The dialysis disequilibrium syndrome (DDS) is characterized by the neurologic turbulence and cerebral edema which occur after hemodialysis and which influence the cardiac and pulmonary systems by inducing changes in the chemicophysical characteristics of the body fluids, such as the pH, temperature, electrolytes composition, and osmolality.1,2 The clinical symptoms of DDS include headache, muscular spasm, nausea, vomiting, brain swelling, seizure, coma, and even death. DDS is actually caused by the disequilibrium of solutes which can result in a subsequent abnormal water distribution during hemodialysis. The mechanism of the disequilibrium of solutes is mainly attributed to differences in the speed of their removal from the tissue and blood. The slow removal of urea from the tissues during hemodialysis establishes a tissue to plasma gradient which promotes water movement from the blood to the tissue and cells, which increases the amount of interstitial and intracellular fluid. Therefore, the abnormal water distribution of DDS is due to an osmotic pressure change.3–5 It has been observed in clinical practice that some of the uremic patients, without any evidence of cerebral or pulmonary edema before their initial hemodialysis, develop acute brain and lung edema only a few hours after hemodialysis.
In this study, the changes in the pulmonary hemodynamics and water content were evaluated while DDS occurred in uremic dogs during hemodialysis. We herein report our findings which indicate that acute pulmonary edema, accompanied by acute brain edema, can be induced by hemodialysis.
Establishment of an acute renal failure (ARF) model in dogs
Four groups of dogs weighing (18±2) kg and aged 18–26 months were studied: Group A, normal control, 12 non-uremic dogs that did not receive dialysis; Group B, uremic nondialyzed control, 12 dogs did not receive dialysis; Group C, hemodialysis, 12 uremic dogs dialyzed for 2 hours 72 hours after bilateral ureteral ligation; Group D, sham hemodialysis control, 6 uremic dogs dialyzed without dialysate for 2 hours 72 hours after the operation. The dogs in groups B, C, and D underwent bilateral ureteral ligation under ketamine anesthesia (5 mg/kg, intravenous injection), and were administrated gentamycin (6000 U/kg) to prevent infection. All of the animals were fed normal food and water after the operation.
Dialysate solution contained Na+140 mmol/L, K+2.8 mmol/L, Cl− 117 mmol/L, Ca2+ 1.75 mmol/L, Mg2+ 1.09 mmol/L, HCO3− 33.1 mmol/L, pH 7.41, osmotic pressure 280 mOsm/kgH2O. Dialysate quality was consistent with the standard of American Association of Medical Instrument (bacteria <200 CFU/ml,endotoxin <2 EU/ml). Before the initiation of dialysis, the dogs were weighed and anesthetized with ketamine (5 mg/kg, intravenous infection). The vascular access was established using a three-lumen catheter inserted through the external jugular vein. A hemodialysis was performed with a blood flow of 150 ml/min, dialysate flow of 300 ml/min for 2 hours, but without ultrafiltration or normal saline to prefill the blood tube. Heparin was used as an anticoagulant during the hemodialysis. A bolus of 1 mg/kg heparin was injected into the blood tube before the dialysis followed by a maintenance dosage (8–12 mg/h). A Nikkiso DBB-22 dialysis machine, blood tube, and FB-70 dialyzer (surface area 0.7 m2, ultrafiltrate coefficient 4.3 ml·mmHg−1·h−1, low flux cellulose diacetic acid membrane, B-Braun Ltd.) were employed in this study. The total volume for the extracorporeal circulation was 45 ml. The sham dialysis was performed under the same conditions mentioned above, but without the dialysate.
Sampling and analysis
The blood samples in group A were drawn from the dogs immediately, while the samples in group B were taken at 72 hours after the operation; all samples were collected from the femoral vein. The samples in group C and D were collected from the external jugular vein 72 hours after the operation (before dialysis), at 1 and 2 hours after the onset of dialysis, as well as 1 hour after the completion of dialysis. The serum levels of urea, creatinine, and Na+ were determined by a biochemical analysis instrument (DT602, American). A blood gas analysis was also performed. The osmolality was monitored by an osmotic pressure analysis machine (Nikkiso, Japan).
Determination of intracranial pressure
Under the ketamine anesthesia, an air sensor connected to a computerized apparatus was surgically implanted under the dura mater in all the dogs. The intracranial pressure was measured in all groups, and it was continuously monitored in groups C and D during the dialysis session. Spinal fluid was collected at the end of dialysis by a lumbar spinal tap.
Determination of pulmonary vascular pressure
A Swan-Ganz catheter was utilized to monitor the right atrial pressure (RAP), right ventricular pressure (RVP), pulmonary artery pressure (PAP), cardiac output (CO), pulmonary capillary wedge pressure (PCWP), and central venous pressure (CVP) of the dogs. The cardiac index (CI, CI = CO/BSA, BSA (m2) = K×(W2/3/10000)), and the total pulmonary resistance index (TPRI = MPAP/CI) were calculated. The blood pressure and heart rate were monitored continuously throughout the dialysis session.
Chest X-ray imaging
Chest X-ray studies were taken for 2 dogs in group C pre- and post-dialysis under the same exposure parameters.
Evaluation of pulmonary and brain water content
The dogs in groups C and D were sacrificed after the 2 hour dialysis, and their lungs were weighed to calculate the pulmonary edema index (lung weight (g)/body weight (kg)). The lung and brain tissue (10 g) were cut into pieces and baked at 120°C for 2 hours to calculate the dry/wet ratio (dry weight of tissue (g)/wet tissue weight (g)). The dogs in groups A and B were also sacrificed and the tissue studies were conducted as in groups C and D.
Pathological findings of tissue
A pathological examination of the lung and brain tissue was separately performed on two dogs from group C.
The data analysis was performed using the statistical software package SPSS15. The data were expressed as the mean ± standard error (SE). The differences among the groups were analyzed by one-way analysis of variance (ANOVA). Comparisons within a group were performed with a t test. The statistical differences were considered significant if the P value less than 0.05.
Establishment of acute renal failure in the dogs and the dialysis
The dogs in groups B, C, and D had developed severe azotemia, acidosis, and electrolyte disorder by 72 hours after the bilateral ureteral ligation. The dogs in group C received hemodialysis for 2 hours, but 2 of them died from the high intracranial pressure (the maximum value reached 25 mmHg) and were not included in our statistical analysis. The other 10 dogs completed hemodialysis. The dogs in group D were treated with a sham hemodialysis for 2 hours. In the uremic animals, the pre-dialysis urea, creatinine, and osmolality were equivalent (Table 1). After the hemodialysis, no variation in the weights of the dogs in groups C and D was found (without ultrafiltration), but the plasma urea and creatinine in group C were reduced by 73.6% and 60.1%, respectively (P <0.001). The osmolality in group C decreased from (359±18) mOsm/kgH2O to (304±6) mOsm/kgH2O, a decrease of (55.0±14.3) mOsm/kgH2O, which was a significant difference (P <0.001). In group D, no differences were found in these parameters in the pre- and post-dialysis testing.
Clinical features of DDS induced by hemodialysis
There was no significant difference in the intracranial pressure between groups B, C and D before dialysis (P >0.05). The intracranial pressure in group C increased from (3.00±0.46) mmHg to (3.94±1.15) mmHg (P=0.0001) at 1.5 hours after the initiation of hemodialysis, and the dogs developed neuropathic manifestations such as dysphoria, dyspnea, striving, roaring, etc. The blood pressures increased from (18.8±1.5)/(10.9±1.0) kPa to (20.3±1.0)/(12.3±1.4) kPa (P <0.05). The heart rates of the pre- and post-dialysis assessment were (129±12) beats/min and (152±11) beats/min, respectively (P <0.05). In group D, there were no abnormal clinical manifestations and no changes of the intracranial pressure, blood pressure, and heart rate after the hemodialysis.
The laboratory evidence of brain edema induced by hemodialysis intracranial pressure increased
The pre-dialysis intracranial pressures in groups B, C and D were significantly higher than those in group A ((3.14±1.46) mmHg, (3.00±0.46) mmHg and (3.45±0.38) mmHg vs (2.57±1.43) mmHg, P <0.05 respectively). The post-dialysis intracranial pressures in group C showed a significant increase ((3.95±1.14) mmHg). But the pre- and post-dialysis intracranial pressures in group D showed no difference (pre-dialysis (3.45±0.38) mmHg vs post-dialysis (3.58±0.33) mmHg, P >0.05).
Brain water content overloaded
After dialysis, the dry/wet ratio of the brain tissue of the dogs in group C was reduced to 0.24 0.02, which was significantly different from that in groups A, B and D (0.36±0.03, 0.34±0.02 and 0.35±0.04, P <0.05). These results indicated that hemodialysis induced an increase in the water content in the brain tissue (Table 2).
Changes of cerebrospinal fluid parameters after dialysis
The ratios of the urea level and the osmotic pressure between the spinal fluid and the plasma increased. But there were no significant changes observed in group D (Table 2). The present results indicated the increases of the intracranial solutes and the brain water content. The decline of the cerebrospinal fluid/plasma HCO3− ratio indicated a much more acidic status in the cerebrospinal fluid than in the blood.
Pathological findings of brain tissue
The post-dialysis brain histological investigation of the dogs in group C showed swollen neurons and glial cells with lightly dyed cytoplasm, engorged blood vessels, and widened vascular and cellular crevices (Figure 1). The dogs in group D showed no significant changes in the brain histological examinations after the sham dialysis.
The laboratory evidence that pulmonary edema was induced by hemodialysis
The indices of the pulmonary vascular resistance of the dogs were monitored in the four groups. As shown in Table 3, no significant differences were found in the hemodynamic parameters among the four groups before dialysis. However, during dialysis the PAP, PCWP, RAP, RVP, CVP, and TPRI tended to increase in group C. These values peaked at 1 hour and then were maintained until the end of the hemodialysis, and were significantly different from those of pre-dialysis monitoring (P <0.05). The parameters listed above did not changed in group D after the dialysis.
Assessment of pulmonary water content
After the dialysis the dry/wet ratio of the lung in group C was reduced to 0.24±0.02, which was significantly different from the three control groups (P <0.05). These results indicated that hemodialysis induced an increase in the water content of the lung tissue. To further evaluate the lung water content, we also measured the index of pulmonary edema. There was no significant difference between groups A, B, C, and D in the pre-dialysis assessment of pulmonary edema. But the index of pulmonary edema increased to 9.58±0.53 in group C after the dialysis (P <0.001), and did not change in group D after dialysis (Table 4). These data clearly demonstrated an increase of the lung water content in the dialyzed dogs.
Results of chest radiography
Chest X-ray studies were performed on 2 dogs in group C before and after dialysis with the same exposure parameters. As is shown in Figure 2, a mild pulmonary congestion was demonstrated in the images after the dialysis.
Pathological findings of lung tissue
The pathological changes of the lung tissue after the hemodialysis were assessed in 2 dogs in group C. The gross inspection revealed lung congestion and volume contraction of the lung tissue. On the surface of the lung tissue, there were some red-white alternative plaques. A mixture of tissue liquid and blood flowed out from the lung section (Figure 3). The microscopic histological examination showed the interstitial congestion and an increased space between the alveoli. There were also scattered pale reddish stains of blood and edema fluid in some of the alveoli (Figure 4).
Kennedy et al6 initially reported the dialysis disequilibrium syndrome in 1962, which occurred as a result of the rapid change of the plasma urea concentration and the slow removal of urea from brain during hemodialysis, which led to the formation of an osmotic gradient between the blood and the brain. The higher concentration of solutes in the brain, relative to the blood, promoted the movement of water from the blood to the brain, and was termed “the urea reverse affect”. In this study, the DDS in the uremic dogs was successfully duplicated; all of the biochemical parameters reflecting the renal function fulfilled the indications for hemodialysis. After a 2 hour dialysis, the plasma urea and creatinine significantly decreased by 73.26% and 60.10% respectively, and the osmolality declined by (55.0±14.3) mOsm/kgH2O. The intracranial pressure gradually increased during hemodialysis, and simultaneously the signs and symptoms related to high intracranial pressure developed, such as dysphoria, hypertension, dyspnea, roaring, and striving etc. These clinical symptoms occurred mainly 1.5 hours after the initiation of hemodialysis.
The data in Table 2 indicate that after the dialysis the intracranial pressure in group C increased from (3.00±0.46) mmHg to (3.95±1.15) mmHg, a ratio of 1.3; the ratios of the cerebrospinal fluid/plasma urea and the dry/wet brain tissue were 1.7 and 0.24, respectively. All of the evidence mentioned above is indicative of an increased water content and solute concentration in the brain tissue. The pathologic examination of the brain revealed a pale stain for the neurons and glial cells, and enlarged spaces among the cells with vascular congestion.
La Greca et al7 reported that the images of the brain by CT evaluation showed that the density of the brain tissue was significantly reduced during and after hemodialysis, which suggested an increased water content of the brain after hemodialysis. Silver et al8 reported that, after 90 min hemodialysis in 9 uremic rats, the plasma urea was reduced by 52.7%, whereas the urea level of the brain tissue was only slightly reduced, by about 13%, and the decline of the plasma osmotic pressure was 8%. In comparision to 15 uremic rats without hemodialysis, the dialyzed rats showed an increase in the water content in the brain tissue of 6% (P <0.01), and a 1.6 times higher urea level in the brain than in the plasma. All the data mentioned above strongly suggested that the increased water content in the brain tissue might be due to the higher level of urea in the brain in the dialysed animal models.
The study of Trinh-Trang-Tan et al9 showed that the molecular basis for the dialysis disequilibrium syndrome was the conjunction of a reduced expression of urea transporter UT-B1 and an increased expression of aquaporin (AQP4 and AQP9) in the brain cells. Because of the low UT-B1 abundance, the urea exit from astrocytes was most probably delayed during the rapid removal of the extracellular urea through hemodialysis. This created an osmotic driving force that promoted water entry into the cells (favored by the abundant AQPs) and subsequent brain swelling. The study of urea kinetics demonstrated that the diffusion capacity of urea in brain tissue was 1.5×10−6 cm/s, considerably slower than water (4.5×10−3 cm/s).10,11 This difference between the diffusion capacity of urea and water easily resulted in an osmotic gradient between the brain and the blood, and consequently, the development of osmotic cerebral edema. It was reported that, every 10 mmol/L change of the plasma urea concentration might induce a 0.03 L fluid exchange between the cell and the interstitial space.
The muscle interstitial urea concentrations during hemodialysis were determined with a microdialysis technique in the pilot study.12 The results demonstrated that the muscle interstitial urea concentration was 19% higher than the plasma urea concentration after 17 minutes, 29% higher after 53 minutes, and 40% higher after 117 minutes, but no difference was observed between them before the hemodialysis. The study by Shirai et al13 demonstrated that the concentrations of urea, creatinine, and uric acid in the plasma were 15%-17% lower than in the edema fluid after hemodialysis, due to the delay of the removal of these solutes from the edema fluid.
All the evidences including the morphology, molecular basis, and urea kinetics suggested that the main pathogenesis of DDS was the “urea reverse effect”, which led to the formation of an osmotic gradient between the blood and interstitial fluid during hemodialysis, and made water move into the tissue interstitium, lung, brain and cells, and resulted in the disequilibrium of multiple systems in the body.
Pulmonary edema is a pathological process in which the formation of fluid is more than the reflux in the lung interstitium; thereby more fluid is accumulated in the pulmonary interstitium and overflows into the alveoli.14 In this study, we observed that the pulmonary vascular pressure parameters gradually increased such as the PAP, PCWP, RAP, RVP, TPRI, and CVP, while the DDS developed during hemodialysis without ultrafiltration and weight change. The results indicated an increased pulmonary vascular resistance and pulmonary congestion. The increased pulmonary water content in the dialyzed dogs was demonstrated by the X-ray sign of pulmonary congestion, increased pulmonary edema index, and reduced dry/wet ratio of the lung tissue. The pathological results showed a decreased lung volume, patchy of pulmonary congestion on the surface and a mixture of bloody fluid that leaked out from the section of lung tissue. The histological examination displayed a certain degree of vessel congestion, an increased interstitial space of the alveoli, and a scattering of bloody fluid in some of the alveoli. The lack of histological control of uremic dogs should be improved in this experiment. The pathogenesis of the increased pulmonary water content was identical to that of the brain; that is the “the urea reverse effect”.
In the early years of the clinical practice of hemodialysis, patients often developed disequilibrium syndrome. Meanwhile, some patients were found to develop acute pulmonary edema within a few hours after hemodialysis. DiFresco et al15 reported the case of a patient who developed a typical disequilibrium syndrome accompanied with respiratory failure during dialysis. They presumed that the respiratory failure was caused by cerebral edema which was confirmed by CT. Zeng et al16 reported 4 cases of uremic patients who developed “adult respiratory distress syndrome” during dialysis sessions. The clinical features of adult respiratory distress syndrome in those patients were as follows: dyspnea, a low partial pressure of oxygen which could not be corrected by routine oxygen supplementation, and the diffuse infiltrative shadows in both lungs on the chest X-ray. The symptoms did not respond to treatment with cardiotonic drugs and they developed immediately after dialysis. The authors considered that this syndrome had clinical features identical to acute pulmonary edema related to dialysis. This adult respiratory distress syndrome was regarded as a special type of disequilibrium syndrome, which shared the same pathophysiological basis with the cerebral edema induced by hemodialysis. According to the idea of these two authors, the dialysis-induced pulmonary complication was related to DDS, and their reports and analysis were consistent with ours.
It was said that the cerebrospinal fluid volume might account for 10% of the intracranial space. In general, a greater than 5% increase of the cerebrospinal fluid volume could induce an increase of the intracranial pressure.17,18 The lung is much more able to endure fluid volume expansion than the brain, and only when the increase of the pulmonary water content reaches up to 30% will an X-ray investigation demonstrate the signs of pulmonary edema. With the significant degree of cardiac compensation and the pulmonary lymphatic reflux (about 200 ml/h), pulmonary edema could hardly be seen. Therefore the symptoms of cerebral edema were easily found and treated in time; however, pulmonary edema might also easily develop in the patients with poor cardiac function, hypoproteinemia, hypoxemia, and severe acidosis.
There were no changes in the plasma level of urea, plasma osmotic pressure, intracranial pressure, pulmonary vascular resistance, or the dry/wet ratios in the brain and lung tissues in the sham dialyzed dogs of group D. The osmotic pressure and blood gas index in the cerebrospinal fluid and plasma were maintained at a steady level during dialysis. The pathological changes of edema of the lung and brain tissues were not observed in the sham dialyzed dogs. Therefore, the present study provided further testimony that the edema of the brain and lung was mainly due to the “urea reverse effect” and was directly associated with hemodialysis. On the other hand, based on the normal histological findings of the lung and brain after the sham dialysis, it could be excluded that the pulmonary edema induced by hemodialysis was caused by complement activation, acute pulmonary hypertension, neurogenic pulmonary edema, or water intoxication. Our experimental result proved that the “urea reverse effect” could result in acute cerebral and pulmonary edema at the same time in the dog model during hemodialysis; pulmonary edema was only a partial manifestation of the abnormal body fluid distribution.
In conclusion, the pathogenesis of DDS has mainly included the “urea reverse effect” and the “brain intracellular acidosis” theories.19,20 The present study indicated that the “urea reverse effect” might induce the DDS during dialysis. The characteristic was cerebral edema and the corresponding clinical manifestations. According to the same pathogenesis, the plasma water transferred across the barrier among the pulmonary capillary vessels, interstitium, and alveoli, and promoted the formation of interstitial and alveolar edema which could be identified by the increase in the pulmonary vascular resistance, the pulmonary edema index, the decrease in the dry/wet ratio of the lung tissue, and by X-ray imaging and the pathological findings of the lung.
1. Sheth KN, Wu GF, Messé SR, Wolf RL, Kasner SE. Dialysis disequilibrium: another reversible posterior leukoencephalopathy syndrome? Clin Neurol Neurosurg 2003; 105: 249–252.
2. Buemi M, Aloisi E, Coppolino G, Loddo S, Crascì E, Aloisi C, et al. The effect of two different protocols of potassium haemodiafiltration on QT dispersion. Nephrol Dial Transplant 2005; 20: 1148–1154.
3. Silver SM, Sterns RH, Halperin ML. Brain swelling after dialysis: old urea or new osmoles? Am J Kidney Dis 1996; 28: 1–13.
4. Tzamaloukas AH, Agaba EI. Neurological manifestations of uraemia and chronic dialysis. Niger J Med 2004; 13: 98–105.
5. Kimura G, Van Stone JC, Bauer JH, Keshaviah PR. A simulation study on transcellular fluid shifts induced by hemodialysis. Kidney Int 1983; 24: 542–548.
6. Kennedy AC, Linton AL, Eaton JC. Urea in cerebrospinal fluid after hemodialysis. Lancet 1962; 1: 410–411.
7. La Greca G, Biasioli S, Chiaramonte S, Dettori P, Fabris A, Feriani M, et al. Studies on brain density in hemodialysis and peritoneal dialysis. Nephron 1982; 31: 146–150.
8. Silver SM, DeSimone JA Jr, Smith DA, Sterns RH. Dialysis disequilibrium syndrome
in the rat:Role of the “reverse urea effect”. Kidney Int 1992; 42: 161–166.
9. Trinh-Trang-Tan MM, Cartron JP, Bankir L. Molecular basis for the dialysis disequilibrium syndrome
: altered aquaporin and urea transporter expression in the brain. Nephrol Dial Transplant 2005; 20: 1984–1988.
10. Fenstermacher JD. Flow of water and solutes across the blood-barrier. In: Dacey RG, eds. Trauma of the Central Neruous System. 1st ed. New York: Raven Press; 1985: 123–140.
11. Fenstermacher JD. Volume regulation of the central nervous system. In: Staub NC, Taylor AE, eds. Edma, 2nd ed. New York: Raven Press; 1984: 81–93.
12. Alquist M, Thysell H, Ungerstedt U, Hegbrant J. Development of a urea concentration gradient between muscle interstiium and plasma during hemodialysis. Int J Artif Organs 1999; 22: 811–815.
13. Shirai D, Kajimoto Y, Ochi S, Tanaka Y, Wada A, Yano F, et al. Evaluation of hemodialysis efficiency by an edema fluid model. Ren Fail 1992; 14: 63–68.
14. Guo Y. Animal models of human's diseases. 2nd ed. Beijing: Renmin Public Health Publication Company; 1995: 163–165.
15. DiFresco V, Landman M, Jaber BL, White AC. Dialysis disequilibrium syndrome
: an unusual cause of respiratory failure in the medical intensive care unit. Intensive Care Med 2000; 26: 628–630.
16. Zeng HB, Sun SL. The Exploration of mechanism and prevention of ARDS after hemodialysis. J Intern Emerg Severe Dis 1999; 1: 6–7.
17. Chen CL, Lai PH, Chou KJ, Lee PT, Chung HM, Fang HC. A preliminary report of brain edema in patients with uremia at first hemodialysis: evaluation by diffusion-weighted MR imaging. Am J Neuroradiol 2007; 28: 68–71.
18. Galons JP, Trouard T, Gmitro AF, Lien YH. Hemodialysis increases apparent diffusion coefficient of brain water in nephrectomized rats measured by isotropic diffusion-weighted magnetic resonance imaging. J Clin Invest 1996; 98: 750–755.
19. Bagshaw SM, Peets AD, Hameed M, Boiteau PJ, Laupland KB, Doig CJ. Dialysis disequilibrium syndrome
: brain death following hemodialysis for metabolic acidosis and acute renal failure — a case report. BMC Nephrol 2004; 5: 9.
20. Silver SM. Cerebral edema after rapid dialysis is not caused by an increase in brain organic osmolytes. J Am Soc Nephrol 1995; 6: 1600–1606.