QJM Advance Access originally published online on July 29, 2005
QJM 2005 98(9):691-703; doi:10.1093/qjmed/hci101
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Masterclasses in medicine |
Acute and fatal hyponatraemia after resection of a craniopharyngioma: a preventable tragedy
From the 1Department of Critical Care Medicine, Hospital for Sick Children and Departments of Anaesthesia and Medicine, University of Toronto, Toronto, Canada, 2Nephrology Unit and Department of Internal Medicine, Stellenbosch University, Cape Town, South Africa, and 3Division of Nephrology, St Michael's Hospital, University of Toronto, Toronto, Canada
Address correspondence to Professor M.L. Halperin, University of Toronto, St Michael's Hospital Annex, Lab #1, Research Wing, 38 Shuter Street, Toronto, Ontario, M5B 1A6, Canada. email: mitchell.halperin{at}utoronto.ca
 |
Summary |
|---|
Top
Summary
IntroductionCentral diabetes insipidus developed for the first time in a 14-year-old female during the resection of a craniopharyngioma. The water diuresis persisted until a vasopressin analogue (dDAVP) was given. Professor McCance was asked to explain why hypernatraemia developed, to anticipate dangers that might develop in the salt and water area with therapy, and to provide insights into why this patient died, due to the subsequent development of hyponatraemia that caused a lethal rise in intracranial pressure. The team specifically wanted Professor McCance's opinions as to why a PNa of 124 mmol/l was uniquely dangerous for this patient, and this was a particularly challenging conundrum. Nevertheless, with the aid of a mini-experiment, a careful chart review, and creative thinking, he was able to offer a novel solution, and to suggest ways to prevent its occurrence in other patients.
|
Introduction |
|---|
In this teaching exercise, the central (imaginary) figure is Professor McCance, a consultant who practiced medicine
70 years ago. His overall objective is to demonstrate that applying principles of integrative physiology at the bedside can be extremely helpful in revealing the pathophysiology of disease, making more accurate clinical diagnoses, and planning optimal therapy. |
The consultation |
|---|
The medical team was apprehensive, because the patient they had tragically lost was to be discussed at the weekly morbidity and mortality meeting. While they had been quite confident about her diagnosis and treatment, the very unfortunate outcome had naturally introduced uncertainties. Although discussions with the nephrology, endocrine, and intensive care services (who had all been involved in the management) had not uncovered any obvious errors, the team wished to ask Professor McCance to provide his analysis of the events. If anyone could come up with novel insights, they had faith that it would be their mentor. Never one to shy away from a challenge, or an opportunity to teach, Professor McCance eagerly agreed to participate.
The patient was a 14-year-old, 40 kg female who appeared to be in excellent health until recently, when headaches became a problem. A small but progressive visual field defect was found on examination. Further studies revealed the presence of a craniopharyngioma, and neurosurgery was planned. Of special importance, she did not have evidence of central diabetes insipidus (DI), including thirst, polydipsia, or polyuria, before surgery.
During anaesthesia, she received 3 l of isotonic saline to prevent a fall in blood pressure, probably because anaesthetic agents lower the tone of smooth muscles in venous capacitance vessels, and hence a larger venous volume is often needed to ensure adequate central venous pressure.1 The volume and composition of this infusion did not differ from that given to other patients undergoing this surgical procedure.2
In the first 4 h after general anaesthesia, her urine output rose dramatically to 10 ml/min. Urine and blood were sent promptly to the laboratory to measure the urine osmolality (Uosm) and the plasma Na+ concentration (PNa), to confirm that the diagnosis was central DI. With this information, the plan was to administer the vasopressin analogue desamino, D-arginine vasopressin (dDAVP) to curtail the water diuresis. Not wishing to infuse too much sodium (Na+) and chloride (Cl–) before receiving the laboratory results, the rate of infusion of isotonic saline was decreased until the diagnosis was confirmed. Nevertheless, another 4 h passed before dDAVP was administered. At this time, a second blood and urine sample was sent to the laboratory. Its PNa was even higher, and there was no change in the urine electrolyte concentrations or the Uosm. The values for the PNa, the urine electrolytes, and the urine volumes are summarized in Figure 1.
|
|
The ‘post mortem’ with Professor McCance |
|---|
The team was confident about their diagnosis of central DI, because of polyuria (urine flow rate 10 ml/min) with a Uosm of 120 mOsm/kg H2O (a water diuresis) in the presence of a physiological stimulus for the release of vasopressin, (PNa 153 mmol/l). Because of the low urine Na+ plus potassium (K+) concentration (UNa + K) (50 mmol/l), the basis for the acute rise in the PNa should be this large water diuresis. Professor McCance raised his hand to issue a caution with this line of reasoning. In his familiar way, he began by posing a question, and as always he wished to have a response in quantitative terms. ‘Examining the data in Figure 1, what is the basis for the initial rise in PNa to 153 mmol/l?’
Question 1. Why did the PNa rise initially to 153 mmol/l in this patient?
Physiology principle 1. Concentration terms have numerators and denominators. Therefore the rise in PNa could be due to a rise in the content of Na+ or a decrease in the volume of water in the extracellular fluid (ECF) compartment (Figure 2).
|
Return to the consultation. The nephrology consultant was quick off the mark, and wished to illustrate how modern specialists determine why the PNa might rise. Balances rather than the output must be examined. Moreover, ‘I always perform an electrolyte-free water (EFW) balance to make this decision,’ he said.3,4 To facilitate his exposition, he illustrated how to convert the composition of the urine in this patient to EFW terms (Figure 3). For example, 3 l of urine with a UNa + K of 50 mmol/l can be thought of as a mixture of 1 l of isotonic saline (containing 150 mmol Na+ + K+/kg of H2O) with 2 l of EFW. Since isotonic saline was infused, there was no input of EFW. This together with an EFW content of the urine of 2 l, means that there was a deficit of 2 l of EFW. The observed rise in the PNa of 13 mmol/l was due to this 2 l deficit of EFW (equation 1, assuming her total body water is 24 l). ‘The accuracy of the calculation in equation 1 can be increased if the PNa values were reported in molal (mmol/kg H2O) rather than molar terms (mmol/l of plasma),’ he stated. Therefore, proper therapy would be to administer 2 l of EFW.
 | (1) |
|
Question 2. Did this patient have a deficit of water?
Return to the bedside. A simple examination of the data at 4 h revealed that the input of water was 3 l by intravenous infusion, while the patient excreted 3 l of urine (left side of Figure 1). Because non-renal water loss should be small over such a short period,5 the patient was in water balance. Hence the only way to account for the rise in her PNa was a positive balance of Na+. Because only 150 mmol of Na+ + K+ were excreted (3 l urine x 50 mmol/l Na+ + K+) while 450 mmol of Na+ were infused (3 l infused with 150 mmol/l Na+), there was a 300 mmol positive balance for Na+. ‘Accordingly, the goal for therapy is now clear, create a negative balance of 300 mmol of NaCl without creating a positive balance for water,’ stated Professor McCance. ‘I call this form of analysis a tonicity balance (Figure 1)—it has the advantage of revealing the specific goals for therapy’.6 In contrast, using the EFW balance approach, a deficit of 2 l of water or a gain of 300 mmol of Na+ are both called a deficit of 2 l of EFW. To make this point even clearer, Professor McCance illustrated another way to achieve the same negative balance for 2 l of EFW—one where this patient did not receive any intravenous fluid (Table 1). Hence there can be many ways to create a negative balance for EFW with the same PNa, but each requires its unique therapy to restore ECF and ICF volumes and compositions to normal.
|
Professor McCance quickly turned to analysing why the PNa rose from 153 to 168 mmol/l between hours 4 and 8 (right side of Figure 1). He began by calculating the water balance, using the data from the chart. Because the patient excreted 3.1 l of urine while receiving 0.1 l of isotonic saline, she now had a 3 l negative balance of water. The next step was to calculate the balance for Na+. The input was 15 mmol (0.1 l x 150 mmol/l) whereas the output was 155 mmol (3.3 l x 50 mmol/l), making a 140 mmol decrease in the positive balance for Na+. The combined results yield a positive balance of 160 mmol of Na+ and a 3 l deficit of water (now 21 l TBW). Hence the current PNa is 168 mmol/l, a value that is very similar to the calculated PNa, and this adds confidence to this analysis.
At this point, the intern looked perplexed. He asked, ‘What is the best way to treat the hypernatraemia in this patient?’
Question 3. What is the best way to treat hypernatraemia in this patient?
Return to the bedside. When therapy is considered in a patient like this with a rise in her PNa, possible options for therapy are to cause a negative balance for Na+ (with Cl– or 
) and/or to cause a positive balance for water. The decision depends on the tonicity balance calculation. In the first 4 h, the best way to treat her hypernatraemia was to create a negative balance of 300 mmol of Na+ (+Cl–). This could be achieved by giving dDAVP to raise the UNa + K and to infuse fluid with a lower Na+ concentration at a rate equal to the urine flow rate. The measured UNa + K was 175 mmol/l, so we could infuse 
-isotonic saline (75 mmol/l) at a rate equal to the urine output until 3 l of urine were excreted. On the other hand, to treat the second rise in PNa during hours 4 to 8, she needed a positive balance of 3 l of water and a deficit of 140 mmol Na+. Because the surplus of 300 mmol of Na+ was not corrected, the goal for the negative balance for Na+ should be adjusted downwards to create a deficit of only 160 mmol, because of the negative balance for Na+ in the second 4-h period.
‘So there is no correct, single therapy for all patients with the same degree of hypernatraemia’ concluded the intern. ‘Clearly, deficits must be replaced and surpluses lost.’ ‘There is no ‘one size fits all’ recipe for the therapy of hypernatraemia,’ agreed McCance. ‘And confidence comes from an understanding based on physiology with a quantitative analysis of the relevant data.’
The endocrine registrar rose and said, ‘One needs to be very careful if you intend to correct the water deficit by infusing 3 l of D5W in a short period of time. You cannot presume that this patient will be able to metabolize all the glucose you infuse.’ In a patient who is under considerable stress such as neurosurgery and has a severe degree of hypernatraemia, there would be a large adrenergic surge,7 she explained. Because one of the 
-adrenergic responses is to inhibit the release of insulin from ß-cells, independent of the plasma glucose concentration (PGlucose),8 this patient may not be able to metabolize all of the infused glucose. Direct measurements reveal that the expected rate of glucose metabolism is only 10–20 g/h (200–300 mg/min) in this setting.9 Since D5W contains glucose at 50 mg/ml, the rate of infusion of D5W should not exceed 6 ml/min (360 ml/h). Much slower rates of infusion of glucose should be given as the PGlucose rises towards 10 mmol/l (180 mg/dl), the renal threshold for glucose.10 Therefore following the recommendation for water therapy for the period from 4–8 h after starting surgery, it should take at least 6 h to replace a water deficit of 3 l. An example of a glucose-induced osmotic diuresis in a patient with hypernatremia and a large water deficit was reported by Skorecki et al.11
Professor McCance thanked the endocrine fellow for her clear description of glucose/insulin physiology and its importance for this patient. This was an excellent example of the need for integrative physiology and a quantitative analysis.
An enthusiastic discussion then ensued about the expected composition of the urine in patients with central DI. The questions raised included, ‘Is there an expected low Uosm in a patient with central DI?’ ‘Does the kidney have a minimum Uosm in a large water diuresis? ‘What determines the maximum urine flow rate in this setting?’ and ‘Will the nature of the urine osmoles be an important factor in determining the urine flow rate?’
Noting their interest, Professor McCance suggested that they perform a little experiment to obtain additional information on the Uosm during a water diuresis of similar magnitude to that in their patient.
|
Learning by self-experimentation |
|---|
To induce a water diuresis, the control group suppressed their release of vasopressin by ingesting a standard water load (20 ml/kg) rapidly (in 15 min). They measured their urine flow rate, Uosm, and the urine electrolyte concentrations. The study was carried out while fasting and at a similar time of day to the patient, to minimize the possible influence of diurnal variation on these parameters.12 To avoid a gender bias, the subjects included males and females and Professor McCance was the volunteer representing a control for age.
McCance had made one other suggestion for the design of this experiment. ‘Blood should be drawn from a brachial vein on one arm and from a vein on the back of the contralateral hand, which would be heated to close to 45°C’. The enthusiastic participants agreed to meet the next afternoon to discuss the results prior to their mortality meeting.
Before leaving, Professor McCance asked them to predict what their Uosm would be when the urine flow rate would be 10 ml/min. For simplicity, he said, ‘Assume that there are 1500 min in a day and you are on an average Western diet, and therefore are excreting 750 mosmol/day.’ The endocrine registrar quickly deduced that the osmole excretion rate would be close to 500 µosmoles/min (750 mosmoles/1500 min). Inserting this information into equation 2, she said that her expected Uosm would be close to 50 mOsm/kg H2O, a value that was appreciably lower than that measured in the patient (120 mOsm/kg H2O). She went on to make a very insightful comment. ‘Because a patient with complete central DI might have a different osmole excretion rate and/or a urine flow rate that was different than 10 ml/min, there should be no normal or expected urine flow rate or Uosm in this setting,’ she declared.
 | (2) |
|
The ‘post mortem’ with McCance: day 2 |
|---|
The intern initiated the discussion by asking: ‘Do we know the basis of the patient's polyuria?’
Question 4. What is the basis of the patient's polyuria?
Physiology principle 2. To understand why polyuria occurs, examine the parameters depicted in equation 2, the urine flow rate, Uosm, and osmole excretion rate.
Return to the bedside. Once again it was the razor-sharp endocrine registrar who took the lead. The patient clearly had a water diuresis due to central DI. There was also a high osmole excretion rate, she added. Before Professor McCance could comment, the eager intern asked, ‘But can an osmotic diuresis and a water diuresis really be present at the same time?’
Question 5. Can an osmotic diuresis and a water diuresis be present at the same time?
Physiology principle 3. When the MCD is not permeable to water because of lack of luminal water channels, the number of osmoles in the luminal fluid cannot influence the rate of reabsorption of water in this nephron segment (Figure 4).
|
Return to the bedside. When the distal nephron lacks luminal aquaporin 2 (AQP2) water channels, the urine flow rate is determined by the distal delivery of filtrate and the amount of water reabsorbed by pathways independent of AQP2 (called basal water permeability13 (equation 3, left side of Figure 4). In contrast, the number of urine osmoles can influence the urine flow rate when the distal nephron is permeable to water (right side of Figure 4).
 | (3) |
The intern had one more question. ‘Your analysis suggests that there is no minimum Uosm in a patient who has a complete lack of actions of vasopressin. Is this true?’ he asked.
Question 6. Is it likely that there is a minimum Uosm during a water diuresis?
Physiology principle 3, restated. In a water diuresis, the Uosm depends on the osmole excretion rate and the urine flow rate.
Return to the bedside. The urine flow rate in the absence of vasopressin is determined by the distal delivery of filtrate. The latter is determined by the GFR and the amount of filtrate reabsorbed in the PCT and the LOH (equation 3). Reabsorption in the PCT can be decreased when the ECF volume is expanded. Reabsorption of water in the LOH occurs in its descending thin limb, driven by a higher osmotic pressure in the medullary interstitial compartment. The latter decreases markedly during a water diuresis (called medullary washout).14 Because the GFR, PCT reabsorption and the medullary interstitial osmolalities are all variables, and because the urea and NaCl excretion rates can be very different from patient to patient depending on their diet, there is no way to predict the maximum urine flow rate and minimum Uosm in a patient with central DI. Professor McCance suggested that they now look at the urine data from their mini-experiment.
|
Urine results from the water load experiment |
|---|
All 12 subjects had consumed the water load in 15 min. The first item to consider was the urine flow rate—it rose promptly 60–90 min after water ingestion began. When the urine flow rate was 10 ml/min, the Uosm was only close to half that of the patient (Table 2), indicating that the patient had a much higher osmole excretion rate. Moreover, all of the ‘extra’ osmoles in the urine were due to a higher rate of excretion of electrolytes while there was little difference in the rate of excretion of urea.
|
Professor McCance thanked the group for their willingness to participate in this study, and put two questions to them. For the first question, he asked ‘What is the reason for the high rate of excretion of electrolytes in the patient?’ He had been reassured beforehand that the patient did not consume a ‘high salt’ diet or take diuretics.
Question 7. Why did this patient have such a high rate of excretion of Na+ + K+?
Physiology principle 4. A large excretion of Na+ is the expected renal response when the effective circulatory volume is expanded and the blood pressure is maintained.
Return to the bedside. ‘There is a clinical entity called cerebral salt wasting (CSW)’, stated the intern15 who had previously done a neurosurgical rotation. ‘Because this patient has a 5-fold higher rate of excretion of Na+ than we did, I assumed that the diagnosis must be CSW because of the CNS lesion,’ he said.
Professor McCance cautioned the intern, ‘I do not believe that you can make this diagnosis at this juncture.’ My reasoning is as follows. Wasting means that the patient is excreting Na+ when there is no physiological stimulus to do so.16 Recall that your patient had an initial 300 mmol positive balance for Na+. This means that her ECF volume was expanded. If the central venous pressure was not low, there should be a stimulus to excrete Na+. Hence the diagnosis of cerebral salt wasting should only be made when there is a low effective circulating volume. For now, my suggestion would be to attribute the natriuresis to the expanded ECF volume. Because I cannot include or exclude a diagnosis of cerebral salt wasting at this point, it is important to keep this in mind for future management.
Professor McCance then posed his second question, which switched the focus to management issues. He pointed out that vasopressin would be needed to diminish the water loss in this patient with central DI. ‘What will the expected urine flow rate be once vasopressin acts?’ he asked.
Question 8. What will the expected urine flow rate be after dDAVP acts?
Physiology principle 5. The urine flow rate is dependent on two factors when the distal nephron is permeable to water—it will vary directly with the osmole excretion rate and inversely with the Uosm (equation 2).
Return to the bedside. The medical consultant, who had a strong interest in clinical epidemiology, offered to tackle this question. ‘The way we should do this is to compare results in this patient to a group of normal subjects who have vasopressin acting. For example, because their urine output can be 1 l/day and there are 1440 min/day, the urine flow rate should be 0.67 ml/min,’ said this expert. Professor McCance was asked if he wished to add to this lucid, modern approach to clinical medicine. Never one to be rude or confrontational, he began by stating the obvious. ‘My emphasis continues to be on an application of principles of physiology at the bedside and to rely on a quantitative analysis wherever possible. Therefore I have a different way to answer question 8. I shall also provide reasons for preferring this physiology-based answer.’
First, there are no normal values for the composition of the urine. Rather normal subjects will excrete metabolic wastes (primarily urea from protein oxidation) and electrolytes ingested in excess of body needs (primarily Na+ and Cl–) in a volume of urine that contains water that is also ingested and exceeds needs. In more detail, if a very large volume of water is ingested, the urine should be very dilute (have as low a Uosm as possible) whereas if there is a water deficit, the Uosm should be as high as possible. Therefore both of these Uosm values are normal—a better word is ‘expected’—for each specific clinical setting. Therefore, I must disagree with analyses that use ‘normal values’ for the urine. Second, on a typical Western diet, the expected number of urine osmoles is 750/day. When I extrapolate her urine flow rate of 10 ml/min to a 24-h period (10 ml/min x 1440 min/day), the 24-h urine volume becomes 14.4 l/day. Multiplying this value by her Uosm of 120 mOsm/kg H2O (1728 mosmoles/day), it is obvious that her osmole excretion rate is more than two-fold higher than in our volunteers. Hence once vasopressin acts, the urine flow rate should be about two-fold higher than in normal subjects. if this were the only factor that was operating.
Notwithstanding, there are two other factors that make the expected urine flow rate after dDAVP acts considerably higher than this initial prediction. First, a prior water diuresis lowers the expected value for the maximum Uosm because of washout of the renal medullary interstitial compartment—‘typical values in this setting are 350 mOsm/kg H2O in my experience’, said Professor McCance.17 The second factor is that not all urine osmoles are effective osmoles, because the tubule becomes permeable to urea when vasopressin acts.18 Accordingly, if electrolytes constitute much more of the urine osmoles, as I suspect they will in this patient, the urine flow rate will be even higher when dDAVP first acts.19
‘Professor, we observed just what you predicted!’ stated the medical intern. The urine flow rate after dDAVP was 5 ml/min and the Uosm was 375 mOsm/kg H2O. The UNa + K was 150 mmol/l, making the Na+ + K+ excretion rate 7-fold higher than prior to surgery (750 vs. 125 µmol/min). ‘The final issue concerns the nature of the urine osmoles’, continued McCance. Only the urine electrolytes (effective osmoles in the urine) should lead to a higher urine flow and (as shown in Figure 1) your patient had a very high electrolyte excretion rate.
In summary, there are three causes for the high urine flow rate when vasopressin acts: the higher osmole excretion rate, the lower medullary interstitial osmolality, and the higher number of effective osmoles in the urine.
|
Hospital course |
|---|
Despite the fact that the ECF volume was expanded, every time there was a change of staff in the ICU, it seems that they ‘detected’ a modest degree of ‘dehydration’ and gave her a bolus of isotonic saline. This, coupled with the fact that the UNa rose to 300 mmol/l several hours after therapy with dDAVP began, resulted in a progressive decline in her PNa (Figure 5) and eventually to her death.
|
‘Let us go back in time to when her PNa had fallen to 140 mmol/l due to the removal of the surplus of NaCl and replacement of the deficit of water,’ said the intern, ‘How could we have changed our therapy to prevent an excessive decline in her PNa?’
Question 9. What therapy would prevent an excessive decline in her PNa?
Physiology principle 2, restated. To keep the PNa and body composition unchanged, the input of water and individual electrolytes must equal their urinary losses during a polyuric state (Figure 1).
Return to the bedside. ‘There are two major types of therapy to achieve Na+ and water balance,’ said Professor McCance. The first can be called ‘input strategy’. To design this therapy, one must measure the urine flow rate and its electrolyte composition. With this information, one knows the ideal electrolyte composition of the intravenous infusion. Therefore, you must infuse 300 mmol NaCl/l at a rate equal to the urine volume to have tonicity balance and there will be no change in the PNa. The danger is that there may be a delay before you are made aware that the actions of dDAVP have worn off. As a result, hypertonic saline will be infused while the UNa is very low. The second strategy is to place leverage in the composition of the urine—‘output strategy’. The aim is to lower the sum of the UNa + UK to equal the PNa in plasma or the intravenous fluids (isotonic saline). To achieve this aim, you must ‘amputate the LOH’. This will prevent the occurrence of a UNa + UK that is appreciably greater than the PNa. Now the infusion will be isotonic saline at a rate equal to the urine flow rate.
Summary. The medical team now understood how to prevent the development of hyponatraemia, but they were still not sure why their patient had died of brain herniation at a PNa where most other patients did not have this very unfortunate complication. Hence, the next step was to see what new issues might be revealed at Mortality Rounds. The reader should stop and think what new information would be helpful to understand why this patient had such a severe degree of brain cell swelling.
|
The mortality meeting |
|---|
The pathologist began by reminding the audience that they had to answer only one question, ‘Why did this patient die from severe brain swelling?’ There was no problem with the diagnosis of the craniopharyngioma, the neurosurgery was skillfully performed, and the only recognizable problem was hyponatraemia. Moreover, the patient did not receive unusual drugs and the standard postoperative orders were followed with precision. At autopsy, the major finding was brain swelling.
‘Let me focus on the medical management for the central DI,’ he said. As with all other patients who developed central DI during neurosurgery, the patient was given dDAVP. Perhaps one could blame the delay in instituting therapy, but this is not a unique occurrence. Her urine output was replaced with an equal volume of 50 mmol/l NaCl on an hourly basis intravenously. One can readily see that this therapy will cause the PNa to fall, because the Na+ concentration being infused is much lower than the UNa, despite the fact that the volume infused is equal the urine output (Figure 2). Clearly this therapy should be changed as of today because with this therapy, her PNa declined steadily towards and past the normal range (Figure 5). All eyes turned to Professor McCance for help, and he was not about to disappoint his colleagues. ‘There are two lines of evidence for my hypothesis to explain why this patient developed fatal brain cell swelling,’ he said. With chalk in hand to illustrate his thoughts, he began with a simple question. He asked, ‘Which source of blood should we rely on to indicate the degree of brain cell swelling: an arterial or a venous sample?’
Question 10. Which source of blood should we rely on to indicate the degree of brain cell swelling: an arterial or a venous sample?
Return to the bedside. There are two points I wish to stress at this point, said McCance. First, to assess the degree of cell swelling in an organ, we should use the PNa in plasma from the capillary blood in that organ. The problem is, we can only draw arterial or venous blood. Usually, arterial and venous PNa values will be equal. Notwithstanding, there may be differences in acute settings when very large infusions or urine flow rates occur. ‘What are the caveats that might make the venous PNa an unreliable indicator of the degree of brain cell swelling?’
Question 11. What are the caveats that might make the venous PNa an unreliable indicator of the degree of brain cell swelling?
Physiology principle 6. The venous PNa indicates the value in the interstitial compartment that surrounds cells in that venous drainage region.
Return to the bedside. ‘The brachial vein PNa tells me about the PNa around muscle cells in the arm, but it may not tell me about the PNa close to brain cells,’ said Professor McCance. This new thought process created a flurry of excitement and a series of questions. The nephrologists asked, ‘Are there any data to back up your speculation, Professor?’ Let me add, I mean no disrespect, but we do live in an era of ‘evidence-based-medicine’.
|
Blood results from the mini-experiment |
|---|
The medical team now knew why their Professor asked them to have simultaneous blood samples drawn from the brachial vein from one arm and a dorsal hand vein from the heated, contralateral arm. The intern from the medical team provided the audience with the data from the blood tests (Figure 6). ‘Please look first at the data before we ingested water,’ she said. There are virtually identical values for the venous and arterialized PNa. Let me assure you that the arterialized blood was red in colour and had a much higher saturation with oxygen than venous blood (95 vs. 50%) as well as a lower PCO2 (42 vs. 49 mmHg). Now look at the values in the first 60 min after water was ingested. The most striking difference is that the arterial PNa was on average 4 mmol/l lower than in blood drawn simultaneously from the brachial vein.20 The peak value for this PNa difference varied in time in individual subjects, but it was present in all of us,’ he concluded. The audience went silent for a moment, and then the first question was asked. ‘Which organ will swell faster after this water load, brain or muscle cells?’
|
Question 12. Which organ will swell faster after this water load, brain or muscle cells?
Physiology principle 7. The organ with the largest blood flow rate per unit weight should have the greatest degree of swelling.
Return to the bedside. Everyone quickly realized that the brain had a much larger blood flow rate per kg mass. Professor McCance added a quantitative dimension to this analysis. While brain and all the muscles at rest each receive 20–25% of the cardiac output, the muscle weighs 20-fold more than the brain. The dangers of early brain cell swelling are much greater, because this swelling occurs within the bony skull. One member of the audience wondered about the physiological relevance of this requirement for a large fall in the arterial PNa to induce a water diuresis. She asked Professor McCance, ‘How do you interpret this information using your favourite topic of integrative physiology?’
Question 13. What are the physiological implications of more rapid brain cell swelling?
Professor McCance began his answer with two general points. First, some look despairingly on speculations where there is not a strong database. ‘I have a different view,’ he said. ‘I think it is useful to convert data into knowledge whenever possible, even if this requires one to speculate.’ This can help drive creative thinking.
Let me summarize what I know about water. Water is essential for survival. We need to have evaporation of water from sweat to dissipate some of the heat generated during metabolism. Moreover, water consumption often occurs at different times from maximum heat generation (exhaustive exercise). Hence, should a large volume of water be ingested causing a significant degree of brain cell swelling, this ‘dangerous water’ should be excreted promptly. Hence the integrative physiology of water has two ‘expected responses’. Smaller volumes (say up to 3% of TBW) should be retained for ‘future sweat’—this is valuable water. In contrast, much of the ‘dangerous’ water (which can potentially cause appreciable brain cell swelling) should be excreted rapidly.20 One of the housestaff asked, ‘What mechanisms might permit a normal subject to retain this ingested ‘valuable water’?’
Question 14. What mechanisms might permit a normal subject to retain this ingested ‘valuable water’?
Physiology principle 8. For ingested water to lower the arterial PNa, we must consider its rate of absorption from the intestinal tract, the cardiac output and the volume of water removed from the circulation each minute, because the blood volume (5 l) makes a complete circuit each minute at rest (cardiac output is 5 l/min).
Physiological analysis. In answering this question in the context of cell swelling, the critical item is the volume of water that enters and leaves the body without effective osmoles—this in essence means water without Na+ and K+ salts. Ingesting water at a slower rate (‘sipping’ rather than ‘gulping’) will minimize the impact on the arterial PNa and allow distribution of this water across all body fluid compartments. There will be less suppression of vasopressin release, and the water is therefore retained. ‘Is there ever a condition in muscle that will permit this very large organ to draw water into myocytes?’ asked the medical intern.
Question 15. Is there ever a condition in muscle that will permit this very large organ to draw water into myocytes?
Physiology principle 1, restated. If the number of effective osmoles in muscle cells were to rise, these cells would import water and the PNa in venous blood draining these muscle cells would be much higher than in the arterial blood.
Return to the bedside. The nephrologist rose and said, ‘Now I understand a puzzling observation that supports what Professor McCance has been saying. Very shortly after a seizure, the venous PNa can rise by 15 mmol/l—this rise is transient.’21 It implies that muscle cells in this venous drainage bed have had a breakdown of large macromolecules into many smaller ones that largely are retained in these cells (Figure 7). Should blood be drawn from a vein very shortly after a seizure, the blood draining this muscle might have a much higher PNa than in arterial blood. In fact, this is just what happened in our patient, because she had a seizure just prior to obtaining the last blood sample.
|
|
The final issue: why did this patient die? |
|---|
Having thought about these nuances, the audience was prepared to return to the issues in the patient. The key question was ‘What are the possible causes of excessive brain swelling?’
Question 16. What are the possible causes of excessive brain swelling?
Physiology principle 9. Given the fixed volume of the skull, this rise in ICP was likely due to a larger than expected degree of brain cell swelling.
Return to the bedside. Consistent with this view, the neurosurgical team reported that there was no evidence of an intracerebral haemorrhage or another lesion in the brain. The medical team commented that the patient now had a normal balance for Na+, so that a rise in the ECF volume in the skull was not likely. Therefore all eyes turned to Professor McCance, expecting one of his novel explanations for this clinical conundrum. Because the time course was so short, he thought that solute gain in brain cell was an unlikely explanation for her excessive brain swelling, but what could it be?
‘Let us not forget that our patient was a young woman,‘ said McCance. ‘Perhaps her youth and her gender conspired to cause this fatal outcome.’ The audience could hardly wait for him to continue. After a long pause to allow his colleagues to ponder on this, he explained that the ratio of brain volume to skull volume was larger in younger patients and therefore, for a given degree of brain cell swelling, they were more likely to have a greater rise in intracranial pressure and serious neurological symptoms. In addition, being female with a small muscle mass, our patient had a smaller total body water compartment. A given water load would therefore cause a greater fall in PNa than in a male with the usual muscle mass. As his colleagues nodded in appreciation, McCance prepared to deliver the coup de grace.
He asked the intern to re-examine the chart for an unusual water input or output prior to the brain herniation. He also asked the ICU team whether the blood sample provided for the PNa measurement was from arterial or venous blood. The answer to the latter question was prompt—it was from a central venous catheter. ‘Aha!’ said our professor, ‘Our mini-experiment provided a necessary clue. If it is supported by the finding I expect to see from the chart, we may have a valid hypothesis for this excessive brain cell swelling.’ The conference room was now buzzing with excitement and anticipation.
Question 17. What did Professor McCance expect to find from the chart review?
Physiology principle 10. When the actions of dDAVP wear off, there will be a prompt and large water diuresis. Because of the absence of AQP-2 water channels, the expected water diuresis will depend on the volume of filtrate delivered to the distal nephron (equation 3 and left side of Figure 4).
Return to the bedside. When a patient has an expanded ECF volume and/or hyperglycaemia, there can be reduced reabsorption of filtrate in the PCT. In addition, if the medullary interstitial compartment is washed out and has a lower osmolality, less filtrate will be reabsorbed in the descending thin limb of the loop of Henle and hence distal delivery of filtrate will rise. Accordingly, I expect to find a very rapid infusion of a Na+-poor solution at a time when the urine output is low. This could cause the arterial PNa to be much lower than the measured venous PNa—an ‘occult’ reason for severe brain cell swelling.
The intern who had done the chart review was amazed. Two hours prior to death, the effects of dDAVP had worn off and polyuria developed—in fact the urine flow rate rose to 25 ml/min, much higher than in any of the volunteers (Figure 8). This had been matched by increasing the intravenous infusion rate, but after the next dose of dDAVP, the urine output fell and there was a markedly positive hourly water balance, just prior to brain herniation.
|
One of the housestaff said, in a loud stage whisper. ‘There he goes again, using principles of physiology (Table 3), a quantitative analysis, and balance data to uncover what was right before our eyes, but we could not see it!’ And the room filled with good-natured laughter and some spontaneous applause. The doctors involved with the case had learnt an important lesson—go slowly and methodically, and apply principles of physiology wherever possible. There was one other important caution. There was no reason to replace the large urine output in the previous hour after dDAVP had been given with hypotonic fluid because the patient was hyponatraemic. Professor McCance was later seen to leave the conference room with a look of contentment—his students had clearly gained valuable new insights and, for a teacher, that was the ultimate reward.
|
|
References |
|---|
1. Mason DT, Bartter FC. Autonomic regulation of blood volume. Anesthesiology 1968; 29:681–92.[CrossRef][Web of Science][Medline]
2. Steele A, Gowrishankar M, Abrahmson S, Mazer D, Feldman R, Halperin ML. Postoperative hyponatremia despite isotonic saline infusion: a phenomenon of ‘desalination’. Ann Intern Med 1997; 126:20–5.
3. Goldberg M. Hyponatremia. Med Clin North Am 1981; 65:251–69.[Web of Science][Medline]
4. Rose BD. New approach to disturbances in the plasma sodium concentration. Am J Med 1986; 81:1033–40.[CrossRef][Web of Science][Medline]
5. Shafiee MA, Bohn D, Hoorn EJ, Halperin ML. How to select optimal maintenance intravenous fluid therapy. Q J Med 2003; 96:601–10.[Web of Science]
6. Carlotti APCP, Bohn D, Mallie J-P, Halperin ML. Tonicity balance and not electrolyte-free water calculations more accurately guide therapy for acute changes in natremia. Intens Care Med 2001; 27:921–4.[CrossRef][Web of Science][Medline]
7. Fraser CL, Arieff AI. Fatal central diabetes mellitus and insipidus resulting from untreated hyponatremia: a new syndrome. Ann Intern Med 1990; 112:113–19.[CrossRef][Web of Science][Medline]
8. Porte DJ. Sympathetic regulation of insulin secretion. Arch Intern Med 1969; 123:252–60.
9. Bjorntorp P, Sjostrom L. Carbohydrate storage in man: speculations and some quantitative considerations. Metabolism 1978; 27:1853–85.[Web of Science][Medline]
10. Deetjen P, Baeyer HV, Drexel H. Renal glucose transport. In: Seldin D, Giebisch G, eds. The Kidney: Physiology and Pathophysiology, 2nd edn. New York, Raven Press, 1992: 2873–88.
11. Halperin ML, Skorecki KL. Interpretation of the urine electrolytes and osmolality in the regulation of body fluid tonicity. Am J Nephrol 1986; 6:241–5.[Web of Science][Medline]
12. Moore-Ede MC, Sulzman FM, Fuller CA. The clocks that time us. Cambridge MA, Harvard University Press, 1982.
13. Lankford SP, Chou C, Terada Y, Wall SM, Wade JB, Knepper MA. Regulation of collecting duct water permeability independent of cAMP-mediated AVP response. Am J Physiol 1991; 261:F554–66.
14. Jamison RL. A patient with polyuria and hyponatremia. Kidney Int 1983; 24:256–67.[Web of Science][Medline]
15. Harrigan MR. Cerebral salt wasting: a review. Neurosurgery 1996; 38:152–60.[CrossRef][Web of Science][Medline]
16. Singh S, Bohn D, Cusimano M, Rutka JT, Halperin ML. Cerebral salt wasting; truths, fallicies,theories and challenges. Crit Care Med 2002; 30:2575–9.[CrossRef][Web of Science][Medline]
17. Davids MR, Edoute Y, Halperin ML. The approach to a patient with acute polyuria and hypernatremia: a need for the physiology of McCance at the bedside. Netherlands J Med 2001; 58:103–10.[CrossRef][Web of Science][Medline]
18. Gamble JL, McKhann CF, Butler AM, Tuthill E. An economy of water in renal function referable to urea. Am J Physiol 1934; 109:139–54.
19. Gowrishankar M, Lenga I, Cheung RY, Cheema-Dhadli S, Halperin ML. Minimum urine flow rate during water deprivation: Importance of the permeability of urea in the inner medulla. Kidney Int 1998; 53:159–66.[CrossRef][Web of Science][Medline]
20. Shafiee MA, Charest AF, Cheema-Dhadli S, Glick DN, Napolova O, Roozbeh J, et al. Defining conditions that lead to the retention of water: the importance of the arterial sodium concentration. Kidney Internat 2005; 67:613–21.[CrossRef][Web of Science][Medline]
21. Welt LG, Orloss J, Kydd DM, Oltman JE. An example of cellular hyperosmolarity. J Clin Invest 1950; 29:935–9.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. A. Onitilo, E. Kio, and S. A. R. Doi Tumor-Related Hyponatremia Clin. Med. Res., December 1, 2007; 5(4): 228 - 237. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. Narins, Participants:, M. Halperin, G. Danovitch, R. Falk, and J. Bargman The Nephrology Quiz and Questionnaire: 2005 Clin. J. Am. Soc. Nephrol., May 1, 2006; 1(3): 592 - 608. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||