case records of the massachusetts general hospital
T h e n e w e ng l a nd j o u r na l o f m e dic i n e
n engl j med 354;10 www.nejm.org march 9, 2006 1065
Founded by Richard C. Cabot
Nancy Lee Harris, m.d., Editor Sally H. Ebeling, Assistant Editor
Jo-Anne O. Shepard, m.d., Associate Editor Christine C. Peters, Assistant Editor
Case 7-2006: A 47-Year-Old Man with Altered
Mental Status and Acute Renal Failure
J. Kimo Takayesu, M.D., M.Sc., Hasan Bazari, M.D., and Michael Linshaw, M.D.
From the Department of Surgery ( J.K.T.)
and the Nephrology Division, Depart-
ment of Medicine (H.B.), Massachusetts
General Hospital; and the Department of
Surgery (J.K.T.), the Nephrology Division,
Department of Medicine (H.B.), and the
Nephrology Division, Department of Pe-
diatrics (M.L.), Harvard Medical School.
N Engl J Med 2006;354:1065-72.
Copyright © 2006 Massachusetts Medical Society.
Pr esen tat ion of C a se
A 47-year-old man was transferred to the emergency department of this hospital at
10 a.m. on a day in early June because of altered mental status and acute renal fail-
ure. At approximately 8 p.m. the previous evening, the patient had been behaving
normally at dinner with his family. After dinner, he went outside to work on his car.
According to his wife, when he later returned to the house, his speech was slurred
and he was lethargic. By 9 p.m., the patient was vomiting and becoming increas-
ingly lethargic. He went to bed, and at 3:30 a.m., his wife found him unresponsive.
She called emergency medical services, and the patient was taken to the emergency
department of a local hospital.
The patient had not had recent fevers, illnesses, or depressive symptoms. He had
irritable bowel syndrome (for which he took atropine and diphenoxylate), chronic
back pain, and anxiety. He was allergic to penicillin. There was no personal or fam-
ily history of major medical problems. He had a 30 pack-year history of smoking
and a history of alcohol abuse, but he had not consumed alcohol in the past year.
His wife did not believe that he used illicit drugs. There were no empty pill bottles
found in the house, and the patient did not have known access to other prescrip-
tion medications.
On arrival in the emergency department, the patient was somnolent and unable
to follow commands; the blood pressure was 85/40 mm Hg, and the heart rate 75
beats per minute. His mental status worsened, and the trachea was intubated for
airway protection with use of rapid-sequence induction with 20 mg of etomidate
and 120 mg of succinylcholine. A chest radiograph showed an infiltrate in the left
lower lobe that was consistent with pneumonia. A computed tomographic (CT) scan
of the head revealed no abnormalities except for an air–fluid level in the left mas-
toid air cells that was suggestive of mastoiditis. An orogastric tube was placed, and
50 g of activated charcoal was given; in addition, 900 mg of clindamycin and 500 mg
of metronidazole were given intravenously for possible aspiration pneumonia. After
endotracheal intubation, the arterial pH was 6.97, the partial pressure of oxygen
182 mm Hg, and the partial pressure of carbon dioxide 34 mm Hg; the level of
carbon monoxide was undetectable. The creatinine level was 3.8 mg per deciliter
(336 μmol per liter), and the white-cell count was 30,000 per cubic millimeter. A
continuous intravenous infusion of sodium bicarbonate (150 mmol per liter) in a
Downloaded from www.nejm.org on January 24, 2010 . Copyright © 2006 Massachusetts Medical Society. All rights reserved.
T h e n e w e ng l a nd j o u r na l o f m e dic i n e
n engl j med 354;10 www.nejm.org march 9, 20061066
solution of 5 percent dextrose in water was started
at a rate of 250 ml per hour, and the patient was
transferred to the emergency department of this
hospital.
On arrival at this hospital, the patient was in-
tubated and sedated and unresponsive to painful
stimuli. The patient’s blood pressure was 137/88
mm Hg, the heart rate 80 beats per minute, and
the temperature 36.2°C; he was ventilated at a rate
of 30 breaths per minute with an oxygen satura-
tion of 99 percent and a fraction of inspired oxygen
of 1.0. His pupils were equal, round, and reactive
at 3 mm. The corneal reflex was present, and the
vestibulo-ocular reflex was absent. An endotra-
cheal tube was in place. The neck was supple
without lymphadenopathy. Rhonchi were present
bilaterally. Cardiac auscultation revealed no abnor-
malities. The abdomen was soft, with active bowel
sounds. A Foley catheter was draining clear urine.
The rest of the examination was normal.
The placement of the endotracheal tube was
confirmed by bedside end-tidal carbon dioxide
calorimetry. A central venous pressure line was
placed in the right internal jugular vein, and in
the process a transient run of ventricular tachycar-
dia occurred, which terminated with retraction of
the guidewire. A chest radiograph obtained with
a portable device showed no pneumothorax and
a small left perihilar opacity. The central venous
pressure was 10 to 13 mm Hg. The white-cell
count was 27,000 per cubic millimeter, with 95
percent neutrophils; the rest of the complete blood
count was normal. The results of other labora-
tory tests are shown in Table 1. After a review of
the CT scan obtained at the first hospital, van-
comycin (1 g) and ceftriaxone (2 g) were admin-
istered intravenously for empirical treatment of
possible meningitis. An electrocardiogram showed
normal sinus rhythm without marked prolonga-
tion of the QRS complex or the QT interval.
Table 1. Results of Laboratory Tests.*
Variable Normal Range Day 1 Day 2 Day 8 Day 20
1 hr, 4 min
after
Admission
1 hr, 9 min
after
Admission
5 hr, 48 min
after
Admission
12 hr, 35 min
after
Admission
Sodium (mmol/liter) 135–145 147 149 153 140 137 136 142
Potassium (mmol/liter) 3.4–4.8 4.1 2.4 2.6 4.8 3.8 5.2 5.1
Chloride (mmol/liter) 100–108 110 107 100 103 108
Carbon dioxide (mmol/liter) 23–31.9 13.6 25.3 31.5 17.8 26.8
Urea nitrogen (mg/dl) 8–25 23 17 15 96 56
Creatinine (mg/dl) 0.6–1.5 2.8 2.7 2.6 12.6 3.6
Glucose (mg/dl) 70–110 208 141 99 163 130 115 90
Calcium (mg/dl) 8.5–10.5 7.6 8.8 8.0 8.6 9.1
Calcium, ionized (mmol/
liter)
1.14–1.30 0.74 0.55 1.22
Phosphorus (mg/dl) 2.6–4.5 4.3 2.8 1.1 7.0 5.7
Magnesium (mEq/liter) 1.4–2.0 1.5 1.3 1.3 2.1 1.6
Fraction of inspired oxygen
(per liter)
No normal value 0.7 0.5
Arterial pH 7.35–7.45 7.10 7.34 7.45
Arterial partial pressure of
carbon dioxide (mm Hg)
35–42 33 28 36
Arterial partial pressure of
oxygen (mm Hg)
80–100 139 234 105
Osmolality (mOsm/kg) 280–296 321
* To convert the values for urea nitrogen to millimoles per liter, multiply by 0.357. To convert the values for creatinine to micromoles per liter,
multiply by 88.4. To convert the values for glucose to millimoles per liter, multiply by 0.5551. To convert the values for calcium to millimoles
per liter, multiply by 0.250. To convert the values for phosphorus to millimoles per liter, multiply by 0.3229. To convert the values for mag-
nesium to millimoles per liter, multiply by 0.500.
Downloaded from www.nejm.org on January 24, 2010 . Copyright © 2006 Massachusetts Medical Society. All rights reserved.
case records of the massachusetts gener al hospital
n engl j med 354;10 www.nejm.org march 9, 2006 1067
Approximately 30 minutes after his arrival at
this hospital, the patient had a generalized tonic–
clonic seizure. Lorazepam was administered in-
travenously every two minutes in doses of 2 mg
up to a total of 20 mg. After 10 minutes of con-
tinuous seizure activity, phenytoin (1 g) was ad-
ministered intravenously at a rate of 50 mg per
minute. Administration of phenobarbital was in
preparation when the seizure activity ceased at 20
minutes. The results of laboratory tests are shown
in Table 1. The anion gap was calculated at 25
mmol per liter. The measured serum osmolality
was 321 mOsm per kilogram, and the patient’s
initial osmolal gap was calculated at 8 mOsm per
kilogram. Calcium gluconate (2 g), potassium
chloride (1 liter of a 40-mmol solution in normal
saline), and magnesium sulfate (2 g) were given.
Toxicologic screening of serum was negative for
ethanol, acetaminophen, and salicylates, and toxi-
cologic screening of urine was also negative. Nu-
merous needle-shaped crystals were seen in the
urinary sediment.
Fomepizole (1 g; 15 mg per kilogram of body
weight) was administered intravenously, and the
nephrology division was consulted for immediate
hemodialysis. A double-lumen hemodialysis cath-
eter was placed in the right femoral vein, and
hemodialysis was initiated. Results of laboratory
tests obtained five hours after admission are
shown in Table 1. The patient was admitted to the
medical intensive care unit for further care.
Differ en ti a l Di agnosis
Dr. J. Kimo Takayesu: I was involved in the case of
this 47-year-old man, who had an acute onset of
altered mental status, acute renal failure, and a
severe metabolic acidosis associated with an an-
ion gap. These findings initially gave rise to a
broad differential diagnosis, which was quickly
narrowed by the additional finding of typical
crystals of calcium oxalate in a urine specimen.
The Anion Gap
The anion gap is the difference between the mea-
sured concentrations of serum sodium and the
total of the main serum anions chloride and bi-
carbonate (anion gap = Na – Cl + HCO3); a normal
anion gap is caused by the presence of negatively
charged proteins that are not measured by serum
analyzers, mainly albumin, and ranges from 3 to
11 mmol per liter.1,2 An elevated anion gap is
caused by either an excess of unmeasured serum
anions in addition to albumin or a paucity of un-
measured cations such as calcium or magnesium.
This patient had a profoundly elevated anion gap
of 25 mmol per liter, which was not explained by
the degree of hypocalcemia and hypomagnese-
mia, and thus prompted a search for the presence
of unmeasured serum anions.
Causes and effects of anion-gap metabolic
acidosis
Anion-gap metabolic acidosis has four main
causes: lactic acidosis, ketoacidosis, renal failure,
and ingested toxins and their metabolites (Table
2). An anion gap of 25 mmol per liter correlates
strongly with the presence of one of these condi-
tions3 and is often due to multiple concurrent con-
ditions. This patient’s acidosis was probably caused
by a combination of lactic acidosis from periph-
eral hypoperfusion and tissue hypoxia, acute re-
nal failure, and an ingested toxin. The physiolog-
ical effects of metabolic acidosis include a
reflexive increase in respiratory drive, depressed
cardiac contractility, peripheral arterial dilata-
tion, increased susceptibility to cardiac arrhyth-
mias, central venoconstriction, a predisposition
to pulmonary edema, depression of the central
nervous system, and glucose intolerance. This pa-
Table 2. Four Main Causes of Anion-Gap Metabolic
Acidosis.
Lactate
Carbon monoxide
Cyanide
Isoniazid (>30 mg/kg)
Iron
Salicylates (cytochrome poisoning)
Metformin
Acute alcohol intoxication
Ketoacidosis
Diabetic ketoacidosis
Alcoholic ketoacidosis
Renal failure
Uremia, decreased secretion of ammonium, hydro-
gen sulfate, hydrogen phosphate
Toxins and metabolites
Toluene
Methanol, ethylene glycol, paraldehyde (metabolized
to formate, oxalate, acetate)
Downloaded from www.nejm.org on January 24, 2010 . Copyright © 2006 Massachusetts Medical Society. All rights reserved.
T h e n e w e ng l a nd j o u r na l o f m e dic i n e
n engl j med 354;10 www.nejm.org march 9, 20061068
tient was mechanically hyperventilated after in-
tubation in an attempt to maximize respiratory
compensation for his metabolic acidosis.4-6 His
hyperglycemia was probably a result of the infu-
sion of bicarbonate in 5 percent dextrose he was
receiving at the time of transfer, as well as a
stress response. The episode of ventricular tachy-
cardia that was precipitated by the guidewire in-
sertion during placement of the central venous
line may have been a result of the acidosis and
the electrolyte abnormalities.
This patient’s obtundation and hypotension
probably resulted in tissue hypoxia with lactate
overproduction, causing lactic acidosis. The treat-
ment for lactic acidosis relies on restoration of
adequate tissue perfusion through intravenous
volume repletion and supplemental oxygen, both
of which this patient received. The use of vaso-
pressors can also be beneficial if tissue hypoper-
fusion persists after intravenous volume repletion,
but these agents were not required in this case.7
A sodium bicarbonate infusion was begun; bi-
carbonate infusions may help to improve cardiac
function when the pH falls below 7.20.8,9 How-
ever, the infusion can also cause fluid overload
and paradoxical tissue acidosis in patients with
limited respiratory reserve, cardiac failure, or ar-
rest. When the patient arrived here, the pH had
corrected to 7.10, and the central venous pressure
and blood pressure were normal.
Diabetic ketoacidosis often occurs in patients
with insulin-dependent diabetes in whom an acute
physiological stressor results in the accumulation
of acetoacetate and β-hydroxybutyrate from fatty-
acid metabolism. This patient had no history of
insulin dependence, making this an unlikely cause
of his acidosis. He did have a remote history of
alcohol abuse, making alcoholic ketoacidosis a
possible contributing diagnosis. An abrupt dis-
continuance of alcohol consumption by a long-
time alcohol user results in a starvation state that
is worsened by subsequent vomiting and dehydra-
tion. The ketosis in this condition is predomi-
nantly from β-hydroxybutyrate, which is not de-
tected by the nitroprusside ketone reaction used
to detect acetoacetate. This patient received ap-
propriate treatment for alcoholic ketoacidosis,
including intravenous hydration with a dextrose
solution and electrolyte replacement; however,
given the acute onset of profound acidosis and
the history of recent sobriety, other diagnoses had
to be considered.
An important feature of this case is the new
onset of renal failure. Acidosis in renal failure is
principally due to an accumulation of acids and a
reduction of ammonium production due to de-
creased nephron mass. Acute renal failure typi-
cally presents with a combination of hyperchlo-
remic acidosis and anion-gap metabolic acidosis.
Bicarbonate levels usually remain greater than 15
mmol per liter, and the anion gap usually does not
exceed 20 mmol per liter. This patient had a large
anion gap, severe acidemia with a relatively low
creatinine level, and no history of previous renal
disease. Thus, although renal failure probably con-
tributed to this patient’s severe acidosis, it was
unlikely to be the primary cause.
Many toxins and drugs can induce acidosis.
Carbon monoxide poisoning causing lactic acido-
sis was ruled out by the other hospital. Salicylates
increase lactate production by poisoning mito-
chondrial cytochromes. However, in this patient
toxicologic screening did not show the presence
of salicylates. Toxic alcohols, such as methanol
and ethylene glycol, would be a primary consid-
eration in this case of anion-gap acidosis. They
are converted by alcohol dehydrogenase to toxic
anionic acids that increase the anion gap. In ad-
dition, these alcohols are osmotically active in
their native forms before they are metabolized,
and they cause an increase in the measured se-
rum osmolality. This osmolal increase is esti-
mated by the difference between the measured
osmolality and the calculated osmolality, deter-
mined as follows: (2 × the serum sodium level) +
(the blood urea nitrogen level ÷ 2.8) + (the blood
glucose level ÷ 18) — known as an osmolal gap.
When the osmolal gap exceeds 10 mmol per li-
ter, the presence of a toxic alcohol must be con-
sidered. This patient had an osmolal gap of only
8 mmol per liter, which could argue against the
presence of a toxic alcohol. However, by the time
he came to medical attention, it is probable that
most of the osmotically active alcohol would have
already been converted to an anionic metabolite,
which would not contribute to the serum osmo-
lality. The needle-shaped crystals found in the
urine on his arrival in the emergency department
of this hospital are characteristic of calcium oxa-
late, a metabolite of ethylene glycol; this finding,
together with the presence of anion-gap acido-
sis, altered mental status, and acute renal tubu-
lar injury, strongly supports the diagnosis of eth-
ylene glycol intoxication.
Downloaded from www.nejm.org on January 24, 2010 . Copyright © 2006 Massachusetts Medical Society. All rights reserved.
case records of the massachusetts gener al hospital
n engl j med 354;10 www.nejm.org march 9, 2006 1069
DR . J. K IMO TA K A Y ESU’S DI AGNOSIS
Ethylene glycol intoxication.
Pathol o gic a l Discussion
Dr. Michael Linshaw: The original sample of urinary
sediment from this patient was not available to
photograph. However, the morphologic structure
of urinary crystals can be helpful in establishing
a diagnosis. Calcium oxalate crystals are of two
types (Fig. 1): calcium oxalate dihydrate crystals,
which are typically octahedrons, and calcium oxa-
late monohydrate crystals, which are needle-shaped.
Calcium oxalate monohydrate crystals are not of-
ten seen in urinary sediment, but they are typical,
and therefore very suggestive, of ethylene glycol
ingestion. Calcium oxalate dihydrate crystals may
be seen in other conditions and are therefore less
specific for the diagnosis. If calcium oxalate crys-
tals are seen within a cast or are associated with
dysmorphic red cells (defined as those showing
marked variation in size and shape), this finding
suggests the possibility of crystal nephropathy.
Discussion of M a nagemen t
Dr. Hasan Bazari: This man presented with the triad
of depression of the central nervous system with
seizures, acute renal failure, and anion-gap met-
abolic acidosis. An evaluation in the emergency
department led to the diagnosis of ethylene gly-
col intoxication, and the nephrology division was
consulted regarding management.
The two alcohols that typically cause both a
severe anion-gap metabolic acidosis and an os-
molal gap are methanol and ethylene glycol.10
The osmolal gap is a useful way to detect the pres-
ence of alcohols, but it has several pitfalls. The
serum osmolality can be measured by determin-
ing either the freezing-point depression or the
boiling-point elevation (the vapor-pressure meth-
od). The latter is unreliable in the presence of
volatile substances such as ethanol and metha-
nol,11 but not with ethylene glycol because it has
a high boiling point.12 Once the osmolal gap is
calculated, if the ethanol concentration is known,
the presence of a second unmeasured osmole can
be deduced. However, ethylene glycol, with a rela-
tively high molecular weight of 62, contributes
fewer osmoles at a given serum concentration than
alcohols of lower molecular weight. Hence, at a
concentration of 21 mg per deciliter, the contri-
bution of ethylene glycol to the osmolality may be
only 4 mOsm per kilogram and at a lethal dose
of 50 mg per deciliter, the contribution to osmo-
lality may be only 8 mOsm per kilogram. Exces-
sive reliance on the osmolal gap can lead one to
omit consideration of ethylene glycol intoxication
if the gap is below that considered to be normal,
as happened in this case. Finally, the time that
has elapsed since the ingestion is also important.
Alcohols that are metabolized, such as ethylene
glycol, will contribute less to the osmolal gap and
more to the anion gap as they are metabolized,
which is what probably occurred in this patient.