Hemodialysis International 2010; 14:158–167
Review Article
Dialysis in the poisoned patient
George BAYLISS
Division of Kidney Disease and Hypertension, Alpert School of Medicine, Brown University, Providence,
Rhode Island, USA
Abstract
Patients who ingest toxic substances may require extracorporeal removal of the poisons or their
toxic metabolites if native renal clearance is not sufficient because of acute kidney injury, acuity of
symptoms, or burden of toxin. Here, a case is presented, and the literature on renal replacement
therapy in the event of acute intoxication is reviewed. Extracorporeal therapy efficacy is examined in
terms of the characteristics of the toxin (molecular size, charge, protein, or lipid binding); the patient
(body habitus and volume of distribution); and the process (membrane effects on extraction ratios
and sieving, role of blood, and dialysate flow rates). The choice of extracorporeal therapy and
hemodialysis prescriptions for specific poisonings are discussed.
Key words: Dialysis, poisoning, hemoperfusion
A CASE
A 51-year-old woman with a history of depression and
chronic pancreatitis from Valproic acid (Depakote) exposure was brought to the emergency department by EMS
after her 11-year-old daughter found her unresponsive on
the bathroom floor. In the emergency department, the
patient was tachycardic, with an initial heart rate of
129 beats/min, but normotensive, with blood pressure
ranging from 139/65 to 135/65 mmHg. She was tachypneic, with a respiratory rate of 44 breaths/min. Oxygen
saturation by pulse oximetry was 97% on 3 L of oxygen
by nasal cannula. She was minimally responsive with dilated, minimally responsive pupils. Cardiac exam revealed normal heart sounds and no murmurs, rubs, or
gallops. Her clear lungs were clear to auscultation. Her
abdomen was soft, her mucus membranes were dry, and
her skin was damp. She had no lower extremity edema. A
chest X-ray showed a normal-sized heart and no infiltrates or pleural effusions. Nurses placed a Foley catheter
in the emergency department, and the patient voided
Correspondence to: G. Bayliss, MD, Division of Kidney
Disease and Hypertension, Alpert School of Medicine, Brown
University, Rhode Island Hospital, Providence, RI 02903,
USA.
E-mail: gbayliss@lifespan.org, geobayliss@gmail.com
158
1.4 L of urine over 4 hours as she was volume resuscitated
with isotonic fluid.
Serum chemistries showed elevated serum creatinine
(1.7 mg/dL, up from her baseline of 0.7 mg/dL), mild
hyperkalemia (potassium = 5.3 mEq/L), serum bicarbonate of 14 mEq/L, and an anion gap of 24. An arterial blood
gas showed pH = 7.38 and pCO2 = 18 mmHg, indicating
both a metabolic acidosis and respiratory alkalosis.
Urine toxicology screen was positive for salicylates, and
the serum salicylate level was 134 mg/dL (therapeutic
o30 mg/dL).
The patient’s initial electrocardiogram revealed sinus
tachycardia, left axis deviation, a narrow QRS complex,
and no prolongation of PR or QT intervals. She subsequently developed supraventricular tachycardia at a heart
rate of 145–160 beats/min that did not respond to
adenosine.
A toxicology consultation was obtained, and the patient was started on a sodium bicarbonate infusion. A renal consult was requested.
DISCUSSION
This patient appeared critically ill and at risk of needing
intubation for airway protection. Her case illustrates the
key circumstances under which extracorporeal removal of
toxins is required to prevent extreme morbidity and
r 2010 The Authors. Journal compilation r 2010 International Society for Hemodialysis
DOI:10.1111/j.1542-4758.2009.00427.x
Dialysis in the poisoned patient
death. In such cases, experts recommend immediate consultation with a medical toxicologist or poison control
center to optimize medical management of patients while
the need for and feasibility of extracorporeal therapy is
evaluated, based on the molecular size of the toxin, its
volume of distribution, water solubility, and proteinbinding characteristics.1
The use of hemodialysis, continuous renal replacement
therapy (CRRT), or hemoperfusion is important in the
management of certain drug overdoses and toxic ingestions when other measures such as activated charcoal,
gastric lavage, specific antidotes, and respiratory support
have failed or are not feasible because of the patient’s
condition.2 Such extracorporeal procedures should be
considered adjunctive, and not substitutes for other measures and supportive care. Furthermore, certain measures
such as gastric lavage, forced alkaline diuresis, or administration of chemical inhibitors may be critical in the early
phases of drug ingestion, obviating the need for extracorporeal therapy or supplementing it.
The circumstances that in this case point to the need
for extracorporeal drug removal include progressive deterioration of the patient’s status despite intensive supportive therapy; decreased level of consciousness with
suppression of midbrain functions; the risk of complications of coma like aspiration pneumonia; impaired native
clearance of drugs and toxins in the setting of reduced
glomerular filtration as a result of acute kidney injury; the
amount of toxin ingested; and faster clearance of the toxin
by extracorporeal means than normal native clearance
would provide.2 In the case of salicylate intoxication,
aggressive volume expansion and forced diuresis could
prove harmful with complications that include pulmonary3
or cerebral edema.4
Extracorporeal techniques include dialysis and hemoperfusion; dialysis therapies include intermittent hemodialysis, CRRT, and peritoneal dialysis. Each has
advantages, depending on the properties of the toxin being removed, the total burden of toxin in the patient, and
the rate at which the extracorporeal method removes toxins from the blood. Therapeutic plasma exchange has also
been used to remove protein-bound drugs as well as
endotoxin.5
Key principles
Molecular size, charge, and binding characteristics help
determine the ease with which toxic substances and
drugs in overdose can be removed from the patient.
Low molecular weight substances (like urea, MW 60 Da)
that are water-soluble pass easily across a dialysis mem-
Hemodialysis International 2010; 14:158–167
brane, while large, protein-bound or lipid-bound molecules are more difficult to remove because of their size
and difficulty in breaking those bonds, thus requiring
dialysis membranes with larger pores (like b2 microglobulin, MW 11,800 Da).
The pharmacokinetics of a drug or a toxin depends on
several factors.6 The first is absorption, with the bioavailability of the drug defined as the percentage of the administered drug that reaches systemic circulation.
Next, the volume of distribution, Vd, of a toxin represents the theoretical dispersion of the substance in the
patient’s body and is defined as the amount of drug in the
body divided by the concentration of the drug in plasma.
This depends on both the characteristics of the patient
and the toxin. Water-soluble molecules will remain more
accessible and toxic levels can be more easily reduced
when the toxin has a small Vd in the patient. Lipid-soluble molecules have a larger Vd, and toxic levels are more
difficult to reduce because the intravascular levels represent only a small fraction of the total body burden of the
toxin.
A related concept is rebound. Intravascular levels of
lipid-soluble molecules may decline quickly after an initial session of dialysis, but increase again as serum levels
re-equilibrate from the extravascular space. Patient characteristics that affect a drug’s Vd include obesity, extracellular fluid volume status, age, gender, thyroid function,
renal function, and cardiac output. In general, a drug’s
plasma concentration correlates inversely with its Vd.6
Clearance of a substance is the theoretical volume of
blood from which the substance is removed per unit of
time.7 Native clearance depends on the ability of a molecule to pass across the glomerular basement membrane
into the urinary space, a function of molecular size and
charge, and the urine flow rate [U]V/[P], expressed in
milliliters per minute. Solute removal occurs first by filtration (convection), but is also influenced by modification in the tubules, the continuous nature of native renal
clearance, and stable serum concentrations in the steady
state.8
In dialysis, clearance is achieved through diffusion and
convection via ultrafiltration and is defined as the product
of the extraction ratio—the percentage of the substance
removed from the blood as it passes over the membrane—
and the blood flow rate [(Cinitial Cfinal)/Cinitial][QB].9 It is
also expressed in terms of milliliters per minute. Intrinsic
characteristics of the dialysis membrane that affect clearance across the membrane include the size and number of
dialysis membrane pores (flux), membrane composition
and thickness, and the total surface area of the dialysis
membrane (efficiency). The intermittent nature of the ther-
159
Bayliss
apy means that solute clearance, mass removal rate, and
serum concentration are not constant.8
Protein binding considerably affects a drug’s pharmacokinetics, with only unbound drugs or active drug metabolites able to exert a pharmacologic effect. Drugs that
are highly protein bound are not removed effectively during dialysis. In general, the volume of distribution of a
drug increases as its protein binding decreases.6
Knowledge of whether renal clearance can be made to
become a major component of total clearance may help
in the management of drug overdoses.10 The native clearance of poisons may be affected by reductions in glomerular filtration rate (GFR) as a result of hypotension or
volume depletion, the large volume of distribution of the
drug or toxin, its charge, size, and binding characteristics.
Renal drug clearance is also influenced by tubular secretion and re-absorption as well as renal epithelial cell
metabolism. Highly protein-bound drugs have either reduced clearance or are secreted into the tubules by active
transport, which is dependent on renal blood flow but
not on GFR. Most drugs and their metabolites are eliminated by first-order kinetics, and the amount of drug
eliminated over time is a fixed proportion of the total
body store. The half-life of a drug thus increases as the
renal function decreases.6
Many toxins and drugs that can build up to toxic levels
are weak acids, like salicylates, barbiturates, and chlorinated acid herbicides. In theory, maintaining them in the
ionized state decreases their re-absorption across
phospholipid cell membranes and potentially increases
their rate of excretion in the urine if the substance is
cleared by the kidneys. Increasing the urinary pH should
in theory increase the ionized fraction of the substance,
which is expressed in terms of the pKa or dissociation
constant. Weak acids have a high pKa. Ionic and nonionic
forms are in equilibrium when pH = pKa. Most drugs are
in the nonionized phase at the physiologic pH = 7.4.
Thus, in those cases where alkalinization is indicated,
experts recommend that the goal should be a urine
pH 7.5 and be maintained at or above that level to
maximize elimination. The same experts distinguish urinary alkalization from forced diuresis, whose goal is
merely to increase urine output.11
Mechanical aspects of dialysis also have an important
effect on the clearance of toxic substances and drugs. In
general, dialytic clearance of a substance is determined by
the intrinsic clearance of the dialyzer membrane, the
length of time of treatment, blood, and dialysis flow rates.
Thus, a high-efficiency dialyzer will provide more surface-area contact for blood and dialysate, allowing greater
diffusion of substances down the concentration gradient
160
from blood to effluent. Larger pores will allow greater
convective solute clearance, or sieving, during ultrafiltration via solvent drag.12
The combination of diffusive and convective clearance
provides more total clearance per unit of membrane surface area than either alone, although the addition of convection, which is more important for larger molecules,
reduces diffusive clearance, which is more significant for
smaller molecules, when replacement fluid is added to
blood prefilter to compensate for increased water permeability of the membrane.13 Older research has shown that
blood contact with high-flux membranes can decrease the
sieving coefficient in symmetric membranes, like the
AN69, through protein adsorption to the membrane surface.14 Increased use is being made of convection to remove middle and high molecular weight solutes. But
while high-flux polysulfone membranes have high middle molecular weight solute clearance, they also contribute to significant loss of albumin.15 Work is under way
to develop a high-flux membrane that does not lead to
albumin loss.16
Modalities of extracorporeal toxin
removal
Hemodialysis is the modality of choice for low molecular
weight, water-soluble molecules, especially those that
have small volumes of distribution and are not protein
bound or lipid bound. It is ideal for molecules that diffuse
easily across dialysis membranes. Its use is limited in patients who are hypotensive because dialysis, even without
volume removal, may reduce the blood pressure further.
The use of dialysis is further limited to drugs and toxins
that reach plasma concentrations that are high enough to
establish a sufficient concentration gradient like alcohols,
lithium, and salicylates.17
Continuous renal replacement therapies have a theoretical advantage in patients who have ingested substances that are highly lipid bound and have large
volumes of distribution with slow intercompartmental
transit times from the extravascular to the intravascular
space. In theory, CRRT is ideal in cases in which serum
levels of toxins rebound after rapid removal as toxins reequilibrate in the vascular space over time from body
stores such as lipid-bound molecules in fat stores. CRRT
is also useful in hemodynamically unstable patients who
are unable to tolerate high blood flow rates. Clearance is
achieved through the longer total dialysis time, rather
than per unit time, given the slower blood flow rates.7
Some researchers have suggested adding albumin to the
Hemodialysis International 2010; 14:158–167
Dialysis in the poisoned patient
dialysate to remove highly protein-bound drugs like valproic acid and carbamezapine during CRRT.18
Hemoperfusion uses a cartridge similar to a dialysis
membrane in which toxins are removed from the blood
by binding to activated charcoal or resin rather than
diffusing out of the blood down a concentration gradient. The procedure is run on a dialysis machine using
regular dialysis pumps but without a dialysate. It is useful
with highly protein-bound and lipid-bound molecules.
Disadvantages include saturation of the cartridge and the
need to change it every 2–3 hours. Hemoperfusion cartridges can be run in series with hemodialysis membranes
in patients who also need solute and fluid removal for
kidney failure.2
Peritoneal dialysis has fallen out of favor as a modality
in the poisoned patient because of its low effectiveness in
removing toxins. It is useful in small children in whom a
peritoneal dialysis catheter can be inserted relatively
quickly and in cases of overdoses further complicated
by hypothermia, in which core rewarming is required.2
Therapeutic apheresis has a theoretical advantage over
dialysis or hemoperfusion in removing substances that
are protein-bound or lipid-bound and in the intravascular
space (4300 Da).19 But there have been no controlled
trials, and recommendations for its use are based on case
reports and series. The American Society for Apheresis
rates toxin removal as a class III indication (suggestion of
benefit but insufficient evidence to establish efficacy or
risk vs. benefit). An exception is plasmapheresis for
phalloid intoxication after ingestion of poisonous mushrooms, particularly Amanita phalloides.20 The ASFA gives
this a class II indication (generally accepted but considered adjunct therapy). Some consider plasmapheresis at
least as effective as hemoperfusion in reducing mortality
following phalloid mushroom ingestion, but there is not
enough evidence to say that it is superior21 (Table 1).
Trends in extracorporeal toxin removal
Hemodialysis has increased sharply as the main modality
for extracorporeal removal of toxins. Cases of poisoning
requiring hemodialysis increased to 707 per million calls
to poison control centers from 231 per million over the
time period 1985–2005.22 In the same period, the number
of cases of poisoning requiring hemoperfusion declined to
12 from 53 per million calls, while cases of poisoning requiring peritoneal dialysis declined to 1.6 from 2.2 per
million calls. Lithium and ethylene glycol remained the
most common toxins removed by hemodialysis over the
20 years, followed by salicylates and valproic acid. Carbamazepine has replaced theophylline as the toxin removed
most frequently by hemoperfusion (Table 2).
Technical aspects and complications
The technical aspects of toxin removal by extracorporeal
means are largely related to access and equipment. In
hemodialysis, CRRT, and hemoperfusion, access is through
a 10–11.5 French dual-lumen dialysis catheter in a central
vein because smaller catheters are unable to support the
necessary blood flows of 200–450 mL/min. Except where
contra-indicated, heparin should be used to anticoagulate
the system to improve dialyzer clearance and prevent
blood loss and the loss of time in replacing the dialysis
circuit if it clots. Heparin should not be used, for example,
in dialyzing patients with methanol toxicity. High-flux,
high-efficiency dialysis membranes should be used to
maximize pore size and surface area for toxin removal by
both diffusion and convection, as discussed above.2
Complications from toxin removal by hemodialysis include hypokalemia and alkalosis due to diffusion of potassium into the dialysate and diffusion of bicarbonate
from the dialysate, particularly given the need for high-
Table 1 Drugs, toxins amenable to extracorporeal removal by extracorporeal modality
Hemodialysis
Lithium
Ethylene glycol
Methanol
Salicylates
Valproic acid
Metformin
Theophylline
CRRT
a
Lithium
Theophyllinea
Valproic acida
Metformina
Pertitoneal
Hemoperfusion
Plasmapheresis
b
Carbamezapine
Theophylline
Paraquat
Phalloids
a
Hemodialysis preferred method when blood pressure permits.
No demonstrated advantage for removal of a particular toxin.
CRRT =continuous renal replacement therapy.
b
Hemodialysis International 2010; 14:158–167
161
Bayliss
Table 2 Hemodialysis vs. hemoperfusion and CRRT in acute intoxication. (Adopted with permission from aBorkan SC. Extracorporeal therapies for acute intoxication. Crit Care Clin 2002; 18: 393–420; Goodman JW and Goldfarb DS. The role of
continuous renal replacement therapy in the treatment of poisoning. bSemin Dial 2006; 19:402–407; Mulder J et al. Platelet
loss across the hemofilter during continuous hemofiltration. bInt J Artif Organs 2003; 96:206–212.)
Parameter
Renal replacement therapy
Restores electrolyte balance
Corrects volume status
Risk of thrombocytopenia
Removes drugs 4300 Da
Removes protein-bound drugs
Removes water-soluble drugs
Removes lipid-bound drugs
Rapid QB improves drug removal
Cartridge saturation
Can be used in hypotensive patient
a
Hemodialysis
a
Hemoperfusion
Yes
Yes
Yes
Unusual
No
No
Yes
No
Yes
No
No
No
No
No
Yes
Yes
Yes
No
Yes
Yes
Yes
No
b
CRRT
Yes
Yes
Yes
Yes
No
Limited
Yes
Theoretical
Yes
No
Yes
CRRT =continuous renal replacement therapy; QB =blood flow; SBP =systolic blood pressure.
efficiency dialysis membranes. The complications of
CRRT include hypocalcemia, particularly when using citrate anticoagulation. Hemoperfusion has been associated
with lymphopenia and thrombocytopenia, although the
mechanism is not clear.2
Dialysis prescription for specific drugs
and toxins
The following section discusses the characteristics of selected drugs and toxins most commonly seen as causes of
overdoses, and the dialysis prescriptions recommended
in the literature for their removal.
Lithium
Lithium carbonate is a highly effective mood stabilizer; this
low molecular weight cation (MW LiHCO3 = 74 Da) is water soluble, with a volume of distribution of 0.7–0.9 L/kg.
It is highly dialyzable.2 Most intoxication is chronic in the
setting of renal failure and volume depletion while the patient is also taking diuretics, nonsteroidal anti-inflammatory drugs, or angiotensin-converting enzyme inhibitors.2
Mild intoxication occurs at a serum lithium concentration of 1.5–2.5 mmol/L, while moderate intoxication occurs at a serum lithium concentration of 2.5–3.5 mmol/L
(therapeutic is o0.6 mmol/L). Both are characterized by
neuromuscular irritability, nausea, and diarrhea. Severe
lithium intoxication occurs at a serum concentration
43.5 mmol/L and is characterized by seizure, stupor,
and permanent neurologic damage.
162
Renal replacement therapy is recommended when:
the serum lithium level is 43.5 mmol/L;
the serum lithium level is 42.5 mmol/L with symptoms and a depressed GFR;
there is moderate intoxication with the expectation
that the lithium levels will increase or that they will
not decline to the therapeutic range within 36 hours.2
Both hemodialysis and CRRT provide far superior
clearance to native renal function even when there is no
evidence of acute kidney injury, with an extraction ratio
during hemodialysis of 90%.17 At blood flows of 250 mL/
min, hemodialysis provides 70–170 mL/min of clearance
compared with 10–40 min/mL in patients with normal
renal function. Hemodialysis is superior to CRRT, which
provides a clearance of around 48–62 mL/min at blood
flows of 200 mL/min and is thus not recommended in the
literature as first-line therapy.2 The literature recommends
dialysis for 6–8 hours with repeat sessions until the level
declines to below 1 mmol/L.2 Other experts suggest, however, that CRRT may be used as the first-line therapy in
the event the physician is concerned about rebound toxicity, envisions a long treatment, or is concerned about
the effects of fluid removal on patient hemodynamics.
The two modalities may be used together: CRRT may be
followed by hemodialysis if hypotension resolves, or
hemodialysis may be followed by a longer, slower session of CRRT. But there have been no head-to-head randomized trials to resolve the issue definitively.23
Hemodialysis International 2010; 14:158–167
Dialysis in the poisoned patient
Ethylene glycol and methanol
Ethylene glycol and methanol are volatile alcohols with
volumes of distribution equal to that of total body water.
Both ethylene glycol (MW 62 Da) and methanol (MW
32 Da) are metabolized by alcohol dehydrogenase to their
toxic metabolites—glycolic acid in the case of ethylene
glycol and formic acid in the case of methanol. Glycolic
acid is further metabolized to oxalate, which deposits in
renal tubules as envelope and needle-shaped crystals. The
exact cause of acute kidney injury is not known. Some
argue that it is related to glycolic acid;24 others argue that
calcium oxalate is the causative agent.25
Co-ingestion with ethanol, which binds competitively to
alcohol dehydrogenase, slows the metabolism of the other
alcohols to their toxic metabolites, but can also conceal
their presence, particularly because all 3 can present with
central nervous system depression and an anion gap
metabolic acidosis.19 Acute kidney injury can delay
native clearance of glycolic acid, leading to coma and
cardiac toxicity, and of formic acid, leading to CNS
depression and blindness from oxidative damage to the
retina. Patients initially develop an osmolar gap and then,
as the alcohol is metabolized to an acid, an anion gap.26
Thus, a normal osmolar gap in the presence of an anion
gap may indicate that the alcohol has been metabolized,
while a normal anion gap in the presence of an osmolar
gap may indicate that the ingestion is recent.
The first line of treatment in the case of ethylene glycol
or methanol poisoning is to block the metabolism of the
alcohols to their toxic metabolites. The antidote 4-methylpyrazole (fomepizole) has emerged as the primary
inhibitor of alcohol dehydrogenase.27 The American
Academy of Toxicology recommends treatment with
fomepizole rather than ethanol unless fomepizole is not
available or causes an allergic reaction. The guidelines for
starting fomepizole are similar for both alcohols:
documented ingestion of the alcohol and a concentration of 420 mg/dL;
documented recent ingestion of the alcohol with an
osmolar gap 410 mOsm/L;
strong clinical suspicion of ingestion of the alcohol
and two of the following—arterial pHo7.3, serum
HCO3 o20 mEg/L, or an osmolar gap 420 mOsm/L
(or urinary oxalate crystals in the case of ethylene
glycol).28,29
Despite the use of fomepizole, dialytic removal of
methanol or ethylene glycol may be necessary in the
event of:
severe metabolic acidosis with the pHo7.25;
Hemodialysis International 2010; 14:158–167
acute kidney injury or electrolyte imbalances that do
not respond to conventional therapy; and
deteriorating vital signs despite intensive care.
The guidelines also indicate dialysis for an alcohol level
450 mg/dL unless fomepizole is administered, the patient is asymptomatic, and the pH is normal. Further metabolism of the toxic metabolites to nontoxic byproducts
should be enhanced by giving pyridoxine and thiamine to
patients with ethylene glycol poisoning and folate or
folinic acid to patients with methanol poisoning.28,29
In the case of ethylene glycol, dialysis is undertaken
with a bicarbonate dialysate bath until the toxic alcohol
level is o20 mg/dL. If one is unable to obtain an alcohol
level, then dialysis should be carried out for at least 8
hours and repeated in another 12 hours.2 Glycolic acid
has a long half-life and a slow elimination rate. But it also
has a low molecular weight (76 Da), is water soluble, and
has a limited volume of distribution (0.55 L/kg). Thus,
even in the event of a low ethylene glycol level, an increased anion gap correlates with glycolic acid levels; dialysis should be undertaken because it increases clearance
considerably.30 Hemodialysis provides a clearance of
200–250 mL/min for ethylene glycol and 170 mL/min
for glycolate at blood flow rates of 250–400 mL/min.
The elimination half-life of glycolic acid on dialysis is
155 min compared with 626 min in patients not treated
by dialysis.30
With very high methanol levels, dialysis may need to
be undertaken for 18–21 hours and the procedure repeated after 24 hours if there is rebound toxicity. Most
experts recommend against using heparin anticoagulation
during dialysis because of the risk of intracranial hemorrhage, particularly in the basal ganglia.31,32 Hemodialysis
provides a clearance of 200 mL/min for methanol and
223 mL/min for formate at blood flows of 100–400 mL/
min. But the endogenous elimination half-life of formate
was not found to be statistically different from the elimination half-life on dialysis (205 90 vs. 150 37 min)
in one study. This may be in part due to other treatments
given to patients—folate or folinic acid—to increase the
metabolism of formic acid to carbon dioxide and water.33
Other researchers, however, have found that hemodialysis significantly reduced formate elimination half-life
compared with endogenous elimination and question
the role of folate.34
Some experts suggest that early intervention with
fomepizole and supportive care like sodium bicarbonate
administration to correct the acidosis may lessen the need
for hemodialysis in selected cases of methanol intoxication—those who present late with acidemia and no
163
Bayliss
detectable methanol level—because elimination of formate through dialysis is not significantly different from
endogenous clearance with supportive care.19 Others argue that dialysis can be avoided in stable patients with
less metabolic acidosis and no visual symptoms who can
be treated with fomepizole alone.35 But fomepizole therapy should be continued even after extracorporeal removal has begun and for several hours after treatment has
been completed to account for rebound.28
Very little data exist to show how CRRT compares with
hemodialysis in clearing volatile alcohols. A series of 3
patients with methanol ingestion found CRRT clearance
of 45 and 48 mL/min at a blood flow rate of 150 mL/min
over 37 and 27 hours in two patients compared with
clearance on hemodialysis of 237 mL/min at a blood flow
rate of 350 mL/min over 4 hours.36
One group has developed an equation to estimate the
time required on dialysis to reduce the levels of methanol
and ethylene glycol to 5 mmol/L. The equation is based
on the assumption that toxic alcohols have a dialysis
clearance similar to that of urea37:
t ¼ ½V ln ð5=AÞ=0:06k;
where t is time, V is the Watson estimate of total body
water,38 A is the initial toxin concentration in mmol/L,
and k is equal to 80% of the manufacturer-specified dialyzer urea clearance in milliliters per minute at the initial
observed blood flow rate. Validated in a small study of 13
patients, the equation requires only 3 measurements of
the toxic alcohol: an initial toxin level, 1 drawn 2 hours
before dialysis is due to stop, and a level drawn 1–2 hours
after dialysis.39
Salicylates
Acute salicylate intoxication is characterized in adults by
a respiratory alkalosis as salicylates stimulate the respiratory center in the medulla and then a metabolic acidosis
as lactic acid builds up from interruption by salicylates of
the mitochondrial electron transport chain.40 CNS symptoms indicate severe poisoning.
An immediate goal of treatment is to reverse the acidosis and alkalosis to prevent respiratory collapse: patients with salicylate intoxication do very poorly on
mechanical ventilation because of the ventilator’s inability to keep up with the patient’s respiratory rate and
the development of respiratory dys-synchronization, although there is little empiric evidence to support this in
the opinion of some experts. One retrospective review of
patients with salicylate intoxication who had been placed
on mechanical ventilation at conventional settings found
164
development of a respiratory acidosis, which led to overall worsening of neurologic toxicity by allowing more
salicylate to pass into the CNS.41 On the other hand, the
older literature cites the case of a child with salicylate
poisoning who was successfully treated with intermittent
positive pressure ventilation after he was paralyzed,42 and
animal experiments on acute respiratory failure following
salicylate-induced hyperventilation found no pulmonary
or arterial blood gas abnormalities in 5 sheep that were
paralyzed and ventilated at normal tidal volume and respiratory rate.43
Alkalinization of the urine is paramount if the patient is
still making large amounts of urine to prevent reabsorption of salicylates,44 and potassium administration prevents hypokalemia and acidification of the urine.9 At a
low pH, more salicylate crosses the blood–brain barrier
and increases CNS intoxication.45 How urinary alkalinization increases salicylate excretion is not entirely clear.
The conventional view is that weak acids like salicylates
become more ionized in an alkaline environment, and
urine alkalinization is recommended for salicylate intoxication unless there are contraindications (renal failure
and heart disease).6 Given its high dissociation constant
as a weak acid, a small increase in urinary pH should
produce a large increase in the ionized fraction of salicylate. But at a urine pH = 5, salicylate is already almost
completely ionized so that increasing pH further would
not increase its ionization further.46
Dialysis is recommended in the event of:
CNS depression at a level 450 mg/dL;
salicylate level 480 mg/dL;
salicylate-induced pulmonary edema.47
Salicylates have a molecular weight of 38–180 Da and a
volume of distribution of 0.17 L/kg. At therapeutic levels,
90% of salicylates are protein bound, but the unbound
fraction increases as the total concentration increases, as
in the case of intentional overdose. Intermittent hemodialysis is preferred over CRRT unless the patient is hypotensive because 50% of protein-bound aspirin is removed
by hemodialysis. The volume of distribution may be difficult to calculate when salicylate is ingested as entericcoated aspirin.48,49
Dialysis is undertaken against a high-flux membrane at
high blood flow rates with continued additional sodium
bicarbonate infusion because patients are often intravascularly depleted and require volume replacement, not
volume removal. Dialysis is continued until the serum
salicylate level is o10 mg/dL if there is concern for rebound toxicity, a possibility if the patient has developed a
bezoar.45
Hemodialysis International 2010; 14:158–167
Dialysis in the poisoned patient
CONCLUSION
Dialysis can be a life-saving therapy in the case of accidental or intentional overdoses of toxins that are amenable to extracorporeal removal such as salicylates, lithium,
and volatile alcohols, with most authorities recommending intermittent hemodialysis with high-efficiency, highflux membranes, except in cases of hypotension, when
CRRT may be necessary because of lower blood flow
rates. Use of sustained low-efficiency dialysis has also
been reported in a case of acute aspirin overdose.51
The patient in the clinical vignette suffered from the
acute effects of aspirin overdose and was at a risk of respiratory collapse if aggressive measures were not taken
to reduce salicylate levels. A temporary dialysis catheter
was placed in the emergency department, and the patient
was transferred to the medical intensive care unit. She
underwent 8 hours of hemodialysis with a blood flow of
350–400 mL/min and a dialysate flow rate of 500 mL/min
against a 35 mEq/L bicarbonate bath. She continued to
receive an infusion of isotonic sodium bicarbonate in 5%
dextrose in water during dialysis.52 She was never intubated. Her breathing improved as hemodialysis progressed, and she was maintained with supplemental
oxygen, first by a face mask and then by a nasal cannula.
Tachycardia also resolved.
Roughly 6 hours after initiation of hemodialysis, the
salicylate level was 30 mg/dL, and at 8 hours, it was
o10 mg/dL, undetectable (see Figure 1), and dialysis was
stopped as the patient became hypotensive. The salicylate
level rebounded slightly but declined to undetectable levels and remained there without further dialysis. The patient became more alert and acknowledged having taken
200 aspirin tablets in a suicide attempt. Her course was
complicated by upper gastrointestinal tract bleeding,
found to be due to erosive gastritis, and requiring transfusion of 2 U of packed red blood cells. Nevertheless, her
renal function returned to baseline, with a total urine
Hemodialysis International 2010; 14:158–167
160
140
Salicylate (mg/dL)
Acetaminophen
Acetaminophen is a moderately water-soluble 151 Da
molecule that is only slightly protein bound. The treatment of choice remains activated charcoal if the patient is
able to take food and fluids by mouth and N-acetyl
cysteine either orally or intravenously if the patient is
unable to swallow. N-acetyl cysteine delivered according
to the Rumack-Matthew nomogram for hepatotoxicity
helps replete stores of glutathione. Case reports suggest
the possibility of using single-pass albumin dialysis using
a high-flux membrane and dialysate to treat the resulting
liver failure.50
134.4
120
100
80
60
40
20
30.5
14.5
10.5
11
10.1
0
9/2/2008 9/2/2008 9/2/2008 9/3/2008 9/3/2008 9/3/2008
0020
1441
2220
0020
0112
0312
Time
Figure 1 Serum salicylate levels as a function of time in a
51-year-old woman presenting after an intentional aspirin
overdose. Initial measurement is from an earlier presentation
to the same emergency room.
output of 2.2 L over her first hospital night. The dialysis
catheter was removed, and the patient was transferred to
a general medical floor in stable condition.
Manuscript received August 2009; revised October 2009
REFERENCES
1 Brent J. Fomepizole for ethylene glycol and methanol
poisoning. N Engl J Med. 2009; 360:2216–2213.
2 Winchester JF, Boldur A, Oleru C, Kitiyakara C. Use of
dialysis and hemoperfusion in treatment of poisoning.
In: Daugirdas JT, et al., ed. Handbook of Dialysis. 4th ed.
Philadelphia: Walters Kluwer; 2007:300–319.
3 Cohen DL, Post J, Ferroggiaro AA, Perrone J, Foster MH.
Chronic salicylism resulting in pulmonary edema requiring hemodialysis. Am J Kidney Dis. 2000; 36:E20.
4 Rauschka H, Aboul-Enein F, Bauer J, et al. Acute cerebral
white matter damage in lethal salicylate intoxication.
Neurotoxicology. 2007; 28:33–37.
5 Kaplan AA. Extracorporeal blood purification in the
treatment of acute renal failure with multi-organ involvement. Blood Purif. 1996; 14:86–93.
6 Olyaei AJ, De Mattos AM, Bennett WM. Use of drugs
in patients with renal failure. In: Schrier RW, et al., ed.
Diseases of the Kidney and Urinary Tract. Philadelphia:
Lippincot Williams & Wilkins; 2007:2765–2807.
7 Goodman JW, Goldfarb DS. The role of continuous renal
replacement therapy in the treatment of poisoning. Semin
Dial. 2006; 19:402–407.
8 Clark WR, Henderson LW. Renal versus continuous versus intermittent therapies for the removal of uremic toxins. Kidney Int. 2001; 59(Suppl 78):S298–S303.
9 Daugirdas JT. Physiologic principles and urea kinetic modeling. In: Daugurdas JT, et al., ed. Handbook of Dialysis.
4th ed. Philadelphia: Walters Kluwer; 2007:25–58.
165
Bayliss
10 Tucker GT. Measurement of the renal clearance of drugs.
Br J Clin Pharmacol. 1981; 12:761–770.
11 Proudfoot AT, Krenzelok EP, Vale JA. Position paper
on urine alkalinization. J Toxicol Clin Toxicol. 2004;
42:1–26.
12 Ledebo I. Does convective dialysis therapy applied daily
approach renal blood purification? Kidney Int. 2001;
59(Suppl 78):S286–S291.
13 Van Laecke S, de Wilde K, Vanholder R. Online hemodiafiltration. Artif Organs. 2006; 30:579–585.
14 Boyd RF, Langsdorf LJ, Zydney AL. A two layer model for
the affects of blood contact on membrane transport in
artificial organs. ASAIO J. 1994; 40:M864–M869.
15 De Smet R, Dhondt A, Eloot S, et al. Effect of the superflux cellulose triacetate dialyser membranes on the removal of non-protein-bound and protein-bound uraemic
solutes. Nephrol Dial Transplant. 2007; 22:2006–2012.
16 Attaluri AC, Huang Z, Belwalkar A, et al. Evaluation of
nano-porous alumina membranes for hemodialysis application. ASAIO J. 2009; 55:217–223.
17 Garella S. Extracorporeal techniques in the treatment of
exogenous intoxications. Kidney Int. 1988; 33:735–754.
18 Churchwell MD, Pasko DA, Smoyer WE, Mueller BA.
Enhanced clearance of highly protein-bound drugs by
albumin-supplemented dialysate during modeled continuous hemodialysis. Nephrol Dial Transplant. 2009;
24:231–238.
19 Borkan SC. Extracorporeal therapies for acute intoxications. Crit Care Clin. 2002; 18:393–420.
20 Winters JL, Pineda AA, McLeod BC, Grima KM. Therapeutic apheresis in renal and metabolic diseases. J Clin
Apheresis. 2000; 15:53–73.
21 Nenov VD, Mrinov P, Sabeva J, Nenov DS. Current
applications of plasmapheresis in clinical toxicology.
Nephrol Dial Transplant. 2003; 18(Suppl 5):v56–v58.
22 Holubek WJ, Hoffman RS, Goldfarb DS, Nelson LS. Use
of hemodialysis and hemoperfusion in poisoned patients. Kidney Int. 2008; 74:1327–1334.
23 Feinfeld DA, Rosenberg JW, Winchester JF. Three controversial issues in extracorporeal toxin removal. Semin
Dial. 2006; 19:358–362.
24 Don BR, Rodriguez RA, Humphreys MH. Acute renal failure associated with pigmenturia or crystal deposits. In:
Shreier RW, et al., ed. Diseases of the Kidney and Urinary
Tract. Philadelphia: Lippincot Williams & Wilkins;
2007:1184–1207.
25 Guo C, Cenac TA, Li Y, McMartin KE. Calcium oxalate,
and not other metabolites, is responsible for the renal
toxicity of ethylene glycol. Toxicol Lett. 2007; 173:8–16.
26 Kraut JA, Kurtz I. Toxic alcohol ingestions: Clinical features, diagnosis and management. Clin J Am Soc Nephrol.
2008; 3:208–225.
27 Brent J, McMartin K, Phillips S, et al. Fomepizole for the
treatment of ethylene glycol poisoning. N Eng J Med.
1999; 340:832–838.
166
28 Barceloux DG, Krenzelok EP, Olson K, et al. American
Academy of Toxicology practice guidelines on the treatment of ethylene glycol poisoning. J Toxicol Clin Toxicol.
1999; 37:537–560.
29 Barceloux DG, Bond GR, Krenzelok EP, et al. American
Academy of Toxicology practice guidelines on the treatment of methanol poisoning. J Toxicol Clin Toxicol. 2002;
40:415–446.
30 Moreau CL, Kerns W, Tomaszewski CA, et al. Glycolate
kinetics and hemodialysis clearance in ethylene glycol
poisoning. J Toxicol Clin Toxicol. 1998; 36:659–666.
31 Phang PT, Passerini L, Mielke B, et al. Brain hemorrhage
associated with methanol poisoning. Crit Care Med.
1988; 16:137–142.
32 Hernandez MA, Holanda MS, Tejerina EE, et al. Methanol poisoning and heparin: A dangerous couple? Am J
Emer Med. 2004; 22:620–621.
33 Kerns W, Tomaszewski C, Mcmartin K, et al. Formate
kinetics in methanol poisoning. J Toxicol Clin Toxicol.
2002; 40:137–143.
34 Hantson P, Haufroid V, Wallemacq P. Formate kinetics
in methanol poisoning. Hum Exp Toxicol. 2005; 24:55–
59.
35 Hovda KA, Jacobsen D. Expert opinion: Fomepizole may
ameliorate the need for hemodialysis in methanol poisoning. Hum Exp Toxicol. 2008; 27:539–546.
36 Kan G, Jenkins I, Rangan G, et al. Continuous haemodiafiltration compared with intermittent haemodialysis in
the treatment of methanol poisoning. Nephrol Dial Transplant. 2003; 18:2665–2667.
37 Hirsch DJ, Jindal KK, Wong P, Fraser AD. A simple
method to estimate the required dialysis time for
cases of alcohol poisoning. Kidney Int. 2001; 60:2021–
2024.
38 Watson PE, Watson ID, Batt RD. Total body water for
adult males and females estimated from simple anthropometric measurements. Am J Clin Nutr. 1980; 33:
27–39.
39 Youssef GM, Hirsch DJ. Validation of a method to predict
required dialysis time for cases of methanol and ethylene
glycol poisoning. Am J Kidney Dis. 2005; 46:509–511.
40 O’Malley GF. Emergency department management of the
salicylate-poisoned patient. Emerg Med Clin N Am. 2007;
25:333–346.
41 Stolbach AI, Hoffman RS, Nelson LS. Mechanical ventilation was associated with acidemia in a case series of
salicylate-poisoned patients. Acad Emerg Med. 2008;
15:866–869.
42 Freier S, Neal BW, Nisbet HIA, Rees GJ, Wilson F. Salicylate intoxication treated with intermittent positivepressure ventilation. Br Med J. 1957; 1:1333–1335.
43 Mascheroni D, Kolobow T, Fumagalli R, et al. Acute respiratory failure following pharmacologically induced
hyperventilation: An experimental animal study. Intensive Care Med. 1988; 15:8–14.
Hemodialysis International 2010; 14:158–167
Dialysis in the poisoned patient
44 Higgins RM, Connolly JO, Hendry BM. Alkalinization
and hemodialysis in severe salicylate poisoning: Comparison of elimination techniques in the same patient.
Clin Nephrol. 1998; 50:178–183.
45 Dargan PI, Wallace CI, Jones AL. An evidence based
flowchart to guide the management of acute salicylate
(aspirin) overdose. Emerg Med J. 2002; 19:206–209.
46 Proudooft AT, Krenzelok EP, Brent J, Vale JA. Does urine
alkalinization increase salicylate elimination? If so, why?
Toxicol Rev. 2003; 22:129–136.
47 Cohen DL, Post J, Ferroggiaro AA, Peron J, Foster MH.
Chronic salicylism resulting in noncardiogenic pulmonary edema requiring hemodialysis. Am J Kidney Dis.
2000; 36:E20.
48 Dugandzic RM, Tierney MG, Dickinson GE, et al. Evaluation of the validity of the Done nomogram in the man-
Hemodialysis International 2010; 14:158–167
49
50
51
52
agement of acute salicylate intoxication. Ann Emerg Med.
1989; 18:1186–1190.
Pierce RP, Gazewood J, Blake RL Jr Salicylate poisoning
for enteric-coated aspirin. Delayed absorption may complicate management. Postgrad Med. 1991; 89:61–62, 64.
Karvellas CJ, Bagshaw SM, McDermid RC, et al. Acetaminophen-induced liver failure treated with single-pass
albumin dialysis: Report of a case. Int J Artif Organs.
2008; 31:450–455.
Lund B, Seifert SA, Matersohn M. Efficacy of sustained
low-efficiency dialysis in the treatment of salicylate toxicity. Neph Dial Transplant. 2005; 20:1483–1484.
Higgins RM, Connolly JO, Hendry BM. Alkalinization
and hemodialysis in severe salicylate poisoning: Comparison of elimination techniques in the same patient.
Clin Nephrol. 1998; 50:178–183.
167