10: Principles and Techniques Applied to Enhance Elimination

Enhancing the elimination of a xenobiotic from a poisoned patient is a logical step after techniques to inhibit absorption such as orogastric lavage, activated charcoal, or whole-bowel irrigation have been considered. Table 10–1 lists methods that might be used to enhance elimination. Some of these techniques are described in more detail in chapters that deal with specific xenobiotics. In this chapter, hemodialysis, hemoperfusion, and hemofiltration are considered extracorporeal treatments because xenobiotic removal occurs in a blood circuit outside the body. Currently, these methods are used infrequently because most poisonings are not amenable to removal by these methods. In addition, because these elimination techniques have associated adverse effects and complications, the risk-benefit analysis suggests a benefit in a relatively small proportion of patients.

Although undoubtedly an underestimate of true use, enhancement of elimination was used relatively infrequently in a cohort of approximately 2.4 million patients reported by the American Association of Poison Control Centers (AAPCC) National Poison Data System (NPDS) in 2011 (Chap. 136).¹² Alkalinization of the urine was reportedly used 10,843 times, multiple-dose activated charcoal (MDAC) 1904 times, hemodialysis 2323 times, and hemoperfusion 14 times. As in the past, there continue to be many instances of the use of extracorporeal treatments that we consider inappropriate, such as in the treatment of overdoses of cyclic antidepressants (CAs).¹²

Although data reporting remains important in comparing the most recent data with past reports (Table 10–2), there is a continued increase in the reported use of hemodialysis, paralleling a decline in reports of charcoal hemoperfusion (Chap. 136). Lithium and ethylene glycol were the most common xenobiotics for which hemodialysis was used between 1985 and 2005. Possible reasons for the decline in use of charcoal and resin hemoperfusion are described in the section Hemoperfusion below. Peritoneal dialysis (PD), a slower modality that should have little or no role in any poisoning, is no longer separately reported (Chap. 136). “Other extracorporeal procedures” in past AAPCC reports may include continuous modalities (discussed below in the section Continuous Hemofiltration and Hemodiafiltration), plasmapheresis, and PD.

Very few prospective, randomized, controlled clinical trials have been conducted to determine which groups of patients actually benefit from enhanced elimination of various xenobiotics and which modalities are most efficacious. For most poisonings, it is unlikely that such studies will ever be performed, given the relative scarcity of appropriate cases of sufficient severity and because of the many variables that would hinder controlled comparisons. Thus, limited evidence predominates. We must therefore rely on an understanding of the principles of these methods to identify the individual patients for whom enhanced elimination is indicated. Isolated case reports in which the kinetics are studied before, during, and after enhanced elimination are also very useful in establishing the efficacy of a method. Fortunately, in the absence of robust evidence, consensus-based recommendations have now been published and are being developed to guide clinical decisions. The American Academy of Clinical Toxicology (AACT) and the European Association of Poisons Centres and Clinical Toxicologists (EAPCCT) published joint position papers on urine alkalinization and MDAC. The Extracorporeal Treatments in Poisoning (EXTRIP) work group, a collaboration of experts from diverse specialties (clinical toxicology, nephrology, pharmacology, critical care, emergency medicine) and represented by more than 30 international societies, is now reviewing the indications for dialysis and other treatments for overdose. Guidelines for several poisons are expected in 2014. See http://extrip-workgroup.org.

Enhanced elimination may be indicated for several types of patients:

• Patients who fail to respond adequately to comprehensive supportive care. Such patients may have intractable hypotension, heart failure, seizures, metabolic acidosis, or dysrhythmias. Hemodialysis and hemoperfusion (although infrequently used) are much better tolerated than in the past and may represent potentially life-saving opportunities for patients with life-threatening toxicity caused by theophylline, lithium, salicylates, or toxic alcohols.

• Patients in whom the normal route of elimination of the xenobiotic is impaired. Such patients may have kidney or hepatic dysfunction, either preexisting or caused by the overdose. For example, a patient with chronic kidney disease associated with long-term lithium use is more likely to develop toxicity and require hemodialysis as treatment.

• Patients in whom the amount of xenobiotic absorbed or its high concentration in serum indicates that serious morbidity or mortality is likely. Such patients may not appear acutely ill on initial evaluation. Xenobiotics in this group may include ethylene glycol, lithium, methanol, paraquat, salicylates, and theophylline.

• Patients with concurrent disease or in an age group (very young or old) associated with increased risk of morbidity or mortality from the ­overdose. Such patients are intolerant of prolonged coma, immobility, and hemodynamic instability. An example is a patient with both severe underlying respiratory disease and chronic salicylate poisoning.

• Patients with concomitant electrolyte disorders that could be corrected with hemodialysis. An example is the metabolic acidosis with elevated lactate associated with metformin toxicity discussed in the Hemodialysis section of this chapter.

Ideally, these techniques will be applied to poisonings for which studies suggest an improvement in outcome in treated patients compared with patients not treated with extracorporeal removal. As previously mentioned, these data are rarely available.^22,26

The need for extracorporeal elimination is less clear for patients who are poisoned with xenobiotics that are known to be removed by the various modalities of treatment but that cause limited morbidity if supportive care is provided. Relatively high rates of endogenous clearance would also make extracorporeal elimination redundant. Examples of such xenobiotics include ethanol and some barbiturates. Both are subject to substantial rates of hepatic metabolism, and neither would be expected to lead to significant morbidity after the affected patient has had endotracheal intubation and is mechanically ventilated. There may be instances of severe toxicity from these two xenobiotics for which enhanced elimination will reduce the length of intensive care unit (ICU) stays and the associated nosocomial risks; extracorporeal elimination may then be a reasonable option.^8,50 Dialysis should be avoided if other more effective modalities are available. For example, a patient with an acetaminophen (APAP) overdose should be initially treated with N-acetylcysteine instead of hemodialysis.

The appropriateness of any modality for increasing the elimination of a given xenobiotic depends on various properties of the molecules in question. Effective removal by the extracorporeal procedures and other methods listed in Table 10–1 is limited by a large volume of distribution (Vd). The Vd relates to the concentration of the xenobiotic in the blood or serum to the total body burden. The Vd can be envisioned as the apparent volume in which a known total dose of drug is distributed before metabolism and excretion occur:

The larger the Vd, the less the xenobiotic is available in the blood compartment for elimination. A xenobiotic with a relatively small Vd, considered amenable to extracorporeal elimination, would distribute in an apparent volume not much larger than total body water (TBW). TBW is approximately 60% of total body weight, so a Vd equal to TBW is approximately 0.6 L/kg body weight.

Ethanol is an example of a xenobiotic with a small Vd approximately equal to TBW. A substantial fraction of a dose of ethanol could be removed by hemodialysis. In contrast, an insignificant fraction of digoxin with a large Vd (5–12 L/kg of body weight) would be removed by hemodialysis. Lipid-soluble xenobiotics have large volumes of distribution, which typically exceed TBW or even total body weight. These high apparent volumes of distribution imply that the xenobiotic is not available to extracorporeal removal because only a small portion would be in the blood and therefore the extracorporeal circuit. In addition to the alcohols, other xenobiotics with a relatively low Vd include phenobarbital, lithium, salicylates, valproic acid, bromide and fluoride ions, and theophylline. Conversely, those with a high Vd (≥1 L/kg of body weight), which would not be removed substantially by hemodialysis, include many β-adrenergic antagonists (with the possible exception of atenolol⁶⁰), diazepam, organic phosphorus compounds, ­phenothiazines, quinidine, and the CAs.

Pharmacokinetics also influence the ability to enhance elimination of a xenobiotic. Kinetic parameters after an overdose may differ from those after therapeutic or experimental doses. For instance, carrier- or enzyme-­mediated elimination processes may be overwhelmed by higher concentrations of the xenobiotic in question, making extracorporeal removal potentially more useful. Similarly, plasma protein- and tissue-binding sites may all be saturated at higher concentrations, making extracorporeal removal feasible in instances in which it would have no role in less significant overdoses. An example is valproic acid, which may be poorly dialyzed at nontoxic concentrations because of high protein binding. However, at higher concentrations, protein-binding sites become saturated and lead to a higher proportion of the drug free in the serum, amenable to removal by hemodialysis at a clinically relevant rate.³⁴ Estimated endogenous elimination of a xenobiotic should be derived by proper toxicokinetic models and not by pharmacokinetic data after therapeutic doses.

When assessing the efficacy of any technique of enhanced elimination, a generally accepted principle is that the intervention is worthwhile only if a large portion of total body drug burden can be eliminated by extracorporeal removal or if total body clearance of the xenobiotic is increased by a ­factor of 2.⁴¹ This substantial increase is easier to achieve when the xenobiotic has a low endogenous clearance. Examples of xenobiotics with low endogenous clearances (<4 mL/min/kg) include valproic acid, lithium, paraquat, ­phenytoin, salicylate, and theophylline. Xenobiotics with high endogenous clearances include many β-adrenergic antagonists, lidocaine, opioids, nicotine, and CAs. Enhancement of elimination is expected to contribute more to overall clearance of the former group than to the latter.

The efficacy of any technique of elimination can be directly assessed by comparing the blood or serum concentrations of the xenobiotic at the beginning and the end of the procedure.⁶⁸ For example, a xenobiotic such as theophylline, with one-compartment kinetics, is essentially limited to the extracellular space. The difference between the theophylline concentration before the procedure, minus the concentration at the end of the procedure, divided by the concentration at the beginning, is the fraction of the body burden of the xenobiotic that is eliminated. However, this calculation would not be appropriate to estimate the effect of treatment on the total body burden of a xenobiotic that distributes in a larger Vd or with multicompartment kinetics because the changes in blood concentration may not reflect the effect of the treatment on the total amount in other compartments.

Certain xenobiotics, such as lithium, distribute in part to the intracellular compartment, and equilibrium is established between the intracellular and extracellular compartments. The latter compartment includes the blood from which xenobiotic elimination occurs and the interstitium. An increase in the elimination rate from the extracellular compartment, such as by hemodialysis, alters this equilibrium. The rate of redistribution of the xenobiotic from the intracellular compartment into the now-­dialyzed extracellular compartment may be slower than the rate of clearance across the dialysis membrane. In that case, the serum concentration may become relatively low despite a substantial intracellular burden. This low serum concentration reduces the concentration gradient for diffusion from serum to dialysate so that dialysance (the dialysis clearance) is reduced. Thus, although serum concentrations may decrease precipitously during the procedure, the total-body burden of the xenobiotic may not be affected significantly if it does not move from cells to the dialyzable extracellular volume. An example is a 60-kg patient who ingests 100 (25-mg) amitriptyline tablets, a CA.²² Assuming that the drug is fully absorbed, the 2500-mg distributes in an apparent volume of 40 L/kg of body weight to achieve a serum concentration of 1000 ng/mL, a potentially toxic concentration. If charcoal hemoperfusion is performed with a blood flow rate of 350 mL/min (plasma flow rate of 200 mL/min, assuming a hematocrit of approximately 43%) and the extraction ratio is 100% (see Charcoal Hemoperfusion later), the clearance of drug is 200 mL/min or 200 µg/min. In 4 hours (240 minutes) of treatment, only 48 mg (48,000 µg), or less than 2% of the amount ingested, will be removed. The treatment would not have affected toxicity.

Further evidence that a xenobiotic is in a slowly equilibrating compartment is provided by monitoring the serum concentrations after discontinuation of the procedure and demonstrating rebound. This rebound indicates postdialysis redistribution (Fig. 10–1). The magnitude of the rebound depends on total-body stores of the xenobiotic. A postdialysis increase in blood concentration may also result from ongoing absorption from the gastrointestinal (GI) tract.

When assessing overall efficacy, the evidence of enhanced elimination must be considered in the context of the clinical response. In some instances, observed improvement is not predicted by the kinetics of the parent xenobiotic and may not result from the intervention. For example, in severe cases of CA poisoning, unexpected improvement during hemoperfusion could be fortuitous because the toxicity is manifested early and ameliorates rapidly during the initial distribution phase.

Beginning a procedure during the initial distribution phase may increase the fraction of the drug burden that can be removed. Later, after the xenobiotic has distributed into fat or has been bound by plasma or tissue proteins, administration of extracorporeal treatment may have much lower clearance rates and benefits.

The extent of xenobiotic removal by enhanced elimination is not necessarily correlated to clinical improvement. For example, enhanced elimination of phenobarbital by MDAC did not affect clinical outcomes compared with those of overdosed, untreated control subjects despite a decrease in their serum concentrations.⁵³ In another study, no difference was observed between patients with lithium poisoning for whom hemodialysis was done and those for whom it was recommended by a poison control center but not done.⁹ Unfortunately, the study was underpowered and limited by confounding by indication. Similarly, in APAP toxicity, the efficacy of administration of N-acetylcysteine makes extracorporeal removal usually unnecessary. Furthermore, although paraquat exhibits all the chemical characteristics necessary for high extracorporeal removal, the outcome of patients presenting late after paraquat ingestion remains dismal despite active elimination procedures.

Occasionally, there may be unanticipated evidence of improvement despite removal of marginal amounts of the xenobiotic and active metabolites. Some authors have attributed this surprising finding to removal of a critical amount of xenobiotic from a shallow “toxic effect” compartment. This ­theory has been advanced to explain the response to hemodialysis of patients overdosed with the antipsychotic chlorprothixene³⁹ or with a combination of diltiazem and metoprolol.³ Such effects may lead to transient improvements that are not sustained as drug redistributes from one pool to another, leading to early benefit but eventual recurrence of symptoms. Much of the relevant literature fails to provide long-term follow-up to demonstrate prolongation of benefits after extracorporeal therapy is completed and xenobiotics have redistributed.

Although the efficacy of or need for removal of many xenobiotics remains controversial, consensus regarding the indications for a number of procedures has developed. This consensus has led to consistent application of several techniques of enhanced elimination for some toxic exposures that occur relatively more frequently. The techniques to enhance xenobiotic elimination most commonly applied over the past decade have been alkalinization of the urine for salicylates and hemodialysis for methanol, ethylene glycol, lithium, and salicylates.

Multiple-Dose Activated Charcoal: “Gastrointestinal Dialysis”

Oral administration of multiple doses of activated charcoal increases elimination of some xenobiotics present in the blood. This modality is discussed in more detail in Antidotes in Depth: A1 and will not be discussed here.

Resins

Resins are sometimes used in poisoning management. They can reduce bioavailability of ingested drugs and act as decontaminants, similarly to activated charcoal. In addition, however, they can enhance elimination of certain xenobiotics by enhancing their back-diffusion from plasma to gut (GI dialysis) and interrupt enterohepatic recirculation.

The most commonly used resins are sodium polystyrene sulfonate (Kayexalate), cholestyramine, and Prussian blue. Sodium polystyrene sulfonate is an ion exchanger that is used regularly for hyperkalemia in patients with chronic kidney disease. Data now exist on its potential use in lithium poisoning, although treatment is limited by hypokalemia.^25,43 Prussian blue is also an ion exchanger used for treatment of thallium poisoning (Antidotes in Depth: A28). Cholestyramine, a bile acid sequestrant, may bind several xenobiotics, including digoxin, ibuprofen, and mycophenolate mofetil, although its application in poisoning is doubtful (Chap. 65).

Forced Diuresis and Manipulation of Urinary pH

Forced diuresis by volume expansion with isotonic sodium–containing solutions, such as 0.9% sodium chloride and lactated Ringer (LR) solution with or without the addition of a diuretic, may increase renal clearance of some molecules. This therapy would theoretically be most useful for xenobiotics such as lithium for which the glomerular filtration rate (GFR), which is the volume of plasma filtered across the glomerular basement membrane per minute, is an important determinant of excretion. In people with normal extracellular fluid (ECF) volume who have not had loss of sodium via renal, GI, or other routes of excretion, the increase in GFR expected with plasma volume expansion is variable and unpredictable and may not lead to significant increases in xenobiotic elimination. The effect is potentially more important in patients who have had contraction of the ECF volume because of sodium loss. Loss of extracellular volume leads to a reduction of GFR partly as a result of decreased cardiac preload and cardiac output, which, in turn, reduces renal plasma flow. This circumstance is also accompanied by activation of angiotensin II, a small peptide that acts as a pressor and stimulates sodium reabsorption in the proximal tubule. Because small molecules such as lithium are both filtered at the glomerulus and reabsorbed by the proximal tubule, especially when sodium depletion has occurred and angiotensin II has been activated, repletion of ECF volume with 0.9% sodium chloride will increase GFR and suppress sodium reabsorption. The result is an increase in excretion of low-molecular-weight xenobiotics such as lithium. After the ECF volume is restored, the continued infusion of 0.9% sodium chloride or LR increases urine volume proportionally more than GFR, which may increase excretion of some small molecules such as urea, but which has little efficacy in the case of most poisonings.

The significant risk of excessive volume repletion is ECF volume overload, manifested by pulmonary and cerebral edema. This complication may be particularly likely in patients with long-standing lithium use in whom chronic tubulointerstitial disease may lead to chronic kidney disease that does not improve with fluid therapy. Other patients with acute kidney injury not mediated by ECF volume depletion are also at risk. Knowing the result of past serum creatinine concentrations may help distinguish acute from chronic kidney disease in such cases. Administration of diuretics such as furosemide along with saline may diminish the risk of ECF volume overload but may complicate the therapy, confuse the assessment of ECF volume, and increase the risk of metabolic alkalosis and hypokalemia. The unproven efficacy of forced diuresis in the management of any overdose has led most experts to abandon its use. On the other hand, the repletion of ECF volume when volume contraction is present, as determined by the history and physical examination, is, of course, appropriate.

Many xenobiotics are weak acids or bases that are ionized in aqueous solution to an extent that depends on the pK_a of the xenobiotic and the pH of the solution. Knowing these variables, the Henderson-Hasselbalch equation (Chap. 9) may be used to determine the relative proportions of the acids, bases, and buffer pairs. Whereas cell membranes are relatively impermeable to ionized, or polar molecules (eg, an unprotonated salicylate anion), nonionized, nonpolar forms (eg, the protonated, noncharged salicylic acid) may cross more easily. As xenobiotics pass through the kidney, they may be filtered, secreted, and reabsorbed. If the urinary pH is manipulated to favor the formation of the ionized form in the tubular lumen, the xenobiotic is trapped in the tubular fluid and is not passively reabsorbed into the bloodstream. This is referred to as ion trapping. Hence, the rate and extent of its elimination can be increased. To make manipulation of urinary pH worthwhile, the renal excretion of the xenobiotic must be a major route of elimination. The 2004 position paper of the AACT and EAPCCT emphasizes that, as discussed above for extracorporeal therapies, enhanced removal does not necessarily translate into a clinical benefit with improved outcomes.⁵⁶

Alkalinization of the urine to enhance elimination of weak acids has a limited role for xenobiotics such as salicylates,⁴⁸ phenobarbital, chlorpropamide, formate, diflunisal, fluoride, methotrexate, and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D). These weak acids are ionized at alkaline urine pH, and tubular reabsorption is thereby greatly reduced. Alkalinization is achieved by the intravenous administration of sodium bicarbonate (1 to 2 mEq/kg rapid initial infusion with additional dosing) to increase urinary pH to 7 to 8.

This degree of alkalinization may be difficult, if not impossible, if metabolic acidosis and acidemia are present, as often is the case in patients with salicylate poisoning. In this situation, bicarbonate (administered as the sodium salt, sodium bicarbonate) is consumed by titration of plasma protons before it can appear in the urine. On the other hand, salicylate poisoning often causes respiratory alkalosis as well. In that case, when the PCO₂ is low, raising serum bicarbonate, equivalent to the induction of metabolic alkalosis, may lead to profound, life-threatening alkalemia. Finally, the risk of ECF volume overload with sodium bicarbonate administration is the same as with the administration of 0.9% sodium chloride or LR. Hypernatremia may also occur after administration of hypertonic sodium bicarbonate. Bicarbonaturia is also associated with urinary potassium losses, so the patient’s serum potassium concentration should be monitored frequently and KCl given liberally as long as GFR is not impaired. A further complication of alkalemia is a decrease of ionized calcium, which becomes bound by albumin as protons are titrated off serum proteins; in this event, tetany may occur.²¹ If these complications can be identified and dealt with judiciously and safely, the renal clearance of salicylate may increase fourfold as urine pH increases from 6.5 to 7.5 with alkalinization. Increasing urine pH by decreasing proximal tubular bicarbonate reabsorption via administration of carbonic anhydrase inhibitors such as acetazolamide is not recommended. Although elimination of a xenobiotic may be increased, metabolic acidosis will ensue unless ample sodium bicarbonate is also administered. In the case of salicylates, metabolic acidosis with acidemia increase distribution into the central nervous system. As with sodium bicarbonate administration, bicarbonaturia is accompanied by urinary potassium losses; hypokalemia may be profound. The role of urinary alkalinization in the management of patients with salicylate poisoning is discussed further in Chap. 39.

Alkalinization is also used to increase the solubility of methotrexate and thereby prevents its precipitation in tubules when patients are given high-dose methotrexate.¹³ Precipitation of sulfonamide antibiotics with kidney stones or kidney failure may also be prevented by alkalinization. Administration of sodium bicarbonate, sodium chloride, or LR also protects the kidneys from the toxic effects of myoglobinuria in patients with extensive rhabdomyolysis. However, because patients with rhabdomyolysis may have acute kidney injury, sodium bicarbonate administration must be used before kidney injury occurs and may lead to ECF volume overload if its administration continues after kidney failure has developed.

Acidification of the urine by systemic administration of HCl or NH₄Cl to enhance elimination of weak bases, such as phencyclidine or the amphetamines, is not useful and is potentially dangerous. The technique has been abandoned because it does not significantly enhance removal of xenobiotics and is complicated by acidemia.

Peritoneal Dialysis

Theoretically, PD enhances the elimination of water-soluble, low-molecular-weight, non–protein-bound xenobiotics with a low Vd. Clearance of xenobiotics by PD is related to the number of exchanges, dwell time of dialysate, the surface area of the peritoneum, and the molecular weight of the compound. The highest clearances are achieved for xenobiotics with molecular weights below 500 Da. The efficacy of PD is markedly decreased when the patient is hypotensive.

Although PD is a relatively simple method to enhance xenobiotic elimination, it is too slow to be clinically useful. Consequently, PD is never the method of choice unless hemodialysis and hemoperfusion are unavailable and transfer to a center that can offer these techniques is not feasible. Besides exchange transfusion, it may be the only practical option in small children when experience with extracorporeal techniques in younger age groups is lacking or until a child can be transported to an appropriate center.

Hemodialysis

Hemodialysis has been employed for 100 years, ever since the discovery by John J. Abel and coworkers that dialysis could remove substantial amounts of salicylates from animals.¹ During conventional hemodialysis, blood and countercurrent dialysate are separated by a semipermeable membrane (dialyzer). Xenobiotics then diffuse across the membrane from blood into the dialysate down the concentration gradients (Fig. 10–2). Blood is pumped through one lumen of a temporary dialysis catheter, passed through the machine, and returned to the venous circulation through the second lumen.

The utility of hemodialysis for the treatment of patients with toxicity caused by lithium, toxic alcohols, salicylates, or theophylline is unquestionable and is not dealt with here; each of these xenobiotics is described in detail in separate chapters that also review their toxicity and indications for extracorporeal therapies. This section describes the hemodialysis procedure in general and its application to some newer situations.

Prompt consultation with a nephrologist is always indicated in the case of poisoning with a xenobiotic that might benefit from extracorporeal removal. Annual AAPCC data consistently suggest that some salicylate-related deaths, for example, could have been prevented if hemodialysis had been instituted earlier (Chap. 136).¹⁹ To perform hemodialysis, a nephrologist must be available along with a nurse or technician. The dialysis machine requires preparation, and a vascular access catheter must be inserted. A delay, ranging from one to several hours before hemodialysis can be instituted should be anticipated, particularly during hours when the hospital’s dialysis unit, if there is one, is closed. If indicated, modalities of treatment such as fomepizole or ethanol for poisoning with toxic alcohols should be administered and other modalities to enhance elimination, such as urinary alkalinization or oral MDAC, should be used when appropriate.

The technical details of performing hemodialysis for treatment of patients with poisonings do not differ markedly from those used in the treatment of patients with acute kidney failure. Several technologic advances have enabled improved xenobiotic clearance while limiting side effects and enabling better tolerance of dialysis.

Vascular access is obtained via a double-lumen catheter that is made of silicon, polyethylene, polyurethane, or Teflon. The catheter is usually inserted either via the femoral, the subclavian, or the jugular vein. The subclavian and internal jugular veins have the added risk of pneumothorax and necessitate radiographic confirmation of catheter position, whereas femoral catheters have increased blood recirculation (ie, lower efficacy of approximately 10% to 15%).⁴² Hemostasis after catheter removal is also more easily achieved at the femoral site. Bleeding or thrombosis at the site used for vascular access occur in approximately 5% of cases but can usually be addressed by adequate post-procedure tamponade of the catheter site. Ultrasonographic placement reduces the risk of complication for all sites and is now recommended in the United States. Larger catheters now permit blood flow rates that can be as high as 450 to 500 mL/min, although 350 mL/min may sometimes be the maximum rate achievable. Nosocomial bacteremia may occur when central catheters are indwelling but are exceedingly rare if they are in place for less than 3 days. This complication is therefore unusual in poisoning where extracorporeal treatment is necessary for a relatively short period of time. Femoral venous lines should always be removed in patients who are mobile and out of bed.

Because new dialysis membranes also have higher water permeability, computerized ultrafiltration control is necessary. These membranes, composed of polysulfone, polyamide, polyacrylonitrile, and other synthetic polymers, have better biocompatibility than older, cellulose-derived membranes. Biocompatibility is measured by the rate of activation or release, after exposure to the membrane, of inflammatory mediators that include white blood cells, platelets, complement, and cytokines. Better biocompatibility means less activation of these potentially damaging mechanisms compared with more bioincompatible membranes. The influence of biocompatibility on the outcomes of patients receiving long-term hemodialysis for chronic kidney failure is still being assessed. Patients with end-stage kidney failure are exposed to these membrane materials during at least three treatments a week for many years. It is unlikely that better biocompatibility will affect outcomes for dialysis of poisoned patients who require only one or two treatments.

Dialysis membrane composition has also continually evolved. The hollow-fiber dialyzer, almost universally used today, is composed of thousands of blood-filled capillary tubes held together in a bundle and bathed in the machine-generated dialysate. Older “conventional” dialyzers, much less frequently encountered today, are made of cellulose-derived polymers, most commonly called cuprophane. Hemodialysis efficacy for poisonings with low-molecular-weight xenobiotics should improve with the use of larger membranes with larger clearances. “High-flux,” synthetic dialyzers have larger pores and increased surface area that allow greater clearance of larger molecules. With the use of high-flux and high-efficiency dialyzers, small molecule clearance has more than doubled, whereas clearance of larger molecules, such as β₂-microglobulin (molecular weight, 11,800 Da) have increased by a factor of five. There are some instances in which high-flux dialyzers might be important in promoting clearance of larger molecules, such as vancomycin (molecular weight, 1449 Da), which are not readily removed by conventional low-flux membranes.²⁴ However, the indications for performing hemodialysis to enhance removal of vancomycin and other xenobiotics with higher molecular weights have not been delineated. Nonetheless, a sound pharmacologic basis for the efficacy of dialysis must still be present; no amount of increased clearance will eliminate a xenobiotic with a large Vd or significant protein or tissue binding. Although clearance rates reported in the literature of the 1970s and 1980s may significantly underestimate currently achievable clearance rates,⁵⁰ these data are still of interest given the relative paucity of more recent reports.

Patients undergoing hemodialysis experience much less hemodynamic instability than in the past; the blood lines and artificial kidney (the dialysis membrane) should be primed with an appropriate volume of fluid to reduce or avoid hypotension when the procedure is started. Furthermore, the source of base in dialysate is now routinely sodium bicarbonate rather than sodium acetate; the latter caused hypotension and decreased cardiac output. Computerized machines allow fine control of ultrafiltration rates to limit volume losses; in the past, imprecise calculations and manipulations led to frequent episodes of hypotension. As a result of such innovations, treatment can be delivered in more instances than was previously possible. Hypotension may still occur in critically ill patients but can often be corrected with 0.9% sodium chloride, colloid, vasopressors, or inotropes.

Full anticoagulation with heparin is usually required to avoid clotting of the circuit. A typical adult heparin dose is 1000 to 5000 units as a bolus followed by 500 to 1000 units hourly, but lower doses can also be given. Alternatively, in patients at high risk of bleeding, periodic 0.9% sodium chloride flushes of the dialysis membrane may suffice. Use of a dialysate that contains citrate may be adequate for anticoagulation and may allow heparin to be avoided. These choices are particularly important when dialyzing patients for toxic alcohol and salicylate exposures who are at risk for increased intracerebral bleeding.

In poisoned patients, hemodialysis is usually performed for 4 to 8 hours. Some precautions are needed for the prescription of the dialysis parameters because poisoned patients have different characteristics than chronic kidney disease patients. In particular, serum potassium, phosphorus, and pH can be markedly different. Assuming that the patient’s serum potassium concentration is normal, a standard bicarbonate-based dialysate with a potassium concentration of 3 or 4 mEq/L and a calcium concentration of 3 mEq/L, flowing at 600 to 800 mL/min, is sufficient. If dialysis is performed in a dialysis unit, the dialysate is a mix of a concentrate with sodium bicarbonate and highly purified water, usually derived by reverse osmosis or deionization. Phosphate can be added to the dialysate bath if needed. Ultrafiltration is rarely required unless oliguric acute kidney injury is present. Dialysis procedures done in ICUs should use portable reverse osmosis machines to generate the water for mixing. Although undesirable, tap water can be used if its chlorine content is less than 0.1 ppm.⁷

Complications specifically related to the dialysis procedure are rare. Furthermore, centers administering dialysis treatments are increasingly common today, and costs of the procedure are on the decline. For these reasons, if dialysis appears required for survival of a poisoned patient, it should usually be attempted.

Table 10–3 lists the characteristics of xenobiotics that make them amenable to hemodialysis. These requirements greatly reduce the number of xenobiotics that can be expected to be cleared by dialysis. During hemodialysis, clearance of a xenobiotic (Cl_x) can be calculated by:

where Q_p is the plasma flow rate and ER is the extraction ratio.

where Q_B is blood flow rate and Hct is hematocrit. The ER is a measure of the percentage of xenobiotic passing through the artificial kidney or charcoal hemoperfusion cartridge. This can be calculated as:

where C_in is the concentration of the xenobiotic in blood entering the system and C_out is the concentration in blood leaving the system.

Thus:

With the recent advances in dialysis therapy, there appears to be a role for high-flux hemodialysis in the clearance of certain xenobiotics that were previously thought to be effectively removed only by charcoal hemoperfusion.²³ The situations in which these modalities are appropriate for these xenobiotics remain exceedingly rare. Because valproic acid is increasingly prescribed for neurologic and psychiatric disorders, the incidence of both intentional and unintentional overdoses of this drug will also increase.⁶⁷ As discussed previously, valproic acid is largely protein bound at therapeutic serum concentrations. Toxic concentrations saturate protein-binding sites, leading to a higher proportion of unbound drug in the serum and a lower apparent Vd, thereby making the drug more dialyzable. Indeed, several case reports have demonstrated that the clearance of valproic acid with high-flux hemodialysis is at least equivalent to, if not greater than, that of charcoal hemoperfusion.^29,38 Although carbamazepine has a low molecular weight and a Vd that would allow for clearance by hemodialysis, its high protein binding and lack of water solubility are expected to impede the efficacy of hemodialysis. Nonetheless, carbamazepine is also effectively cleared by high-flux hemodialysis.⁶¹ In a patient who underwent high-flux hemodialysis followed by charcoal hemoperfusion for carbamazepine toxicity, removal rates were similar for the two modalities.⁶⁹ Unlike the case of valproic acid, the rationale for these anecdotal reports for enhanced carbamazepine elimination is unclear. Whether the pharmacokinetics of the drug are altered at toxic serum concentrations, as are those of valproic acid, is not known. Because of its potential efficacy and availability and its insignificant effect on cost and adverse events, high-flux hemodialysis should probably replace charcoal hemoperfusion as the treatment modality of choice when extracorporeal elimination is to be performed. As stated above, however, effective clearance is not necessarily a surrogate for improved outcomes.

In addition to removing xenobiotics, hemodialysis can correct acid–base and electrolyte abnormalities such as metabolic acidosis or alkalosis, hyperkalemia, and ECF volume overload. Consequently, hemodialysis is preferred, if not essential, for poisonings characterized by these disorders, especially when clearance rates resulting from hemoperfusion and hemodialysis are relatively similar. Examples include salicylate poisoning, which is often associated with metabolic acidosis,³² and propylene glycol toxicity, which is often associated with lactic acidosis, especially in the presence of renal or hepatic impairment.⁵² Valproic acid toxicity causes hyperammonemia, which is reduced by hemodialysis, possibly contributing to the benefit of the procedure.⁷⁰

A more controversial question is the role of dialysis in the treatment of metformin-associated lactic acidosis. Although debated, present evidence now suggests that metformin intoxication itself may induce metabolic acidosis with elevated lactate concentration without predisposing factors.^2,55 Here, the small molecular weight and negligible plasma protein binding of metformin may allow for adequate drug removal despite a relatively large Vd. Endogenous clearance of metformin is quite high and dialysis has therefore less appeal in patients with intact kidney function. In addition, hemodialysis would rapidly correct acidosis via administration of sodium bicarbonate without the complication of volume overload. Clinical improvement with hemodialysis may result as much or more from rapid correction of the acidosis as from removal of the drug.

In addition, hemodialysis increases the elimination of some drugs administered therapeutically, such as folic acid and other water-soluble vitamins and xenobiotics. Doses of these drugs should be increased during dialysis or administered immediately afterward. Similarly, the rate of ethanol infusions used in the treatment of patients with toxic alcohol ingestions must be increased. Fomepizole also has a low molecular weight (molecular weight, 82.1 Da) and a low Vd (0.6–1.0 L/kg), so it can also be dialyzed. If necessary, it should be redosed after dialysis (Antidotes in Depth: A30). When fomepizole is not available, ethanol removal can be limited in such cases by enriching the dialysate with ethanol to a concentration of 100 mg/dL.¹⁴ Similarly, hypophosphatemia after more prolonged high-flux hemodialysis can be avoided by adding sodium phosphate salts (eg, Fleet Phospho-Soda) to the dialysate or by administering phosphate intravenously.

Hemoperfusion

During hemoperfusion, blood is pumped via a catheter through a cartridge containing a very large surface area of sorbent, either activated charcoal or a resin, on to which the xenobiotic can be directly adsorbed (Fig. 10–2). The activated charcoal sorbent is coated with a very thin layer of polymer membrane such as cellulose acetate (Adsorba: Gambro, Lakewood, CO), heparin-hydrogel (Biocompatible Hemoperfusion Systems: Clark, New Orleans, LA) or polyHEMA (2-hydroxyl methacrylate or Hemosorba: Asahi, Tokyo, Japan). The membrane prevents direct contact between blood and sorbent, improves biocompatibility, and helps prevent activated charcoal embolization. There may be a further theoretical advantage to the heparin–hydrogel coating to diminish platelet aggregation.

Other adsorptive resins were used for hemoperfusion in the past, such as the synthetic Amberlite XAD-2 and XAD-4 and anion exchange resins such as Dow 1X-2. None of these columns is currently approved or available for use in the United States, but remain available in other countries. The literature regarding their efficacy is scant and relatively anecdotal. Although in vitro evidence suggests that these resins may have greater adsorptive capacities than activated charcoal, there are few, if any, meaningful comparisons in a clinical setting.

The adsorptive capacity of the cartridge is reduced with use because of deposition of cellular debris and blood proteins and saturation of active sites by the xenobiotic in question. Hemoperfusion is usually performed for 4 to 6 hours at flow rates that can usually not exceed 350 mL/min because of the risk of hemolysis.⁵⁸ The technique can be used in adults¹⁵ or children.⁵¹

The characteristics of xenobiotics that make them amenable to hemoperfusion (Table 10–3) differ slightly from those for hemodialysis in that hemoperfusion is not as limited by plasma protein binding, although this is less true with the advent of new dialysis membranes which now permit removal of significant amounts of highly bound xenobiotics.²⁷ Some xenobiotics are poorly adsorbed by activated charcoal, including the alcohols, lithium, and many metals (Antidotes in Depth: A1), making hemoperfusion inappropriate in their management. Hemodialysis and hemoperfusion have been performed in series for procainamide, thallium, theophylline, and carbamazepine overdoses, with greater apparent clinical efficacy than with either procedure alone.^10,17,37 In this technique, blood circulates first through the hemodialysis membrane and then through the activated charcoal cartridge. If blood traverses the dialysis membrane first, some of the xenobiotic is dialyzed, and the activated charcoal cartridge has less drug to adsorb.³¹

Added to the questionable benefits of hemoperfusion, multiple limitations make their use less attractive when compared to hemodialysis. A practical problem limiting the use of charcoal hemoperfusion is the availability of the cartridges. Many dialysis units do not currently stock these cartridges.⁶⁴ Activated charcoal cartridges were more available in hemodialysis units when they were more frequently needed for the treatment of patients with chronic aluminum toxicity. Saturation of the cartridge is apparent within 1 hour of use and markedly decreases absorptive capacity at 2 hours, whereas no such decreased performance is apparent with hemodialyzers. Compared with hemodialysis, patients must be anticoagulated with greater amounts of heparin. The cartridges are expensive ($350–$425 compared with the maximal cost for a high-flux dialysis membrane at about $40–$50), especially considering cartridges usually need to be replaced every 2 hours. Some have expiration dates, limiting their shelf life. Others, such as the Clark system, have an indefinite shelf life but must be autoclaved before use, which may affect their availability in an emergency. Complications of hemoperfusion are more common than with hemodialysis: thrombocytopenia, leukopenia, and hypocalcemia.⁶⁵ Finally, hemoperfusion cannot correct acid–base or electrolyte abnormalities and cannot provide ultrafiltration if volume overload occurs.

Although hemoperfusion has historically been considered the preferred method to enhance the elimination of carbamazepine, phenobarbital, phenytoin, and theophylline (Table 10–4), recent improvements in hemodialysis technology may make older comparisons of hemodialysis and hemoperfusion clearance rates obsolete. The most frequent indication for charcoal hemoperfusion in the past was theophylline toxicity, and theophylline is rarely used today in the treatment of obstructive lung disease and asthma and is consequently less often implicated in acute and chronic poisoning. Along with the rarity of aluminum toxicity in chronic hemodialysis patients, the infrequency of theophylline prescription accounts for the diminished availability of activated charcoal cartridges and the relative infrequency with which the procedure is performed. In a survey of New York City ­hospital dialysis units taking 911 calls, only 10 of 34 units had cartridges.⁶⁴ In a review of the AAPCC annual data, theophylline was the most common xenobiotic removed by hemoperfusion from 1985 to 2000, but carbamazepine became the most frequent xenobiotic removed by hemoperfusion during 2001 to 2005.³⁰ As in the case of hemodialysis, doses of drugs used therapeutically may need to be increased if they are removed by hemoperfusion.

Liver Dialysis

A newer concept for poisonings is that of liver dialysis. These procedures are currently available in the United States, mostly for the treatment of liver failure. Several techniques have been developed; the following three are the most common. (1) Single pass albumin dialysis (SPAD) is similar to hemodialysis but has albumin added to the dialysate in counter-directional flow and then discarded after passing through a filter. (2) The Molecular Adsorbents Recirculation System (MARS) is identical to SPAD, but the albumin-enhanced dialysate (with the adsorbed xenobiotics) is itself recycled after going through another dialysis circuit and through both resin and activated charcoal cartridges. (3) The Prometheus system is a device that combines albumin adsorption with high-flux hemodialysis after selective filtration of the albumin fraction through a polysulfone filter. In all of these techniques, the dialysate bathing the fibers contains human serum albumin that serves as a sorbent to bind the xenobiotic of interest and maintain concentrations of the free xenobiotic at zero. A steep concentration gradient from blood to dialysate is established so that even highly protein-bound xenobiotics can be removed from the plasma. The membrane is impermeable to albumin, which remains in the dialysate.

These extracorporeal devices are all able to remove protein-bound xenobiotics, but their use in poisoning remains limited to rare case reports. These devices are mostly used in patients with hepatic encephalopathy and liver failure and as a bridge to hepatic transplantation.^47,66 In toxicology, their use is mostly described in hepatic failure following toxicity of APAP or Amanita spp. exposure (Chap. 120). The reasons why the procedure might offer significant benefits are not completely understood, although it does remove both water-soluble and protein-bound compounds. It is not known which protein-bound molecules are removed from the blood, such as bile salts and aromatic amino acids, to account for the therapeutic advantage in hepatic failure. Specific uses for liver dialysis for elimination enhancement has been described in poisoning with phenytoin,⁶³ theophylline,⁴⁰ valproic acid,¹⁸ and Amanita mushrooms.³⁶ Whether this relatively expensive (in excess of $4000 per treatment), complicated, and nonspecific procedure would offer benefit in a handful of instances in which protein-binding limits removal of xenobiotics is not known. The report of the initial use of a sorbent-based hemodiabsorption device in 10 cases of CA overdoses claimed benefit, although the limitations of extracorporeal therapy for this particular class of drugs are discussed above.⁶ Devices using powdered sorbent are not currently available in the United States.

Continuous Hemofiltration and Hemodiafiltration

Hemofiltration is the movement of plasma across a semipermeable membrane in response to hydrostatic pressure (convection). The addition of dialysate on the other side of the membrane further enhances elimination of xenobiotics or endogenous uremic toxins. A review of continuous modalities of dialytic therapy concluded that they are still relatively unproven for the treatment of poisoning.²⁸ These techniques find relatively widespread usage for the treatment of patients with acute kidney injury in the ICU, and in this context, they are referred to collectively as modalities of continuous renal replacement therapy (CRRT). The clearances of either urea or xenobiotics that are achieved with these techniques are significantly lower than those achieved with hemodialysis. But as continuous modalities, what they lack in clearance they can make up for in time.

There are several possible advantages of continuous modalities. One is the capability to continue therapy for 24 hours every day, permitting hemofiltration to be instituted after hemodialysis or hemoperfusion to further remove a xenobiotic after it redistributes from tissue to blood.⁴⁶ This is an attractive modality for slow, continuous removal of drugs, such as lithium, which distribute slowly from tissue-binding sites or from the intracellular compartment (Fig. 10–1). Other xenobiotics with volumes of distribution that are large enough to preclude use of dialysis or hemoperfusion might also be eliminated with longer courses.

Table 10–3 summarizes the properties of xenobiotics that make them amenable to hemofiltration. However, the rate of removal with this form of therapy may be insufficient to benefit critically ill patients. Patients who can tolerate slower clearance rates may not require this enhanced elimination therapy at all. Whether slow treatment to avoid redistribution of intracellularly distributed lithium, for example, is preferable to repeating conventional hemodialysis is unclear.^45,71 Rebound of serum concentrations indicates that the drug is moving out of the intracellular compartment, where it is toxic, into the extracellular compartment, where it is susceptible to removal by hemodialysis. Therefore, although the continuous modalities are suggested to prevent rebound, the importance of this property, during which poison in the toxic compartment is actually decreasing while plasma concentrations actually increase, remains unproven and may in fact be desirable. Despite many case reports demonstrating significant xenobiotic clearance, no data have demonstrated that these continuous techniques affect prognosis or mortality in treatment of patients with xenobiotic toxicity. In most cases, hemodialysis should be considered the preferred initial mode of therapy.

The continuous modalities may be best suited for patients with hypotension who cannot tolerate conventional hemodialysis, although this situation is infrequent unless oliguric acute kidney injury is present.¹¹ These modalities may also have a role after hemodialysis to avoid the need for a repeat session after equilibration of the xenobiotics into plasma. The continuous modalities may also have a slight advantage over the use of high-flux hemodialysis membranes, in being able to clear larger molecules such as myoglobin (molecular weight, 17,000 Da). Growing evidence suggests that a significant part of the clearance of many molecules occurs because of adsorption of the molecule to synthetic membranes.¹⁶ However, quantification of the contribution of adsorption to total clearance is difficult and is not usually accomplished in published studies. Adsorption of molecules responsible for adverse effects is being studied as a potential benefit in the management of cytokines and liver failure and could be a property useful in the management of poisoning.⁵⁷ At present, this property is difficult to quantify and of uncertain benefit. Another practical advantage of CRRT is that the procedure is usually done now in ICUs by ICU nurses, and when available in such units, might not require dialysis personnel. Familiarity of ICU staff with the procedure is critical to its availability when needed; it is most likely to be used effectively in hospitals with higher incidence rates of acute kidney injury.

In pure hemofiltration, sometimes called slow continuous ultrafiltration (SCUF), there is no dialysate solution on the other side of the dialysis membrane (Fig. 10–2). Molecules are transported across the membrane with plasma water, a mechanism known as convective transport or bulk flow. The ECF volume status of the patient determines whether replacement of all or some of the filtered plasma with physiologic electrolyte solution (LR or other commercially available preparations) is indicated. Although this technique can be done intermittently using a hemodialysis machine, it has been adapted for use in ICUs as a continuous form of treatment, particularly when removal of ECF is indicated. The clearance of low-molecular-weight solutes such as urea is relatively limited.

Solute clearance may be significantly enhanced by having a dialysate solution bathe the blood-filled capillaries running countercurrent to the blood flow, in order to add diffusion to convection. The combination of hemofiltration with dialysis is known as hemodiafiltration. Addition of dialysis also usually suffices to treat supervening acute kidney failure or preexisting chronic kidney failure. Hemodiafiltration, similar to hemodialysis, requires that blood perfuse hollow-fiber dialysis membranes made of synthetic plastics such as polysulfone or polyamide.²⁰ For all of these procedures, the patient usually must be fully anticoagulated, but some hemofilters are available that may not require anticoagulation. Anticoagulation may be achieved either with heparin or with citrate. The hydrostatic pressure required for hemofiltration is derived from a blood pump. Continuous venovenous hemofiltration (CVVH) uses a blood pump to maintain adequate flow rates and has replaced continuous arteriovenous hemofiltration (CAVH), which required arterial puncture with a large-bore catheter. However, the need for a blood pump also necessitates an experienced ICU team to be continuously present for more than the 4 to 6 hours needed for acute hemodialysis or hemoperfusion. Both the expense and the complexity of the xenobiotic-removal procedure are thereby increased. The addition of a dialysate bathing solution to the hemofiltration apparatus changes CVVH, or hemofiltration, to the augmented continuous venovenous hemodialysis (CVVHD), or, with greater convective volumes, hemodiafiltration (CVVHDF) (Fig. 10–2).

Ultrafiltrate flows of 100 to 6000 mL/h across the membrane can be achieved. Fluid and electrolyte losses must be replaced carefully. Depending on the filter, xenobiotics with a molecular weight of up to 40,000 Da, as well as water, urea, creatinine, and sodium, pass into the ultrafiltrate. Heparin, myoglobin, insulin, and vancomycin are examples of larger molecules cleared with relative efficiency.

Attention must be paid to the undesirable removal of therapeutic drugs such as antimicrobials via these continuous modalities. Drug clearances with different synthetic membranes are available in the literature; the doses necessary to maintain therapeutic drug concentrations can also be determined.⁵

Plasmapheresis and Exchange Transfusion

Plasmapheresis and exchange transfusion are intended to eliminate xenobiotics with large molecular weights that are not dialyzable. This includes xenobiotics and endogenous molecules with molecular weights greater than 150,000 Da typified by immunoglobulins. The xenobiotic to be eliminated should also have limited endogenous metabolism to make pheresis or exchange worthwhile.³⁵ By removing plasma proteins, both techniques offer the consequent potential benefit of removal of protein-bound molecules such as Amanita toxins,³⁰ thyroxine, vincristine, monoclonal antibodies, or complexes of digoxin and antidigoxin antibodies. However, there is little evidence that either technique affects the clinical course and prognosis of a patient poisoned by any of these or other xenobiotics.

Pheresis is particularly expensive, and both pheresis and exchange transfusion expose the patient to the risks of infection with plasma- or blood-borne diseases. Replacement of the removed plasma during plasmapheresis can be accomplished with fresh-frozen plasma, albumin, or combinations of both. The former is associated with hypersensitivity reactions, such as fever, urticaria, wheezing, and hypotension, in as many as 21% of cases.³⁵

A different setting in which exchange transfusion may be an appropriate technique is in the management of small infants or neonates in whom dialysis or hemoperfusion may be technically difficult or impossible. Anticoagulation and MDAC may be hazardous and therefore contraindicated in patients in the neonatal nursery, where the risk of intracerebral bleeding and necrotizing enterocolitis is high. In premature neonates, a single volume exchange appeared to alleviate manifestations of theophylline toxicity.⁴⁹ The therapy has been successfully used to treat other pediatric patients with poisonings, including severe salicylism.⁴⁴

Further discussions of some of the techniques to enhance elimination that are not discussed here may be found in Chap. 31 (exchange transfusion), Special Considerations: SC3 (cerebrospinal fluid drainage and replacement), Chap. 65 (toxin-specific antibodies), and Chap. 96 (chelation). All of these techniques have limited, very specific indications, and the effect of these interventions on the overall body burden of the xenobiotic is usually small.

Unlike patients who receive acute hemodialysis once or twice in the management of poisoning, patients with chronic kidney failure are repeatedly exposed to large volumes of water derived from municipal reservoirs during the course of their hemodialysis treatments. If an “average” regimen consists of three treatments of 4 hours each week, with dialysate flows of 800 mL/min, patients will be exposed to nearly 600 L of water separated from them only by a semipermeable membrane designed to allow solute passage in either direction. Problems with dialysate generation therefore have the potential to be lethal to this population by exposing them to significant quantities of toxins. There are two potential sources of dialysate contamination: the municipal reservoirs and water treatment plants and the dialysis unit.⁵⁴ The quality of water used for dialysate generation is regulated in the United States by the Association for the Advancement of Medical Instrumentation (AAMI).⁷ This organization regularly revises its guidelines to incorporate new technology and data.

Contamination of dialysate from the municipal water supply may occur as a result of xenobiotic runoff into reservoirs or as a result of inadvertent or intentional addition of a chemical by the municipality. Chlorine and chloramine are frequently added to municipal water supplies to control bacterial populations. However, chlorine may combine with nitrogenous compounds and form chloramine, which may cause nausea, vomiting, methemoglobinemia, and hemolytic anemia.⁷² Recently, chloramine was blamed for decreased bone marrow sensitivity to erythropoietin. Aluminum is present in some municipal water supplies, and before it was recognized as a problem, aluminum led to encephalopathy characterized by seizures, myoclonus, and dementia as well as osteomalacia and microcytic anemia.

Water from the municipal supply entering the dialysis unit is first treated with a water softener to remove calcium and magnesium. It is then run through an activated charcoal bed to adsorb chloramine. The potential toxicity (hemolysis and death) from this compound has caused AAMI to mandate a redundancy in the carbon beds; when the active sites in one carbon bed are exhausted, a second will ensure that no toxicity will occur. Most commonly, water for dialysate is then generated by reverse osmosis, a process that requires that water, in response to applied hydrostatic pressure (and against the osmotic gradient), cross a membrane that is relatively impermeable to solutes, leaving them behind. Alternatively, but less commonly, water can also be purified using deionization, a technique that runs water over an exchange resin, releasing hydroxyl ions in exchange for charged species in the water. Deionization is inferior to reverse osmosis for removal of aluminum and may be associated with release of lethal concentrations of fluoride when exchange sites are exhausted.⁴ General water chemistry testing is mandated annually; testing for chlorine and chloramine must be done before each dialysis shift.

Current requirements are that water be highly purified, but not sterile, because bacteria cannot cross from the dialysate into the blood. However, small quantities of endotoxin (molecular weight, 5–15,000 Da) can cross, particularly in situations that include the use of high-flux membranes. Endotoxin is suspected of contributing to activation of circulating cytokines, malnutrition, fever, and other syndromes such as carpal tunnel syndrome, which are associated with chronic inflammation. Recommendations for the frequency of testing for endotoxin and the maximum amounts of endotoxin tolerated are continually debated and continue to become more stringent. Water distribution systems are cleaned at least monthly with bleach, peracetic acid, or other sterilants. Care must be taken that all of these potential xenobiotics are thoroughly flushed from the system before dialysis is restarted. Other products of bacterial metabolism, such as volatile sulfur-containing compounds, generated on the dialysate side of the membrane may be responsible for symptoms in hemodialysis patients.⁶²

Unusual microbes have also been associated with serious toxicity. Untreated water at one center in Brazil demonstrated growth of Cyanobacteria (blue-green algae) and production of microcystins, cyclic peptides that cause serious hepatic toxicity; patients dialyzed with the contaminated water had a dramatic rate of death from liver failure.³³ Water contamination should especially be suspected when multiple dialysis patients experience similar symptoms nearly simultaneously. Dialysate distribution systems are always made of polyvinyl chloride (PVC) or other inert plastics, rather than copper, which may also leach into the water and cause hemolytic anemia.

Besides water for dialysate, another potential source of poisoning in hemodialysis units is the process of reusing dialysis membranes. Until recently, up to 70% of dialysis units in the United States reused membranes because of cost considerations. Reuse is becoming less common today, in part because the cost of dialyzers has continued to fall. Each dialysis membrane is sterilized with peracetic acid, formaldehyde, or glutaraldehyde. Careful quality assurance programs ensure that there is no significant exposure of the patients to these molecules during the dialysis procedure. Nonetheless, reuse programs are associated with a variety of syndromes such as pyrogenic reactions attributed to patient exposure to germicides or endotoxin caused by inadequate sterilization procedures.⁵⁹ Controversial data over the years have suggested, but not proven, higher mortality rates with reuse. Occupational exposure of dialysis personnel to the relevant sterilants has also been monitored closely.

• Urinary alkalinization and many of the other techniques listed in Table 10–1 can be instituted quickly in the emergency department.

• The extracorporeal methods of xenobiotic removal, including hemodialysis, sorbent hemoperfusion, and continuous hemofiltration, all require consultation with a nephrologist or intensivist.

• Timely use of these techniques requires mobilization of a competent team and preparation of the requisite equipment.

• Rapid identification of a toxic exposure for which these techniques are appropriate and the presence of more ominous prognostic features should lead to prompt notification of the appropriate consult services so that application of these techniques can proceed in an expeditious manner.

• The applicability of these techniques to new xenobiotics should be considered based on the principles discussed, so that these and newer treatment modalities are not used indiscriminately.

• The literature regarding these techniques in general, and for the treatment of specific xenobiotic exposures in particular, should be read critically and with appropriate skepticism.

Acknowledgment

Daniel Matalon, MD, contributed to this chapter in previous editions.

References