Have Advances in Extracorporeal Removal Techniques Changed the Indications for Their Use in Poisonings?

Fiona M. Garlich and David S. Goldfarb

During the past 25 years, numerous changes have taken place in the use of hemodialysis as a therapeutic modality. Advances in technologies and a progression in our collective understanding of the pharmacokinetics of certain xenobiotics have resulted in alterations in the indications, effectiveness, and safety of hemodialysis. However, these changes have not necessarily been reected in the current published data regarding treatment of intoxications. Reported clearance rates often reect what was achievable in the 1970s and 1980s, and more recent reports are frequently lacking. Our goal in this review is to summarize the changes in hemodialysis and in other extracorporeal removal technologies and highlight the effects of these changes on the current indications for hemodialysis of the poisoned patient. Changes in dialysis performance that are reviewed in this article include the use of high-efciency and high-ux dialysis membranes, improved hemodynamic stability because of ultraltration control, and the use of bicarbonate as a source of base. We review the indications for hemodialysis for removal of specic toxins, including vancomycin, methotrexate, carbamazepine, and valproic acid. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. Key Words: Carbamazepine, Hemoperfusion/trends, Methotrexate, Overdose, Poisons, Poisoning/therapy, Sorption detoxication/methods, Valproic acid

he use of extracorporeal removal (ECR) techniques to enhance the elimination of certain toxins has been used for decades in the treatment of the severely poisoned patient. Between 1985 and 2005, a total of 19,351 poisoning cases in which an ECR technique was used were reported to the American Association of Poison Control Centers. The number that received hemodialysis increased signicantly in this period, from 231 per million calls in 1985 to 707 per million calls in 2005.1,2 During the past 25 years, numerous changes have taken place in the use of hemodialysis as a therapeutic modality.3 Advances in technologies and a progression in our collective understanding of the pharmacokinetics of certain xenobiotics have resulted in alterations in the indications, effectiveness, and safety of hemodialysis. However, these changes have not necessarily been reected in the current published data regarding treatment of intoxications. Reported clearance rates often reect what was achievable in the 1970s and 1980s,4 and more recent reports are frequently lacking. In essence, the literature is old. Our goal in this review is to summarize the changes in hemodialysis and other ECR technologies, and highlight the effects of these changes on the current indications for hemodialysis of the poisoned patient.

Advances in ECR Techniques

Although many changes in the technical aspects of hemodialysis have occurred in the last decades, we will highlight only those we consider relevant to treatment of poisonings in this review, and not those related only to treatment of end-stage kidney disease. When we write about conventional hemodialysis, we are referring to the generic form of dialysis which was routinely used 3 times a week in the treatment of patients with chronic kidney failure before the more widespread use recently of larger and more porous membranes.

High-Efciency Hemodialysis Membranes

High-efciency is a term used to describe hemodialysis membranes that achieve higher clearances of the traditional low-molecular-weight marker of dialysis adequacy, urea. The denition of high-efciency hemodialysis is usually considered urea clearance greater than 210 mL/min.5 Higher ow rates of blood and dialysate are required to achieve these clearance rates. With the impetus in end-stage kidney disease, both in the United States and worldwide, to deliver higher urea clearances, the requirements of high-efciency hemodialysis became standard of care and are likely to be used in the treatment by a nephrologist of any life-threatening intoxication. To some extent, however, blood ow rates may be limited by hypotension and low-molecular-weight clearance thereby impaired.

From New York City Poison Control Center, NYU School of Medicine, New York, NY; Department of Emergency Medicine, NYU School of Medicine, New York, NY; Nephrology Section, NY Harbor V Medical Center, New York, A NY; and Division of Nephrology, NYU School of Medicine, New York, NY. Address correspondence to David S. Goldfarb, MD, Nephrology Section, New York Harbor V Health Care System, New York DV A AMC, 423 E 23rd St./111G, New York, NY 10010. E-mail: david.goldfarb@va.gov Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. 1548-5595/$36.00 doi:10.1053/j.ackd.2011.01.009

High-Flux Membranes
Traditionally, hemodialysis has been considered to be effective only for low-molecular-weight xenobiotics with molecular weight (MW) ,500 Daltons (Da). In the 1990s, the desire to remove amyloidogenic b2-microglobulin and other middle molecules in the setting of end-stage kidney disease led to the development and popularization

172

Advances in Chronic Kidney Disease, Vol 18, No 3 (May), 2011: pp 172-179

Advances in Extracorporeal Removal Techniques

173

of high-ux dialysis membranes.5 As opposed to the older, conventional, usually cellulose-derived membranes, the high-ux membranes are composed of various synthetic polymers, such as polysulfone, polyamide, polyacrylonitrile, and other proprietary compounds.6 Advances in techniques of mass production contribute to signicant reductions in price, which has also led to their wide availability. Their larger pore size allows for the clearance of larger molecules. As compared with conventional low-ux dialysis, dialysis with high-ux membranes demonstrates improved efcacy in the removal of middle molecules ranging in size from 1000 to .15,000 Da.7 The clearance of b2-microglobulin (MW, 11,800 Da) is often used to express the high-ux dialysance of a membrane. High-ux membranes are those that achieve .20 mL/min of b2-microglobulin clearance.5 The widespread availability of such membranes has made the consideration of clearance of more high MW xenobiotics possible.

hypotension, metabolic acidosis, and decreased cardiac output, and has now been entirely replaced by sodium bicarbonate.8 Older literature, before the late 1980s, reporting on hemodialysis as a treatment of poisoning may be limited by hemodynamic instability caused by acetate. The use of sodium modeling, a technique by which the sodium concentration of the dialysate can be programmed to vary during the course of the procedure, may also act to promote hemodynamic stability. Although bicarbonate-based hemodialysis is now the standard of care, sodium modeling as a useful technique in hemodynamic stabilization during ECR of xenobiotics has not been explored.

Higher Blood Flow Rates

Historically, blood ow rates during hemodialysis ranged from 250 to 400 mL/min. Now, higher blood ow rates, of a maximum of 400 to 500 mL/min, are routinely achieved and are considered as the standard of care in the treatment of end-stage kidney disease, in which higher urea clearUltraltration Control ance rates have been emUltraltration (also called phasized in recent years. CLINICAL SUMMARY hemoltration) is the moveThese higher blood ows ment of uid across a memare achievable because of  High ux dialysis membranes now allow dialysis of larger molecules, though efcacy remains limited to molecules brane in response to the improved hemodywith relatively low volumes of distribution and low hydrostatic pressure gradinamic parameters that arise protein binding. ents. Dialysis membranes from bicarbonate-based di Despite advances in membrane technology, the number of used for treatment of endalysate, ultraltration conpoisonings for which hemodialysis is indicated has stage kidney disease have trol, and the use of larger expanded relatively little with scant evidence of clinical had progressively higher uldual-lumen catheters. The signicance. traltration coefcients, direct result of these inno Instead of focusing on the technology of dialysis a measure of the amount of vations is the capacity for membranes, attention should be focused on improving uid moved across a memimproved clearance of access to dialysis in a timely manner in cases in which it brane in response to a prestoxins, which may have sigwill have the most clinical impact. sure gradient per unit time. nicant clinical effect in High-ux membranes have certain settings. Higher dihigh ultraltration coefcients. In the past, frequent hyalysate ow rates to values beyond those usually used curpotensive episodes complicated dialysis therapy, due in rently might maintain concentration gradients and large part to the need for repeated calculations and masomewhat increase clearance of protein-bound solutes.9 nipulations of the ultraltration rate. The development of computerized hemodialysis machines that register Hemodialtration small variations in the inow and outow of dialysate alHemodialtration is hemodialysis and ultraltration perlows ne control of ultraltration rates and the ability to formed together, as with conventional hemodialysis, but limit extracellular uid volume losses despite the memwith the difference that the rate of removal of extracellular branes high ultraltration coefcient. Ultraltration conuid volume vastly exceeds that needed to remove excess trol has become more important with the use of high-ux sodium and water.10 This technique is now widely used in membranes, as their higher water permeability requires Europe, but not in the United States, for management of more precise volume control. end-stage kidney disease. Augmentation of ultraltration leads to signicant enhancement of solute clearance by the Improved Hemodynamic Stability addition of convective removal of plasma to add to the difIn addition to ultraltration control, other changes in fusive clearance offered by hemodialysis. The removal of hemodialysis technology have allowed patients to expelarge volumes of plasma necessitates infusions with large rience less hemodynamic instability than that in the volumes of manufactured replacement uid. Clearance is past. One such change is in the source of base in the dialthen determined by the total volume of both the dialysate ysate. The use of sodium acetate caused vasodilation, and replacement solution. Because data suggest higher

174

Garlich and Goldfarb

clearances for b2-microglobulin, it is likely that the technique could be useful beyond high-ux hemodialysis in treatment of intoxication. However, data examining its improved efcacy for this indication are sparse. One case report of hemodialtration for treatment of a phenobarbital overdose demonstrates rapid removal but compares that successful treatment only with the historical clearance rates from data published in the 1980s.11 Thus, the benet in enhancing outcome compared with conventional hemodialysis is not demonstrated. Nonetheless, as hemodialtration becomes widely available, and perhaps if it gains traction in the United States, its anecdotal use for treatment of poisonings may be more reported.

Continuous Venovenous Hemo(dia)ltration

Continuous venovenous hemo(dia)ltration (CVVH or CVVHDF) is now widely used in the management of acute kidney injury. Its broad availability has no doubt contributed to its growing application to management of intoxications. Although the technique is ideally used to treat kidney failure and extracellular uid volume expansion in hemodynamically unstable patients, its efcacy for management of poisoning is limited by its relatively slower clearance rates when compared with hemodialysis. If hypotension limits the performance of hemodialysis, CVVH may be preferable to mere supportive care. The technique also has a logistic advantage as its frequent use in intensive care units often allows it to be initiated without dialysis unit staff mobilization. We have recently reviewed the use of these continuous modalities for treatment of poisoning.2,12

decreased dramatically. According to the data submitted to the American Association of Poison Control Centers, hemoperfusion was reported to have been used in only 16 cases in 2007.1 Several factors contribute to this decline in use, the most signicant of which is the lack of availability of the charcoal cartridges, which are expensive and now rarely stocked.13 In the past, when aluminum carbonate and aluminum hydroxide were more widely used as phosphate binders, chronic dialysis patients were at higher risk of developing aluminum toxicity, which was treated with charcoal hemoperfusion. Because calcium salts and sevelamer resins have supplanted aluminum salts as phosphate binders, the need for charcoal cartridges has declined. Furthermore, the prevalence of theophylline poisoning, historically the most common indication for charcoal hemoperfusion, has decreased in the United States.14,15 The net result is that if sorbent hemoperfusion was useful in management of some poisonings, its lack of indication for treatment of endstage kidney disease has led to its near vanishment. Charcoals superiority in management of salicylate or theophylline intoxication was of minor importance once the small-molecule clearance of hemodialysis was enhanced, as described earlier in the text. Even their concomitant use, by having dialysis membranes in series with charcoal hemoperfusion cartridges, seems unimportant today.

Changes in Indications for Dialysis in the Poisoned Patient

The cumulative effect of recent innovations in hemodialysis technology is an improvement in the efcacy and hemodynamic safety of dialysis, and an increase in the size of molecules that are amenable to dialysis. Additionally, some highly protein-bound molecules that were previously only amenable to removal by charcoal hemoperfusion, are adequately cleared by newer highux dialysis techniques.16-20 This means, in theory, that the use of high-ux dialysis could be expanded to include many other toxins with larger MWs or higher protein binding. Xenobiotics that can now be included in the list of dialyzable toxins include vancomycin, carbamazepine, methotrexate, valproic acid (VPA), and insulin. However, despite an increase in membrane pore size, a pharmacokinetic basis for the efcacy of hemodialysis must still be present for the technique to be applied. In addition, a compelling indication for application of the technique should be present. Hemodialysis can only be effective in management of a toxic xenobiotic if the molecule of interest has a low volume of distribution (,1 L/kg), low tissue binding, high water solubility, and poor or impaired endogenous clearance. Perhaps most importantly, hemodialysis should only be prescribed if a clinical benet to ECR has been demonstrated or at least is hypothesized. Despite improved plasma clearance, high-ux hemodialysis may simply not remove a high enough percentage of total

Peritoneal Dialysis
Peritoneal dialysis has also been the object of technologic advances. However, it has little indication for use in treatment of poisonings because of its relatively slow rate of clearance of even low-molecular-weight solutes. Advances in automated technology and changes in dialysate composition, as well as nancial considerations and incentives, may contribute to wider use of peritoneal dialysis for treatment of end-stage kidney disease. However, it seems unlikely that innovations in the technique will lead to peritoneal dialysis becoming an important modality for treatment of intoxications. Its use is limited to rare instances when hemodialysis is not available, when transport to a center with hemodialysis is not possible, and perhaps in some children who are too small to undergo hemodialysis in facilities that do not routinely dialyze children.

Trends in Use of Charcoal Hemoperfusion

Use of charcoal hemoperfusion, a modality traditionally used to enhance the elimination of highly protein-bound xenobiotics that display high charcoal adsorption, has

Advances in Extracorporeal Removal Techniques

175

body stores of some drugs to offer signicant clinical benet. The limitations of hemodialysis, which warrant caution in its application, must also be reviewed. Foremost among several is the obvious need for vascular access, most often accomplished in the acute setting by femoral cannulation with a double lumen hemodialysis catheter. Femoral cannulation has a low, but real, rate of complication, including bleeding and retroperitoneal hemorrhage. Despite the improvements highlighted earlier in the text in maintaining hemodynamic stability, even with care, blood pressure may decrease and make adequate dialysis difcult to deliver. Anticoagulation with heparin remains the standard regimen, and although low doses are often sufcient, the absolute avoidance of systemic anticoagulation is often not successful. Finally, the logistics are often not simple, requiring that if treatment during offhours is needed, a nephrologist must be consulted, a dialysis nurse called in from home, a machine prepared, and cannulation achieved, leading to delays that may affect the response to therapy and the prognosis.21 These possible complications, the logistical issues, and the cost, add up to the need for a thorough analysis before deciding that hemodialysis is indicated in the management of poisonings. Whether prognosis will be affected remains the last question to be decided. The fact that larger molecules can be removed through high-ux membranes does not mean that they should be removed if their removal will not change the patients course. In the following section, we review a variety of xenobiotics that might now be removable by hemodialysis and consider whether the available data in fact warrant their removal by hemodialysis in case of intoxication. Given the relative rarity of many of these poisonings, it is unlikely that randomized controlled trials will be performed to determine whether the use of any extracorporeal modality affects patient outcomes, and then which modality of ECR is best. In reporting cases or performing meta-analyses, careful measurement of clearances and detailed reporting of relevant variables should be emphasized. In addition, a healthy level of skepticism regarding the benets of such therapies should be maintained by the critical reader.

was too large to be dialyzed with conventional membranes, and treatment of overdose was limited primarily to forced diuresis.27 With a MW of 1485 Da and a volume of distribution of 0.2 to 1.25 L/kg, vancomycin is now amenable to removal by dialysis with high-ux membranes.28 The use of intermittent hemodialtration and CVVH with high-ux membranes has been reported in overdose.22,24 The removal of 60% to 67% of vancomycin from the intravascular compartment during high-ux hemodialysis has been reported, with a dialysis half-life of approximately 2 hours.27 Although this represents an improvement over previously reported removal by charcoal hemoperfusion,29 signicant postdialysis rebound of serum vancomycin concentrations was also demonstrated. Despite the effective clearance by highux hemodialysis, it remains unclear whether the clinical benets of enhanced extracorporeal elimination of vancomycin outweigh the risks of dialysis. It can be argued that the minimal toxicity of the drug,23,25,26 even at very high serum concentrations, does not justify the expense and safety concerns of hemodialysis, especially in the patient with functioning kidneys. In a patient with kidney failure caused by concomitant illness or preexisting chronic kidney disease, hemodialysis with high-ux membranes will result in decreased serum concentrations.

Carbamazepine
The anticonvulsant carbamazepine has a MW of 236 Da, a volume of distribution of 0.8 to 1.8 L/kg, and a protein binding of 75% to 78% in plasma. It is a triple-ringed tertiary amine, structurally similar to tricyclic antidepressants, and shares some of their adverse effects.30 In overdose with serum concentrations .16 mg/L, central nervous system depression, hypotension, tachycardia, cardiac conduction abnormalities, and QRS and QT interval prolongation (sequential EKG annotations [P-QR-S-T]) can occur.30,31 Massive overdose with serum concentrations .40 mcg/mL has been associated with life-threatening cardiac dysrhythmias, respiratory failure, coma, seizures, and death.32 Carbamazepine is a lipophilic drug with rapid tissue distribution after gastrointestinal absorption, which can be slow, unpredictable, and significantly delayed in the setting of massive ingestion.30 Carbamazepine is hepatically metabolized to an active epoxide metabolite, and displays zero-order kinetics at high serum concentrations. These factors contribute to a prolonged elimination half-life that can exceed 20 hours after acute ingestion.30,33 Noninvasive treatment of carbamazepine toxicity is sufcient in the vast majority of patients, and involves the use of multidose activated charcoal (MDAC) to reduce enterohepatic and enteroenteric circulation. One randomized trial showed a reduction in elimination halflife from 28 to 13 hours, along with corresponding

Indications for Dialysis of Selected Specic Toxins Vancomycin

Vancomycin is a commonly-prescribed antibiotic that is primarily excreted by the kidney. In the setting of deteriorating kidney function, decreased elimination of vancomycin can result in elevated serum concentrations. Inadvertent iatrogenic overdose has also been reported.22-24 Clinical manifestations of toxicity can include ototoxicity and nephrotoxicity, especially when used in combination with an aminoglycoside antibiotic.25,26 Historically, the vancomycin molecule

176

Garlich and Goldfarb

reduction in duration of coma, mechanical ventilation, and intensive care unit stay, with MDAC alone.34 However, in cases of severe, life-threatening toxicity, or when MDAC is unavailable or unfeasible, ECR may be of benet. Because of the high degree of protein binding, the treatment of carbamazepine toxicity has traditionally involved the use of charcoal hemoperfusion.12 The side effects associated with charcoal hemoperfusion, namely thrombocytopenia, coagulopathy, hypocalcemia, and hypothermia, prompted the exploration of the use of hemodialysis. The successful removal of carbamazepine by high-efciency hemodialysis (using a highly permeable high-ux membrane with increased dialysate ow) has been reported with comparable efcacy when compared with charcoal hemoperfusion.16,17 In one reported case, high-efciency hemodialysis in an 18-month-old comatose convulsing child with a serum carbamazepine concentration of 27 mcg/mL, resulted in a reduction in apparent elimination half-life from 17.3 to 2.7 hours with associated neurologic recovery during the 4.5 hour dialysis procedure.16 Several factors may explain the dialyzability of the relatively highly protein-bound carbamazepine. Its protein binding can be variable, and the proportion of free drug increases slightly at higher concentrations.35,36 The metabolite, carbamazepine 10,11-epoxide, is considered to be responsible for some of the clinical effects of toxicity, and has a lower protein binding of only 50%.37 Additionally, the improved effectiveness of drug clearance achievable by high-efciency hemodialysis may result in enough rapid clearance of free drug so that clinical improvement can occur.

hemoperfusion, but more recently as effective monotherapy.19,20 In one report, high-ux hemodialysis alone reduced the serum half-life from 23.41 to 2.74 hours, with concomitant improvement in mental function and hemodynamic status.19 Continuous venovenous hemodialtration is ineffective in severe toxicity.39 The majority of cases of VPA toxicity improve with supportive care and discontinuation of medications that affect hepatic VPA metabolism.30 L-carnitine is available as antidotal therapy for patients with evidence of hyperammonemia or hepatotoxicity. However, in patients with severe, lifethreatening toxicity demonstrating signicant neurologic and hemodynamic compromise, high-ux hemodialysis may be indicated. This therapy has the added benet of removing ammonia, and thus may be particularly useful in the setting of severe hyperammonemia after VPA overdose.30,40

Methotrexate
Methotrexate is a nephrotoxic chemotherapeutic agent that is commonly used in the treatment of several malignant, rheumatologic, dermatologic, and obstetric conditions. Its therapeutic effects are based on the inhibition of dihydrofolate reductase and thymidylate synthetase, which impair deoxyribonucleic acid (DNA) and ribonucleic acid (RNA) synthesis. Methotrexate is primarily excreted unchanged in the urine.41-43 At high doses, methotrexate and its insoluble metabolites accumulate and can precipitate in renal tubules, causing acute tubular necrosis. High dosage methotrexate regimens, which are used in treatment of certain neoplasms at doses of a maximum of 1000 mg/m2, are associated with impaired kidney function in 48%,44,45 and signicant nephropathy in 1.8% of the patients.43 Impaired kidney function can lead to decreased drug clearance, which causes drug accumulation and amplication of toxic effects. In addition to nephrotoxicity, these adverse effects include gastritis, hepatitis, myelosuppression, and central nervous system dysfunction.41-43 Methotrexate displays a triphasic plasma clearance. After an initial plasma distribution half-life of 0.75 hours, the second phase, representing renal clearance, typically has a half-life of 2 to 3 hours in the setting of normal kidney function. In the third phase, extensive tissue redistribution takes place with a half-life of 8 to 10 hours. Once intracellular, methotrexate is metabolized to polyglutamate derivatives that do not easily diffuse out of the cell, causing persistent cytotoxic effects and limiting attempts at drug removal.46,47 Management strategies to reduce toxicity include gastrointestinal decontamination with activated charcoal in single or multiple doses, aggressive hydration, urinary alkalinization, leucovorin rescue, or carboxypeptidaseG2 ([CPDG2] glucarpidase). CPDG2 is an investigational recombinant bacterial enzyme that hydrolyzes

Valproic Acid
VPA is a carboxylic acid anticonvulsant that has been increasingly prescribed for neurologic and psychiatric disorders, resulting in a dramatic increase in unintentional and intentional exposures since its introduction in 1978.38 Clinical manifestations of VPA toxicity include central nervous system depression, hyperammonemic encephalopathy, coma, cerebral edema, bone marrow suppression, metabolic derangements, and rarely, hepatotoxicity, pancreatitis, and impaired kidney function.30,38 VPA has a MW of 144 Da, and a volume of distribution of 0.1 to 0.4 L/kg,30 but its high protein binding (.90%) at therapeutic concentrations and poor water solubility led to previous assumptions that ECR was limited to charcoal hemoperfusion.18 However, at supratherapeutic concentrations, protein binding sites become saturated, leading to an increasing proportion of unbound drug in the serum that is amenable to hemodialysis. At serum concentrations .300 mg/mL, for example, only 35% of the drug is protein-bound.19,38 High-ux hemodialysis has been used successfully to increase VPA clearance, initially in combination with

Advances in Extracorporeal Removal Techniques

177

methotrexate to inactive metabolites, and can lower plasma concentrations by 70% to 99% within minutes when administered to patients with methotrexateinduced nephrotoxicity.43 It remains unapproved by the United States Food and Drug Administration (FDA) but is available from the manufacturer through compassionate use requests. Given the high potential for toxicity, methods of ECR have been extensively explored. Methotrexate has a MW of 454 Da, a volume of distribution of 0.4 to 0.9 L/kg,45,47 and 50% plasma protein binding,45,48 suggesting its dialyzability. Previous reports documented that methotrexate clearance by conventional hemodialysis was poor or ineffective,42,49 but several recent investigations showed the effective removal of methotrexate by high-ux hemodialysis. In a review of all dialysis-based methods of methotrexate removal between 1977 and 2002, high-ux hemodialysis resulted in the greatest decrease in serum concentrations (median, 75.7%; range, 42%-94%) within the shortest time (median, 4 hours; range, 4-12 hours).43 In one case series, the mean plasma clearance of methotrexate achieved with high-ux hemodialysis was 96.3 6 10.5 mL/min, which represented a 4-fold increase over renal excretion over the same time interval.42 However, because of redistribution from tissues, signicant rebound of serum concentrations after dialysis is nearly universal (a maximum of 221% of postdialysis concentrations has been reported43), making a prolonged course of recurrent intermittent hemodialysis41-43 or continuous venovenous hemodialysis45 a necessity. The availability of other, highly effective treatments of methotrexate toxicity, most notably leucovorin and CPDG2, combined with postdialysis rebound after tissue redistribution greatly limit the utility of hemodialysis in methotrexate toxicity. In patients with signicantly impaired kidney function, those who cannot tolerate aggressive intravenous uid administration, and in situations in which leucovorin and CPDG2 are unavailable, early high-ux hemodialysis before tissue distribution may be considered. Thus, it is important to emphasize that even in xenobiotics with large volumes of distribution, ECR may be effective after an acute exposure or ingestion while the drug remains in the plasma, before complete tissue distribution. In the case of methotrexate, this gives an effective window that is limited to a period of several hours, often rendering dialysis unrealistic or impractical later. If delayed hemodialysis is instituted after a period of several hours after acute exposure, postdialysis rebound and a prolonged course of treatment should be anticipated.

modalities (dextrose therapy) are far easier and less invasive. Because of its larger MW of 5800 to 6000 Da, insulin was not cleared by low-ux dialysis, but the larger pore size of high-ux membranes allowed insulin clearance.51,52 The clinical applicability of this is limited to situations in which a patient already on hemodialysis for chronic kidney disease intentionally or inadvertently is administered supratherapeutic doses of insulin. In cases of refractory hypoglycemia, early high-ux hemodialysis may prevent the hypervolemia associated with repeated doses of concentrated dextrose.

Myoglobin
Many poisoned patients develop rhabdomyolysis, often in the setting of severe serotonin syndrome, neuromuscular malignant syndrome, cocaine intoxication, or sedativehypnotic overdose with prolonged down-time, and rarely in salicylate toxicity and Tricholoma equestre mushroom toxicity.53 The release of myoglobin from muscle breakdown can cause nephrotoxicity by producing renal tubular obstruction, lipid peroxidation within tubular cells, and renal vasoconstriction.54 Treatment by massive intravenous volume expansion and forced diuresis are frequently used, but renal replacement therapies are often instituted when kidney function declines. Previous attempts to remove myoglobin from the intravascular compartment using conventional hemodialysis55 and continuous venovenous hemodialtration56,57 were met with very limited success. This should be of little surprise, given that myoglobin is a large molecule with a MW of 17,000 Da. Effective and relatively rapid clearance of myoglobin by hemoltration with a hyperpermeable super-high-ux membrane (molecular cutoff point, 100,000 Da) has been reported in a case of acute rhabdomyolysis caused by serotonin syndrome.58 A reduction in serum myoglobin concentration from .100,000 mg/L to 16,542 mg/L was achieved in 48 hours, representing a clearance that was 5 times greater than that achieved with conventional hemoltration in the same patient.

Conclusion
In essence, despite the advent of high-ux membranes and high-efciency techniques, there has been relatively little change in the indications for dialysis. There are several xenobiotics that can now be dialyzed, but the clinical benet of doing so is either questionable or highly limited to situations of severe, life-threatening toxicity when other effective, less-invasive treatment modalities are unavailable or unfeasible. These situations are rare. Examples include prolonged, life-threatening carbamazepine toxicity when MDAC is unavailable or ineffective, severe VPA toxicity with hemodynamic compromise, and possibly the early treatment of life-threatening

Insulin
Hemodialysis is not traditionally used as a treatment for insulin overdose,50 primarily because other treatment

178

Garlich and Goldfarb

methotrexate toxicity in a patient with kidney failure when rescue therapies are unavailable. The use of hemodialysis in the latter scenario is, understandably, controversial. In each of these clinical settings, the risks associated with hemodialysis must be weighed against the potential benet of enhanced extracorporeal elimination, on an individual case-by-case basis. It is worth emphasizing that more effective clearance does not necessarily equal improved outcomes. Thus, although improvements in middle molecule clearance may be important to the benets of chronic hemodialysis for the patient with end-stage kidney disease, the importance of increased large molecule clearance to the management of poisoning is less dramatic and perhaps more controversial. Only carefully analyzed clinical data, even if short of a randomized clinical trial, can reveal whether these changes in technology have positively inuenced care. The enhanced clearance of certain drugs in the patient who is already on maintenance hemodialysis for chronic kidney disease poses a unique clinical situation. Iatrogenic, intentional, or inadvertent overdoses of larger MW drugs like vancomycin and insulin may be effectively treated with earlier or routine high-ux hemodialysis. However, the same overdose in a patient with functioning kidneys could be adequately treated with supportive, noninvasive care. The use of super high-ux dialysis membranes to treat the myoglobinemia associated with rhabdomyolysis is promising, and warrants further clinical exploration. In general, the improved hemodynamic prole that is associated with newer hemodialysis technologies is benecial to all patients who require dialysis. Poisoned patients benet from these improvements as well, most signicantly by making dialysis safer in the patient with hemodynamic compromise. An improvement in efciency of drug clearance may in part explain the ability of high-ux dialysis to clear highly protein-bound drugs like carbamazepine that were previously only amenable to charcoal hemoperfusion. In cases of severe toxicity, a small reduction in serum concentration may be sufcient to result in clinical improvement. The indications for hemodialysis of some particularly dangerous toxins (such as salicylates, toxic alcohols, and lithium) have not changed. Innovations in ECR techniques have not affected the already excellent clearance rates of those xenobiotics. Despite the fact that hemodialysis can be life-saving or morbidity-reducing, its underutilization (eg, in salicylate intoxication) is probably a greater limitation than any aspect of the techniques efcacy.21 Perhaps the improved safety prole will increase the use in poisoned patients who are critically ill. Instead of focusing on the technology of dialysis membranes, attention should be focused on improving access to dialysis in a timely manner in cases in which it will have the most clinical effect.

References
1. Holubek WJ, Hoffman RS, Goldfarb DS, et al. Use of hemodialysis and hemoperfusion in poisoned patients. Kidney Int. 2008;74:1327-1334. 2. Fertel BS, Nelson LS, Goldfarb DS. Extracorporeal removal techniques for the poisoned patient: a review for the intensivist. J Intensive Care Med. 2010;25:139-148. 3. Lameire N, Van BW, Vanholder R. Did 20 years of technological innovations in hemodialysis contribute to better patient outcomes? Clin J Am Soc Nephrol. 2009;4(suppl 1):S30-S40. 4. Palmer BF. Effectiveness of hemodialysis in the extracorporeal therapy of phenobarbital overdose. Am J Kidney Dis. 2000;36:640-643. 5. Ambalavanan S, Rabetoy GM, Cheung AK. High-efciency and high-ux hemodialysis. In: Henrich W, Bennet W, eds. Philadelphia, PA: Current Medicine; 1999:1-3. 10. 6. Bowry SK. Dialysis membranes today. Int J Artif Organs. 2002;25:447-460. 7. Cheung AK, Levin NW, Greene T, et al. Effects of high-ux hemodialysis on clinical outcomes: results of the HEMO study. J Am Soc Nephrol. 2003;14:3251-3263. 8. Diamond SM, Henrich WL. Acetate dialysate versus bicarbonate dialysate: a continuing controversy. Am J Kidney Dis. 1987;9:3-11. 9. Luo FJ, Patel KP, Marquez IO, et al. Effect of increasing dialyzer mass transfer area coefcient and dialysate ow on clearance of protein-bound solutes: a pilot crossover trial. Am J Kidney Dis. 2009;53:1042-1049. 10. Thomas G, Jaber BL. Convective therapies for removal of middle molecular weight uremic toxins in end-stage renal disease: a review of the evidence. Semin Dial. 2009;22:610-614. 11. van de PA, Stolk L, Verhoeven MA, et al. Successful treatment of acute phenobarbital intoxication by hemodialtration. Clin Toxicol (Phila). 2006;44:93-94. 12. Kim Z, Goldfarb DS. Continuous renal replacement therapy does not have a clear role in the treatment of poisoning. Nephron Clin Pract. 2010;115:c1-c6. 13. Shalkham AS, Kirrane BM, Hoffman RS, et al. The availability and use of charcoal hemoperfusion in the treatment of poisoned patients. Am J Kidney Dis. 2006;48:239-241. 14. Litovitz TL, Klein-Schwartz W, Caravati EM, et al. 1998 annual report of the American Association of Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg Med. 1999;17:435-487. 15. Bronstein AC, Spyker DA, Cantilena LR Jr, et al. 2006 Annual Report of the American Association of Poison Control Centers National Poison Data System (NPDS). Clin Toxicol (Phila). 2007;45:815-917. 16. Schuerer DJ, Brophy PD, Maxvold NJ, et al. High-efciency dialysis for carbamazepine overdose. J Toxicol Clin Toxicol. 2000;38:321323. 17. Kielstein JT, Schwarz A, Arnavaz A, et al. High-ux hemodialysis an effective alternative to hemoperfusion in the treatment of carbamazepine intoxication. Clin Nephrol. 2002;57:484-486. 18. Dharnidharka VR, Fennell RS III, Richard GA. Extracorporeal removal of toxic valproic acid levels in children. Pediatr Nephrol. 2002;17:312-315. 19. Kane SL, Constantiner M, Staubus AE, et al. High-ux hemodialysis without hemoperfusion is effective in acute valproic acid overdose. Ann Pharmacother. 2000;34:1146-1151. 20. Johnson LZ, Martinez I, Fernandez MC, et al. Successful treatment of valproic acid overdose with hemodialysis. Am J Kidney Dis. 1999;33:786-789. 21. Fertel BS, Nelson LS, Goldfarb DS. The underutilization of hemodialysis in patients with salicylate poisoning. Kidney Int. 2009;75:1349-1353. 22. Akil IO, Mir S. Hemodialtration for vancomycin overdose in a patient with end-stage renal failure. Pediatr Nephrol. 2001;16:1019-1021.

Advances in Extracorporeal Removal Techniques

179

23. Miner LJ, Faix RG. Large vancomycin overdose in two premature infants with minimal toxicity. Am J Perinatol. 2004;21:433-438. 24. Goebel J, Ananth M, Lewy JE. Hemodialtration for vancomycin overdose in a neonate with end-stage renal failure. Pediatr Nephrol. 1999;13:423-425. 25. Mellor JA, Kingdom J, Cafferkey M, et al. Vancomycin toxicity: a prospective study. J Antimicrob Chemother. 1985;15:773-780. 26. Rybak MJ, Albrecht LM, Boike SC, et al. Nephrotoxicity of vancomycin, alone and with an aminoglycoside. J Antimicrob Chemother. 1990;25:679-687. 27. Bunchman TE, Valentini RP, Gardner J, et al. Treatment of vancomycin overdose using high-efciency dialysis membranes. Pediatr Nephrol. 1999;13:773-774. 28. Quale JM, OHalloran JJ, DeVincenzo N, et al. Removal of vancomycin by high-ux hemodialysis membranes. Antimicrob Agents Chemother. 1992;36:1424-1426. 29. Panzarino VM, Feldstein TJ, Kashtan CE. Charcoal hemoperfusion in a child with vancomycin overdose and chronic renal failure. Pediatr Nephrol. 1998;12:63-64. 30. Doyon S. Anticonvulsants. In: Nelson LS, Lewin NA, Howland MA, Hoffman RS, Goldfrank LR, Flomenbaum NE, eds. Goldfranks Toxicologic Emergencies. New York, NY: McGraw-Hill; 2011:698-710. 31. Apfelbaum JD, Caravati EM, Kerns WP, et al. Cardiovascular effects of carbamazepine toxicity. Ann Emerg Med. 1995;25:631-635. 32. Hojer J, Malmlund HO, Berg A. Clinical features in 28 consecutive cases of laboratory conrmed massive poisoning with carbamazepine alone. J Toxicol Clin Toxicol. 1993;31:449-458. 33. Winnicka RI, Topacinski B, Szymczak WM, et al. Carbamazepine poisoning: elimination kinetics and quantitative relationship with carbamazepine 10,11-epoxide. J Toxicol Clin Toxicol. 2002; 40:759-765. 34. Brahmi N, Kouraichi N, Thabet H, et al. Inuence of activated charcoal on the pharmacokinetics and the clinical features of carbamazepine poisoning. Am J Emerg Med. 2006;24:440-443. 35. Spiller HA. Management of carbamazepine overdose. Pediatr Emerg Care. 2001;17:452-456. 36. Hooper WD, Dubetz DK, Bochner F, et al. Plasma protein binding of carbamazepine. Clin Pharmacol Ther. 1975;17:433-440. 37. Vree TB, Janssen TJ, Hekster YA, et al. Clinical pharmacokinetics of carbamazepine and its epoxy and hydroxy metabolites in humans after an overdose. Ther Drug Monit. 1986;8:297-304. 38. Sztajnkrycer MD. Valproic acid toxicity: overview and management. J Toxicol Clin Toxicol. 2002;40:789-801. 39. Kay TD, Playford HR, Johnson DW. Hemodialysis versus continuous veno-venous hemodialtration in the management of severe valproate overdose. Clin Nephrol. 2003;59:56-58. 40. Al AZ, Yalamanchili P, Gonzalez E. Extracorporeal management of valproic acid toxicity: a case report and review of the literature. Semin Dial. 2005;18:62-66.

41. Saland JM, Leavey PJ, Bash RO, et al. Effective removal of methotrexate by high-ux hemodialysis. Pediatr Nephrol. 2002;17:825-829. 42. Wall SM, Johansen MJ, Molony DA, et al. Effective clearance of methotrexate using high-ux hemodialysis membranes. Am J Kidney Dis. 1996;28:846-854. 43. Widemann BC, Balis FM, Kempf-Bielack B, et al. High-dose methotrexate-induced nephrotoxicity in patients with osteosarcoma. Cancer. 2004;100:2222-2232. 44. Green MR, Chamberlain MC. Renal dysfunction during and after high-dose methotrexate. Cancer Chemother Pharmacol. 2009; 63:599-604. 45. Vilay AM, Mueller BA, Haines H, et al. Treatment of methotrexate intoxication with various modalities of continuous extracorporeal therapy and glucarpidase. Pharmacotherapy. 2010;30:111. 46. Chabner BA, Allegra CJ, Curt GA, et al. Polyglutamation of methotrexate. Is methotrexate a prodrug? J Clin Invest. 1985;76:907-912. 47. Wang RY. Antineoplastics: methotrexate. In: Nelson LS, Lewin NA, Howland MA, Hoffman RS, Goldfrank LR, Flomenbaum NE, eds. Goldfranks Toxicologic Emergencies. New York, NY: McGraw-Hill; 2011:778-782. 48. Bunchman TE, Ferris ME. Management of toxic ingestions with the use of renal replacement therapy. Pediatr Nephrol. In press. 49. Howell SB, Blair HE, Uren J, et al. Hemodialysis and enzymatic cleavage of methotrexate in man. Eur J Cancer. 1978;14:787-792. 50. Roberge RJ, Martin TG, Delbridge TR. Intentional massive insulin overdose: recognition and management. Ann Emerg Med. 1993;22:228-234. 51. Abe M, Kikuchi F, Kaizu K, et al. The inuence of hemodialysis membranes on the plasma insulin level of diabetic patients on maintenance hemodialysis. Clin Nephrol. 2008;69:354-360. 52. Abe M, Kaizu K, Matsumoto K. Plasma insulin is removed by hemodialysis: evaluation of the relation between plasma insulin and glucose by using a dialysate with or without glucose. Ther Apher Dial. 2007;11:280-287. 53. Nieminen P, Mustonen AM, Kirsi M. Increased plasma creatine kinase activities triggered by edible wild mushrooms. Food Chem Toxicol. 2005;43:133-138. 54. Zager RA. Rhabdomyolysis and myohemoglobinuric acute renal failure. Kidney Int. 1996;49:314-326. 55. Hart PM, Feinfeld DA, Briscoe AM, et al. The effect of renal failure and hemodialysis on serum and urine myoglobin. Clin Nephrol. 1982;18:141-143. 56. Nicolau D, Feng YS, Wu AH, et al. Myoglobin clearance during continuous veno-venous hemoltration with or without dialysis. Int J Artif Organs. 1998;21:205-209. 57. Ronco C. Extracorporeal therapies in acute rhabdomyolysis and myoglobin clearance. Crit Care. 2005;9:141-142. 58. Naka T, Jones D, Baldwin I, et al. Myoglobin clearance by super high-ux hemoltration in a case of severe rhabdomyolysis: a case report. Crit Care. 2005;9:R90-R95.