post
Left hemisphere FTTJ Ts-TJ PJ-Ts FTP J
1.48 0.67 0.75 0.99
± ± ± ±
0.26 0.14 0.16 0.20
1.27 0.76 0.77 1.00
± ± ± ±
0.21 0.14 0.14 0.22
1.49 0.84 0.91 1.02
± ± ± ±
0.27 • 0.23 0.25 0.29
1.17 1.00 0.98 1.13
± ± ± ±
0.17 0.24 0.22 0.26
Right hemisphere Fa-T 4 T6-T 4 P4 -T 6 Fa-P 4
1.05 0.51 0.61 0.79
± ± ± ±
0.13 0.11 0.16 0.16
0.99 0.63 0.66 0.85
± ± ± ±
0.17 0.11 0.14 0.18
1.54 0.59 0.73 1.00
± ± ± ±
0.46 0.19 0.27 0.31
1.36 0.73 0.82 0.96
± ± ± ±
0.24 0.17 0.24 0.21
All values are expressed as means ± SEM. No statistical difference was observed when comparing pre- v post-H HDF, pre-H HDF v pre-HD, post-H HDF v post-HD. 'Only one significant difference (P < .02) was found when comparing pre- v post-HD; paired Student's t test.
468
BASILE ET AL
riorate during dialysis. Lewis et al,27 however, have been able to show that evoked potential latencies tend to improve in the first and in the following hours after dialysis. Our EEG data are in agreement with those of Kiley et al in uremic patients 28 and those of Arieff et al, 7 who were unable to observe any significant alteration in the EEG spectral analysis in uremic dogs during and after rapid HD. To the contrary, many studies have shown an EEG deterioration during dialysis.1I14.16 As it has already been pointed out, 7 though, not all the patients in these reports suffered an EEG deterioration during dialysis12.13.16; furthermore, the assessment of these reports shows that in no instance was a systematic examination of the EEG performed, and no standardized criteria were used for evaluating or quantitating deterioration of the EEG. In conclusion, this study shows neither a post-
dialysis change in brain density and ventricular size nor a postdialysis EEG deterioration in a group of stable uremic patients undergoing both a rapid and a standard dialysis treatment. Thus, DDS must be considered a peculiarity of rapid HD in patients, particularly children, affected by acute renal failure or at the start of their program of mai!ltenance HD. 3 Most of the uremic stable patients on maintenance HD are highly unlikely to develop DDS under any circumstances. 29 Furthermore, the diagnosis of DDS has become recently a "wastebasket" for a number of disorders that can occur in patients with renal failure. 30 ACKNOWLEDGMENT The authors wish to thank the personnel of the Dialysis Unit at the University of Alberta Hospitals for their assistance and cooperation, C. Ness for her aid in preparing the manuscript, and S. Breitkreuz and 1. Buicliu for their skillful technical assistance.
REFERENCES I. Wakim KG: The pathophysiology of the dialysis disequilibrium syndrome. Mayo Clin Proc 44:406-429, 1969 2. Arieff AI: Neurological complications of uremia, in Brenner BM, Rector FC (eds): The Kidney. Philadelphia, Saunders, 1981, pp 2306-2343 3. Arieff AI: Dialysis disequilibrium syndrome, in Massry SG, Glassock RJ (eds): Textbook of Nephrology. Baltimore, Williams & Wilkins, 1983, pp 24-26 4. Rosomoff HL, Zugibe FT: Distribution of intracranial contents in experimental edema. Arch Neurol 9:26-34, 1963 5. Pappius HM, Oh JH, Dossetor 18: The effects of rapid hemodialysis on brain tissues and cerebrospinal fluid of dogs. Can J Physiol Pharmacol 45: 129-147, 1967 6. Arieff AI, Guisado R, Massry SG, et al: Central nervous system pH in uremia and the effects of hemodialysis. J Clin Invest 58:306-311, 1976 7. Arieff AI, Lazarowitz VC, Guisado R: Experimental dialysis disequilibrium syndrome: Prevention with glycerol. Kidney Int 14:270-278, 1978 8. La Greca G, Biasioli S, Chiaramonte S, et al: Studies on brain density in hemodialysis and peritoneal dialysis. Nephron 31: 146-150, 1982 9. Krane EJ, Rockoff MA, Wallman JK, et al: Subclinical brain swelling in children during treatment of diabetic ketoacidosis. N Engl J Med 312:1147-1151, 1985 10. Villafana T: Physics and instrumentation, in Lee SH, Rao KCVG (eds): Cranial Computed Tomography. New York, McGraw-Hill, 1983, pp 1-46 II. Kennedy AC, Linton AL, Luke RG, et al: Electroencephalographic changes during hemodialysis. Lancet 1:408411, 1963 12. Hampers CL, Doak PB, Callaghan MN, et al: The electroencephalogram and spinal fluid during hemodialysis. Arch Intern Med 118:340-346, 1966 13. Jacob JC, Gloor P, Elwan OH, et al: Electroencephalo-
graphic changes in chronic renal failure. Neurology 15:419429, 1965 14. Hampl H, Klopp HW, Michels N, et al: Electroencephalogram investigations of the disequilibrium syndrome during bicarbonate and acetate dialysis. Proc Eur Dial Transplant Assoc 19:351-359, 1982 15. Arieff AI, Massry SG, Barrientos A, et al: Brain water and electrolyte metabolism in uremia: Effects of slow and rapid hemodialysis. Kidney Int 4:177-187, 1973 16. Port FK, Johnson WJ, Klass DW: Prevention of dialysis disequilibrium syndrome by use of high sodium concentration in the dialysate. Kidney Int 3:327-333, 1973 17. Basile C, Bornstein RA, Elleker MG, et al: The chronic efficacy and safety of hypertonic hemodiafiltration: A randomized cross-over study (submitted) 18. Basile C, Di Maggio A, Curino E, et al: Pharmacokinetics of netilmicin in hypertonic hemodiafiltration and standard hemodialysis. Clin Nephrol 24:305-309, 1985 19. Basile C, Di Maggio A, Manente G, et al: Solute kinetics in hypertonic hemodiafiltration and standard hemodialysis. Am J Kidney Dis 7:483-489, 1986 20. Basile C, Di Maggio A , Longo S, et al: Sodium balance in hypertonic hemodiafiltration. Blood Purif 2:70-75, 1984 21. Basile C, Di Maggio A, Scatizzi A: Long-term experience with hypertonic hemodiafiltration. Proc Eur Dial Transplant Assoc 22:303-307, 1985 22. Kirkwood RG, O'Brien M, Butruille Y, et al: Initial clinical evaluation of the Hospal artificial kidney system. Trans Am Soc Artif Intern Organs 24:509-515, 1978 23. International Federation of Societies for Electroencephalography and Clinical Neurophysiology: The ten-twenty electrode system. Am J EEG Technoll:13-19, 1961 24. Teschan PE, Ginn HE, Bourne JR, et al: Quantitative indices of clinical uremia. Kidney Int 15:676-697, 1979 25. Teschan PE, Bourne JR, Reed RB, et al: Electrophys-
DIALYSIS DISEQUILIBRIUM SYNDROME
iological and neurobehavioral responses to therapy: the National Cooperative Dialysis Study. Kidney Int 23:S58-S65 , 1983 (suppl 13) 26. Bourne JR, Teschan PE: Computer methods, uremic encephalopathy and adequacy of dialysis. Kidney Int 24:496-506, 1983 27. Lewis EG, O'Neill WM, Dustman RE, et al: Temporal effects of hemodialysis on measures of neural efficiency. Kidney Int 17:357-363, 1980
469 28. Kiley JE, Woodruff MW, Pratt KL: Evaluation of encephalopathy by EEG frequency analysis in chronic dialysis patients. Clin Nephrol 5:245-250, 1976 29. Arieff AI, Massry SG: Dialysis disequilibrium syndrome, in Massry SG, Sellers AL (eds): Clinical Aspects of Uremia and Dialysis. Springfield, IL, Thomas, 1976, pp 34-52 30. Mahoney CA, Arieff AI: Uremic encephalopathies: Clinical, biochemical, and experimental features. Am J Kidney Dis 2:324-336, 1982