Symptomatic Hyponatremia with Hypoxia Is a Medical Emergency

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

see original article on page 1319

rection of serum sodium concentration Symptomatic hyponatremia with and one who died in the absence of any therapeutic interventions to correct her hypoxia is a medical emergency hyponatremia.4 We now understand that much heterogeneity exists among 1 JP Kokko patients with hyponatremia. Some patients will tolerate a relatively con- Patients presenting with the combined finding of severe symptomatic servative approach, whereas others are at hyponatremia and hypoxia have such high mortality rates that they high mortality risk if their symptomatic should be admitted to an intensive care unit and intubated sooner rather hyponatremia is not corrected rapidly. than later. Without delay, these patients need rapid correction of their Many clinicians have made important serum sodium by 8–10 mequiv. per liter and an increase in their partial contributions to this area; perhaps the studies of Ayus, Arieff , and their col- pressure of arterial oxygen to values above 70 mm Hg. leagues are most direct in pointing out Kidney International (2006) 69, 1291–1293. doi:10.1038/sj.ki.5000252 that the patients at highest risk are the ones with the combined fi nding of symptomatic hyponatremia and hypoxia.5–7 It Hyponatremia is a frequent electro- sodium concentration was rapidly cor- should be noted that neither the early lyte abnormality among hospitalized rected from 106.3 to 151.8 mequiv. per clinical observations (before 1992) nor patients. In general, it is classified as liter with the use of hypertonic saline. animal studies addressed the possible acute (less than 48 hours in duration) Similar fi ndings of CPM could not be adverse eff ect of hypoxia on outcomes in versus chronic (more than 48 hours in duplicated by Ayus and colleagues2 symptomatic hyponatremia. To under- duration). It is further characterized as when hyponatremia in their rats was stand why hypoxia is a severe comorbid asymptomatic versus symptomatic (agi- increased by 14–25 mequiv. per liter in factor requires a paragraph on brain cell tation, depressed sensorium, grand mal the fi rst 24 hours, that is, not overcor- volume regulation. seizures). Th is simple, albeit superfi cial, rected rapidly to values above normal. Essentially all cells have the capacity classifi cation is important, as it suggests However, physicians are taught not to to regulate their volume. Th e regulation an appropriate therapeutic approach to a harm their patients, and therefore the of cell volume by the brain cells is criti- given patient. Before the 1980s, many of fear of CPM caused such concern that cally important because acute swelling us who were in clinical practice felt that many physicians became afraid to of the brain cells is limited by the fi xed if we saw a laboratory abnormality, it treat symptomatic hyponatremia, even skull, which will cause pressure necrosis should be corrected rapidly. However, it when they recognized that failure to and/or transtentorial herniation. On the was becoming evident that patients with treat was associated with high mortal- other hand, rapid shrinkage in response symptomatic hyponatremia had an unac- ity rates. Th is conservative view against to an external hypertonic gradient can ceptably high mortality rate whether rapid correction of plasma sodium con- cause adverse cellular events including they were left untreated or whether centration was supported by a careful putative disruption of controlled mye- their sodium concentration was rapidly paper by Sterns,3 in which he observed lin synthesis and degradation. Cell vol- corrected. In a classic paper published among his 62 patients that the previ- ume regulation has been studied more in 1981,1 Kleinschmidt-DeMasters and ously noted high mortality rates with extensively when cells are challenged to Norenberg reported an experimental symptomatic hyponatremia were not protect their volume by exposure to a study in rats that was based on their pre- nearly as high as some of the literature hypo-osmotic external milieu than when vious clinical experience showing that then suggested if the low sodium con- they are exposed to a hypertonic envi- central pontine myelinolysis (CPM) is an centration was corrected at low versus ronment such as a rapid rise in plasma ‘iatrogenic disorder’ caused by rapid cor- high rates. Nevertheless, many patients sodium concentration. When cells are rection of hyponatremia. Although their were still dying when symptomatic exposed to a hypo-osmotic environment conclusion that CPM is an iatrogenic hyponatremia was left untreated and (hyponatremia), they swell unless there disorder was correct, it should be noted allowed to correct with spontaneous is an appropriate outward movement that in their experiments the rats’ serum water diureses. Th is led Kidney Interna- of intracellular solutes. Th is is largely tional to devote a “Nephrology Forum” accomplished by activation of leak path- 1Department of Medicine, Emory University School to “Treating hyponatremia: damned if way channels that initially allow outward of Medicine, Atlanta, Georgia, USA we do and damned if we don’t,” in which diff usion of potassium and chloride and Correspondence: JP Kokko, Emory University School of Medicine, 69 Jesse Hill, Jr. Drive, Suite 205, Berl described being presented with two later allow outward diff usion of organic Atlanta, Georgia 30303, USA. patients with hyponatremia, one who solutes. It is beyond the scope of this E-mail: [email protected] developed CPM following rapid cor- Commentary to review these processes;

Kidney International (2006) 69 1291 commentary

the interested reader is directed to an in those patients who had their serum remaining diff erence in sodium concen- excellent review of this topic by Pas- sodium concentration corrected at rates tration. I advise not to raise the sodium antes-Morales and colleagues.8 Many below 0.55 mequiv. per liter per hour.10 concentration much over 120 mequiv. biologic regulatory processes are not So what is the clinician to conclude per liter in the fi rst 24 hours and then symmetric. This certainly is the case from all these studies when faced with to correct the remaining defi cit at a rate when cellular processes are activated to a severely hyponatremic patient? Th ere that improves serum concentration protect the volume of cells exposed to is no question that many asymptomatic each 24 hours by 50% of the desired a hypertonic milieu that would tend to patients with normal arterial oxygen fi nal sodium concentration. One of the shrink them. Under these circumstances saturation and severe hyponatremia do practical issues I have faced in the emer- the cell must gain net solute against elec- well without rapid correction of their gency room is the unavailability of 3% trochemical gradients and requires much serum sodium concentration; how- NaCl or an enormous delay in getting it more oxygen-dependent energy than is ever, it is now clear that symptomatic there. Given the lack of any evidence to provided by activation of outward leak hyponatremic patients, especially if the contrary, I have recommended push- pathway channels. Th us it is evident that they are female and hypoxic with arte- ing an ampoule of NaHCO3 in these conditions in which oxygen availability rial partial pressure of oxygen below circumstances (an ampoule usually con- is limited are just those conditions in 70 mm Hg, should receive rapid cor- tains either 44 or 50 mequiv. NaHCO3); which cells are unable to have suffi cient rection of their sodium concentration this will allow time to get the 3% NaCl inward transport rates of solutes to pro- to safe levels. Th us, to use Berl’s color- delivered. I feel that the weight of the tect against cell shrinkage. ful choice of words,4 one must conclude literature supports initiation of the cor- As they report in this issue,9 Ayus, that we are damned if we don’t, or at rection of the osmolality as soon as it is Armstrong, and Arieff have carried least the patient is. feasible. I make these same therapeutic out an important series of studies that Now comes the important question of recommendations also for patients who extends many of their previous obser- how to administer the sodium, and at have symptomatic hyponatremia but vations. Th ey show in three groups of what rate it will decrease the very high normal oxygenation. Although these rodents that hypoxia interferes with mortality associated with the combi- patients do not have the same high risk normal brain cell volume adaptation in nation of hypoxia and hyponatremia. of mortality as hypoxic patients, they response to hyponatremic challenges. In The literature supports the view that nevertheless are at increased risk of hyponatremic animals, hypoxia aggra- the rate of correction should be at least dying. If, on the other hand, the patient vated cerebral edema, decreased cerebral 0.55 mequiv. per liter per hour. It is is hypoxic for any reason (central res- perfusion, and led to histologic abnor- impractical in an acute hospital setting piratory depression secondary to brain malities in the cerebellum, thalamus, to try to calculate accurately how much edema, pneumonia, or any other cause), reticular formation, and basal ganglia.9 sodium to give. First, it is diffi cult to or if the patient’s arterial oxygen begins Many of these histologic fi ndings are obtain accurate patient weights, good to drop during the course of treatment, similar to those that have been noted estimates of total body water, and esti- then he or she should be hospitalized previously to occur when hyponatremia mates of bone weight (approximately in an intensive care unit and intubated has been rapidly corrected. It is interest- 43% of body sodium is in functionally sooner rather than later if the arterial ing that many of these histologic abnor- non-exchangeable bone pool); and sec- partial pressure of oxygen cannot be malities were not seen when hypoxia ond, although administered sodium raised to over 70 mm Hg by an increase was not present. Although similarities freely moves between plasma, inter- in fraction of inspired oxygen (FIO2). in histologic changes do not necessarily stitium, and, thus, lymph, the rate of If comorbid conditions are recognized refl ect similarities in functional altera- sodium equilibration between intrac- early in symptomatic hyponatremic tions, taken together with the other ellular compartments, connective tis- patients, and if treatment is initiated rap- observations of Ayus et al. they do sug- sue, and so on is slower. Nevertheless, idly, then we can reasonably expect to gest that “hypoxia appears to be a major some guidelines can be suggested. I decrease the very high mortality other- co-morbidity factor in the outcome of have recommended ordering 250 ml of wise reported. We should not dismiss the patients with hyponatremic encephalo- 3% NaCl (514 mequiv. per liter), giv- possibility of CPM, but our therapeutic pathy.”9 In support of this conclusion is ing 50 ml by intravenous push (which approaches should not be paralyzed by a large metaanalysis that was reported at should raise the plasma sodium con- fear that CPM will develop. the 2005 American Society of Nephrology centration by 1–2 mequiv. per liter), REFERENCES meeting (A.I. Arieff , M.S. Etheridge and and letting the remaining 200 ml drip 1. Kleinschmidt-DeMasters BK, Norenberg MD. E.L. Bradley, J Am Soc Nephrol 16: 44A, for the next 4–6 hours. Th is generally Rapid correction of hyponatremia causes 2005) of 344 patients with symptomatic raises the serum sodium concentration demyelination: relation to central pontine myelinolysis. Science 1981; 211: 1086–1070. hyponatremia over a 30-year period that by 8–10 mequiv. per liter and gets the 2. Ayus JC, Krothapalli RK, Armstrong DL, Norton concluded that associated hypoxia was a patient out of the acute neurologic dan- HJ. Symptomatic hyponatremia in rats: effect of major determinant of adverse outcome gers of hyponatremia. Th is will then buy treatment on mortality and brain lesions. Am J Physiol 1989; 257: F18–F22. and that mortality was 3.5 times greater time in which to plan how to correct the 3. Sterns RH. Severe symptomatic hyponatremia:

1292 Kidney International (2006) 69 commentary

treatment and outcome. Ann Intern Med 1987; failure. Chest 1995; 107: 517–521. identifying high-risk individuals, and to 107: 656–664. 7. Ayus JC, Arieff AI. Chronic hyponatremic improving treatment and overall progno- 4. Berl T. Treating hyponatremia: damned if we encephalopathy in postmenopausal women: do and damned if we don’t. Kidney Int 1990; 37: association of therapies with morbidity and sis. Besides the non-genetic factors, drug 1006–1018. mortality. JAMA 1999; 281: 2299–2304. responses are also known to be infl uenced 5. Vexler ZS, Ayus JC, Roberts TPL et al. Ischemic 8. Pasantes-Morales H, Franco R, Ordaz B, Ochoa by inherited factors;3 this fi eld is known as or hypoxic hypoxia exacerbates brain injury LD. Mechanisms counteracting swelling in brain 3 associated with metabolic encephalopathy cells during hyponatremia. Arch Med Res 2002; pharmacogenomics. Th e basic concept is in laboratory animals. J Clin Invest 1994; 93: 33: 237–244. simple: a drug interacts with its target, for 256–264. 9. Ayus JC, Armstrong D, Arieff AI. Hyponatremia 6. Ayus JC, Arieff AI. Pulmonary complications of with hypoxia: effects on brain adaptation, instance an enzyme or a receptor. When hyponatremic encephalopathy: noncardiogenic perfusion and histology in rodents. Kidney Int genetic alteration leads to modifi ed target pulmonary edema and hypercapnic respiratory 2006; 69: 1319–1325. availability or function, the drug response is modifi ed as well. However, this concept also involves more complex relations see original article on page 1438 such as interplay between several genetic sites during therapy and the direct eff ects of drugs on gene transcription. Basically, Time to consider ACE insertion/ pharmacogenomics is the integrated eff ects of the genome on drug response. In deletion genotypes and individual this issue, So et al.4 provide more evidence supporting the importance of pharmacog- renoprotective treatment in enomics in diabetic patients by reporting an impact of the angiotensin-converting diabetic nephropathy? enzyme (ACE) insertion/deletion polymorphism in the renin–angiotensin sys- PK Jacobsen,1 L Tarnow1 and H-H Parving1,2 tem on response to ACE inhibition in 2089 Chinese type 2 diabetic patients with vari- One reason for the inadequacy of current renoprotective therapy and ous degrees of albuminuria. the persistent poor renal prognosis in diabetic nephropathy is the large Activation of the renin–angiotensin– interindividual variation in response to treatment. Genetic as well as aldosterone system is important in the non-genetic factors are known to influence treatment efficacy. This initiation and progression of diabetic nephropathy.1 Th e ACE insertion (I)/dele- Commentary summarizes the impact of the angiotensin-converting tion (D) polymorphism infl uences the enzyme (ACE) insertion/deletion polymorphism in the ACE gene on systemic and renal activity of the renin– initiation and progression of diabetic nephropathy. angiotensin system, as patients carrying Kidney International (2006) 69, 1293–1295. doi:10.1038/sj.ki.5000283 the D allele have increased systemic and renal ACE levels, whereas patients with the II genotype have the lowest ACE concen- Antihypertensive treatment has greatly diabetic nephropathy is still the most com- trations.1 A recent metaanalysis of 14 727 improved the renal prognosis and survival mon cause of end-stage renal disease in the diabetic patients included in case-control of diabetic patients with nephropathy over Western world. One reason for the inad- studies from 1994 to 2004 confi rms a sta- the past few decades.1 More patients die equacy of current renoprotective therapy tistically signifi cant protective role of the of other causes, primarily cardiovascu- and the persistent poor renal prognosis II genotype in the development of diabetic lar, with well-preserved kidney function. is the large interindividual variation in nephropathy;5 the eff ect was most pro- However, despite therapy targeting ele- response to fi rst-line therapy including nounced in Asians with type 2 diabetes, vated blood pressure, albuminuria, hyper- drugs blocking the renin–angiotensin– followed by Caucasians with types 1 and glycemia, and lipid abnormalities, patients aldosterone system, as is demonstrated 2 diabetes.5 Th e relation between the ACE with diabetic nephropathy still on average by the wide range in loss of glomerular I/D polymorphism and progression of have a rate of decline in renal function filtration rate (GFR) in these subjects. diabetic nephropathy has been studied three to six times that seen in individuals Indeed, this ranges from 0 to 24 ml per during the past 10 years. Minimal data without renal disease.1,2 Consequently, minute per year during treatment.1,2 Well- exist on the impact of ACE gene poly- known non-genetic risk factors, namely morphism during the natural course of 1Steno Diabetes Center, Gentofte, Denmark; and elevated blood pressure, albuminuria, diabetic nephropathy without any anti- 2Faculty of Health Science, University of Aarhus, poor glycemic control, and hypercholes- hypertensive treatment. In 59 normoten- Aarhus, Denmark terolemia, explain about 30%–50% of this sive type 1 diabetic patients with diabetic Correspondence: PK Jacobsen, Steno Diabetes 1,2 Center, Niels Steensens Vej 2, DK-2820 Gentofte, variation. Characterization of additional nephropathy and stable renal function Denmark. risk factors is the key to understanding the (decline in GFR of about 1 ml per minute E-mail: [email protected] progression of diabetic nephropathy, to per year), we found no relation between

Kidney International (2006) 69 1293