Sie sind auf Seite 1von 12

The n e w e ng l a n d j o u r na l of m e dic i n e

review article
Disorders of Fluids and Electrolytes
Julie R. Ingelfinger, M.D., Editor

Physiological Approach to Assessment

of AcidBase Disturbances
Kenrick Berend, M.D., Ph.D., Aiko P.J. de Vries, M.D., Ph.D.,
and Rijk O.B. Gans, M.D., Ph.D.

From the Department of Internal Medi nternal acidbase homeostasis is fundamental for maintaining
cine, St. Elisabeth Hospital, Willemstad, life. Accurate and timely interpretation of an acidbase disorder can be lifesav-
Curaao (K.B.); and the Division of Ne
phrology, Department of Medicine, Leiden ing, but the establishment of a correct diagnosis may be challenging.1 The three
University Med ical Center, and Leiden major methods of quantifying acidbase disorders are the physiological approach,
University, Leiden (A.P.J.V.), and the De the base-excess approach, and the physicochemical approach (also called the Stew-
partment of Internal Medicine, University
of Groningen, University Medical Center art method).2 This article reviews a stepwise method for the physiological approach.
Groningen, Groningen (R.O.B.G.) The physiological approach uses the carbonic acidbicarbonate buffer system.
both in the Netherlands. Address reprint Based on the isohydric principle, this system characterizes acids as hydrogen-ion
requests to Dr. Berend at the Department
of Internal Medicine, St. Elisabeth Hospi donors and bases as hydrogen-ion acceptors. The carbonic acidbicarbonate sys-
tal, Breedestraat 193, Willemstad, Curaao, tem is important in maintaining homeostatic control. In the physiological ap-
or at proach, a primary change in the partial pressure of carbon dioxide (Pco2) causes
N Engl J Med 2014;371:1434-45. a secondary adaptive response in the bicarbonate concentration and vice versa;
DOI: 10.1056/NEJMra1003327 further changes in carbon dioxide or bicarbonate reflect additional changes in
Copyright 2014 Massachusetts Medical Society.
acidbase status. The four recognized primary acidbase disorders comprise two
metabolic disorders (acidosis and alkalosis) and two respiratory disorders (acidosis
and alkalosis).
The hydrogen-ion concentration is tightly regulated because changes in hydro-
gen ions alter virtually all protein and membrane functions.2-6 Since the concen-
tration of hydrogen ions in plasma is normally very low (approximately 40 nmol
per liter), the pH, which is the negative logarithm of the hydrogen-ion concentra-
tion, is generally used in clinical medicine to indicate acidbase status.3-5,7 The
terms acidemia and alkalemia refer to states in which the blood pH is abnor-
mally low (acidic) or abnormally high (alkaline). The process in which the hydro-
gen-ion concentration is increased is called acidosis, and the process in which the
hydrogen-ion concentration is decreased is called alkalosis.3,4 The traditional de-
termination of acidbase values is based on the HendersonHasselbalch equation
(in which pK denotes the acid dissociation constant):

pH=pK+log10 (bicarbonate [HCO3][0.03partial pressure of arterial carbon dioxide (Paco2)]),

where bicarbonate is in millimoles per liter and Paco2 is in millimeters of mer-

cury.6,7 An acidbase disorder is called respiratory when it is caused by a primary
abnormality in respiratory function (i.e., a change in the Paco2) and metabolic when
the primary change is attributed to a variation in the bicarbonate concentration.

His t or y a nd Ph ysic a l E x a minat ion

The first step in assessment of an acidbase disorder is a careful clinical evaluation.

Various signs and symptoms often provide clues regarding the underlying acid
base disorder; these include the patients vital signs (which may indicate shock or

1434 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

sepsis), neurologic state (consciousness vs. un-

Table 1. Primary AcidBase Disturbances with a Secondary (Compensatory)
consciousness), signs of infection (e.g., fever), Response.*
pulmonary status (respiratory rate and presence
Metabolic acidosis
or absence of Kussmaul respiration, cyanosis,
and clubbing of the fingers), and gastrointestinal pH <7.38 and bicarbonate [HCO3] <22 mmol per liter
symptoms (vomiting and diarrhea). Certain un- Secondary (respiratory) response: Paco2=1.5[HCO3]+82 mm Hg or
[HCO3] + 15 mm Hg
derlying medical conditions such as pregnancy,
diabetes, and heart, lung, liver, and kidney dis- Complete secondary adaptive response within 1224 hr
ease may also hint at the cause. The clinician Superimposed respiratory acidosis or alkalosis may be diagnosed if the calcu
should determine whether the patient has taken lated Paco2 is greater or less than predicted
any medications that affect acidbase balance Metabolic alkalosis
(e.g., laxatives, diuretics, topiramate, or metfor- pH >7.42 and [HCO3] >26 mmol per liter
min) and should consider signs of intoxication Secondary (respiratory) response: Paco2=0.7([HCO3]24)+402 mm Hg
that may be associated with acidbase distur- or [HCO3]+15 mm Hg or 0.7[HCO3]+20 mm Hg
bances (e.g., acetone fetor as a sign of diabetic Complete secondary adaptive response within 2436 hr
ketoacidosis or isopropyl alcohol intoxication, Superimposed respiratory acidosis or alkalosis may be diagnosed if the calcu
and visual disturbance as a symptom of metha- lated Paco2 is greater or less than predicted
nol intoxication). Respiratory acidosis
pH <7.38 and Paco2 >42 mm Hg
De ter minat ion of the Pr im a r y Secondary (metabolic) response
Acid B a se Disor der a nd the Acute: [HCO3] is increased by 1 mmol/liter for each Paco2 increase of
Sec onda r y R e sp onse 10 mm Hg above 40 mm Hg
Chronic: generally [HCO3] is increased by 45 mmol/liter for each Paco2
The second step is to determine the primary increase of 10 mm Hg above 40 mm Hg
acidbase disorder and the secondary response. Complete secondary adaptive response within 25 days
The range of pH that is compatible with life is
Superimposed metabolic alkalosis or acidosis may be diagnosed if the calcu
7.80 to 6.80 (a hydrogen-ion concentration [H+] lated [HCO3] is greater or less than predicted
of 16 to 160 nmol per liter).3 For the purposes
Respiratory alkalosis
of this review, the reference value for pH is
pH >7.42 and Paco2 <38 mm Hg
7.400.02, for Paco2, 382 mm Hg, and for
[HCO3], 242 mmol per liter. The four major Secondary (metabolic) response
acidbase disturbances are defined as primary Acute: [HCO3] is decreased by 2 mmol/liter for each Paco2 decrease of
10 mm Hg below 40 mm Hg
acidbase disorders (Table 1 and Fig. 1). Empirical
observations suggest that the homeostatic re- Chronic: [HCO3] is decreased by 45 mmol/liter for each Paco2 decrease
of 10 mm Hg below 40 mm Hg
sponse to acidbase disorders is predictable and
can be calculated.9-18 In response to metabolic Complete secondary adaptive response in 25 days
acidbase disturbances, changes in the respira- Superimposed metabolic alkalosis or acidosis may be diagnosed if the calcu
lated [HCO3] is greater or less than predicted
tory rate develop quickly, and a new steady-state
Paco2 is reached within hours. In cases of persis- * Reference values for arterial blood gases are the following: pH, 7.40.02, par
tent respiratory abnormalities, metabolic com- tial pressure of arterial carbon dioxide (Paco2), 402 mm Hg, and bicarbonate,
pensation develops slowly, and 2 to 5 days are 242 mmol per liter. Reference values for venous blood gases are the following:
pH, 7.36 to 7.38, Pvco2, 43 to 48 mm Hg, and bicarbonate, 25 to 26 mmol
required for the plasma bicarbonate concentra- per liter. To convert the values for PCO2 to kilopascals, divide by 7.5006.
tion to reach a new steady-state level. A respira- This formula is also known as the Winters formula.
tory change is called acute or chronic de- These calculations are easy to make at the bedside but are not reliable at all
bicarbonate concentrations. Data are from Berend.8
pending on whether a secondary change in the The secondary respiratory response is difficult to predict in metabolic alkalosis.
bicarbonate concentration meets certain criteria
(Table 1). Mixed acidbase disorders are diag-
nosed when the secondary response differs from are used to assess acidbase status are approxi-
that which would be expected.13,18-23 mations based on nearly 40-year-old studies in-
There are several caveats concerning compen- volving humans and dogs.1 Experimental studies
satory changes. Blood gas values can always be of severe chronic hypocapnia and hypercapnia in
explained by two or more coexisting acidbase humans are not ethically feasible; thus, data are
disorders.12 The current prediction equations that insufficient to construct confidence limits for

n engl j med 371;15nejm.orgoctober 9, 2014 1435

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

pH <7.38

Metabolic acidosis Respiratory acidosis

HCO3 <22 mmol/liter PaCO2 >42 mm Hg

Secondary (respiratory) response Secondary (metabolic) response

Calculate expected PaCO2 :1.5[HCO3 ]+82 mm Hg Observe measured [HCO3]
Observe measured values If there is a change in [HCO3] of
PaCO2 lower than expected: additional respiratory 1 mmol/liter increase per 10 mm Hg PaCO2
alkalosis increase above 40 mm Hg: acute respiratory
PaCO2 higher than expected: additional acidosis
respiratory acidosis <1 mmol/liter increase per 10 mm Hg PaCO2
increase above 40 mm Hg: additional metabolic
45 mmol/liter increase per 10 mm Hg PaCO2
increase above 40 mm Hg: chronic respiratory
>5 mmol/liter increase per 10 mm Hg PaCO2
increase above 40 mm Hg: additional metabolic

Anion gap: ([Na+][Cl][HCO3 ])

(reference value is analyzer-specific) Aa difference (Aa gradient) in mm Hg
Correct for albumin: for every 1 g/dl albumin At sea level (ambient air):
decrease, increase calculated anion gap 150PaO2 1.25PaCO2
by 2.5 mmol/liter

Normal anion gap: High anion gap Aa difference 10 mm Hg Aa difference >10 mm Hg

Calculate urinary anion gap (e.g., lactate, keto acids, (20 mm Hg in elderly) (>20 mm Hg in elderly)
([Na+ ]+[K+ ][Cl]) toxic alcohols) Hypoventilation without Hypoventilation with
If urinary pH>6.5, or urinary intrinsic lung disease intrinsic lung disease,
[Na+] <20 mmol/liter: ventilationperfusion
evaluate urinary osmolal gap mismatch, or both

Urinary anion gap Urinary anion gap DeltaDelta () Osmolal gap

negative positive: RTA Ketoacidosis: (measuredcalculated osmolality) >10 mOsm/kg
(e.g., diarrhea, sodium AG[HCO3 ] (e.g., toxic alcohols)
Type 1: serum [K+]
infusion, proximal RTA Lactic acidosis: Calculated serum osmolality:
decrease, urinary
[often hypophospha- Compute the value (2[Na+])+[glucose, in mg/dl]/18+(blood urea
pH >5.5
temia, hyperuricemia, of [0.6 AG] nitrogen, in mg/dl)/2.8
Type 4: serum [K+]
renal glucosuria]) [(HCO3 )] In standard units
increase, urinary
pH >5.5 in (mmol/liter)=(2[Na+])+[glucose]+[urea]
If the result is 5 to 5
mmol/liter for
either of the above:
only high anion-
gap metabolic
>5 mmol/liter:
high anion-gap
acidosis as well as
metabolic alka-
<5 mmol/liter:
high anion-gap
metabolic acidosis
as well as normal
anion-gap acidosis

1436 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

concentrations and largest variations in concen-

Figure 1. Assessment of Acidosis.
tration are used to calculate the excess of un-
Reference values for the alveolararterial (Aa) oxygen
tension difference are less than 10 mm Hg in young per measured anions in metabolic acidosis that con-
sons and less than 20 mm Hg in the elderly. AG denotes stitutes the anion gap, which is calculated as
delta anion gap, Paco2 partial pressure of arterial carbon [Na+][Cl][HCO3].
dioxide (mm Hg), Pao2 partial pressure of arterial oxy A true ion gap, however, does not exist in
gen (mm Hg), and RTA renal tubular acidosis. To con
vivo, because the sum of the positive and nega-
vert the values for Paco2, Pao2, and the alveolararterial
difference to kilopascals, multiply by 0.1333. tive ion charges in plasma must be equal. Wide
reference ranges of 3.0 to 12.0 mmol per liter up
to 8.5 to 15.0 mmol per liter in the anion gap
severe chronic respiratory alkalosis and acidosis. have been reported,33-36,43 owing to differences
It is generally accepted that compensatory pro- in laboratory methods.23,45 Consequently, clini-
cesses may normalize the pH only in chronic cians should know the reference range for their
respiratory alkalosis. In contrast with older data, own laboratory.
data from a more recent study13 indicate that the
pH in chronic respiratory acidosis may be nor- High-Anion-Gap Metabolic Acidosis
mal and, in individual cases, higher than gener- There are many causes of high anion-gap meta-
ally recognized (pH >7.40).13,17,24 Furthermore, bolic acidosis (Table 2). A useful mnemonic for
the usual compensatory changes in the Paco2 may the most common causes is GOLD MARRK (gly-
be limited in cases of severe hypoxemia. Instru cols [ethylene and propylene], 5-oxoproline [pyro
ments used for the measurement of blood gas glutamic acid], l-lactate, d-lactate, methanol,
and electrolytes may differ, affecting results.25-27 aspirin, renal failure, rhabdomyolysis, and keto-
Indeed, studies involving the use of modern ana- acidosis).46 The anion gap increases when the
lyzers show pH reference values (7.40 to 7.44)28-30 concentration of bicarbonate decreases relative
and secondary responses that differ from those to levels of sodium and chloride because of over-
published in textbooks.12,21,31 Although these production of acid (in ketoacidosis, lactic acido-
differences are small, a reappraisal of the pre- sis, and drug and alcohol-related intoxication),
diction equations may be needed. underexcretion of acid (in advanced renal failure),
cell lysis (in massive rhabdomyolysis), or other
circumstances (e.g., the use of penicillin-derived
E va luat ion of the Me ta bol ic
C omp onen t of a n Acid B a se antibiotics).
Disor der
Uses and Limitations of the Anion Gap
The third step in an evaluation is to consider the Lactic acidosis accounts for about half the cases
metabolic component of the acidbase disorder. of a high anion gap33-49 and is often due to shock
or tissue hypoxia.44,47 However, the anion gap is
Metabolic Acidosis a relatively insensitive reflection of lactic acidosis
Calculation of the anion gap is useful in the roughly half the patients with serum lactate
initial evaluation of metabolic acidosis.32-45 levels between 3.0 and 5.0 mmol per liter have an
The sum of the positive and negative ion charg- anion gap within the reference range.39,40 The
es in plasma are equal in vivo: [Na+]+[K+]+ anion gap, which has a sensitivity and specificity
[Ca2+]+[Mg2+]+[H+]+unmeasured cations=[Cl] below 80% in identifying elevated lactate levels,
+[HCO3]+[CO32]+[OH]+albumin+phosphate cannot replace a measurement of the serum lac-
+sulfate+lactate+unmeasured anions (e.g., in- tate level.39,40,47-50 Nevertheless, lactate levels are
organic anions).35-44 Routine measurement of all not routinely measured or always rapidly avail-
the ions in plasma is generally unnecessary. A able, and a high anion gap can alert the physician
more practical approach takes advantage of the that further evaluation is necessary.34,39,43 Un
fact that most plasma ions are normally present fortunately, a baseline value of the anion gap is
at relatively low concentrations and that varia- generally not available for an individual patient.
tions into the pathologic range are quantitatively In addition, the anion gap should always be ad-
small. The three ions with the highest plasma justed for the albumin concentration, because this

n engl j med 371;15nejm.orgoctober 9, 2014 1437

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. The Anion Gap in Relation to Common Medical Conditions with Metabolic Acidosis.*

High anion gap

Overproduction of acid
Ketoacidosis (diabetic ketoacidosis, alcoholic ketoacidosis, starvation)
Lactic acidosis
L-Lactic acidosis
Type A hypoxic (septic shock, mesenteric ischemia, hypoxemia, hypovolemic shock, carbon monoxide
poisoning, cyanide)
Type B nonhypoxic (thiamine deficiency, seizure, medications [nonnucleoside reverse-transcriptase in
hibitors, metformin, propofol, niacin, isoniazid, iron], intoxication [salicylate, ethylene glycol, propylene
glycol, methanol, toluene ingestion (early), paraldehyde])
D-Lactic acidosis in the short-bowel syndrome
Underexcretion of acid (advanced renal failure)
Impaired lactate clearance in liver failure (also type B acidosis)
Cell lysis (massive rhabdomyolysis)
Use of penicillin-derived antibiotics
Pyroglutamic acid (5-oxoproline)32
Normal anion gap
Loss of bicarbonate
Gastrointestinal conditions (diarrhea, ureteral diversions, biliary or pancreatic fistulas)
Renal conditions (type 2 [proximal] renal tubular acidosis, toluene ingestion [late in the process of toluene intoxication],
conditions associated with medications [ifosfamide, tenofovir, topiramate, carbonic anhydrase inhibitors
such as acetazolamide])3,41
Decreased renal acid excretion
Early uremic acidosis
Type 1 renal tubular acidosis (e.g., due to amphotericin, lithium, Sjgrens syndrome)3
Type 4 renal tubular acidosis (hypoaldosteronism or pseudohypoaldosteronism)
Other causes: fluid resuscitation with saline, hyperalimentation (lysine, histidine, or arginine hydrochloride), adminis
tration of hydrochloride, ammonium chloride, cholestyramine, hippuric acid, sulfuric acid

* An anion gap of more than 10 mmol per liter above the upper limit of the reference value is highly suggestive of organic
acidosis. A minor increase in the anion gap is less helpful in diagnosing metabolic acidosis.
Advanced renal failure is indicated by a glomerular filtration rate below 20 ml per minute.

weak acid may account for up to 75% of the track the resolution of ketosis9,15,23,33 and diag-
anion gap.36,39,40 Without correction for hypoal- nose a normal anion-gap acidosis if large vol-
buminemia, the estimated anion gap does not umes of isotonic saline are administered.50
reveal a clinically significant increase in anions A high anion gap with a normal lactate level
(>5 mmol per liter) in more than 50% of cases. in a patient with alcoholism may be an impor-
For every decrement of 1 g per deciliter in the tant clue for the diagnosis of alcoholic ketoaci-
serum albumin concentration, the calculated dosis. This diagnosis may be missed because the
anion gap should be increased by approximately test widely used to assess ketonuria (the nitro-
2.3 to 2.5 mmol per liter.9,36,39,40 Nevertheless, prusside test) reacts only with acetoacetate, not
the albumin-corrected anion gap is merely an with -hydroxybutyrate, the primary keto acid
approximation, since it does not account for ions seen in alcoholic ketoacidosis. The pH may also
such as magnesium, calcium, and phosphate ions. be misleadingly normal or elevated because of
The anion gap can help to establish the diag- concomitant metabolic alkalosis from vomiting
nosis of diabetic ketoacidosis. In patients with or respiratory alkalosis from liver disease, preg-
this condition, the anion gap can be used to nancy, high temperature, or sepsis.18,51-53

1438 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

The anion gap can also aid in the diagnosis tylsalicylic acid, D-lactic acid, and large quanti-
of d-lactic acidosis in patients with the short- ties of penicillin).9 Furthermore, the acidifica-
bowel syndrome, because the standard lactate tion of the urine requires adequate distal delivery
level (l-lactate) remains normal while the anion of sodium; thus, the usefulness of the urinary
gap increases.49 anion gap is questionable when the urinary so-
A low or negative anion gap is observed when dium level is less than 20 mmol per liter.12 In
hyperchloremia is caused by high levels of cat- such cases, the urinary osmolal gap is generally
ions, as seen in lithium toxicity, monoclonal IgG more reliable.
gammopathy, or disorders characterized by high The urinary osmolal gap determines the dif-
levels of calcium or magnesium. A negative an- ference between measured and calculated uri-
ion gap is caused by pseudohyperchloremia in nary osmolality. The urinary osmolality is calcu-
bromide or iodide intoxication.33,36,54 lated as follows:

Normal Anion-Gap Acidosis (2[Na+]+2[K+])+(blood urea nitrogen

Chloride plays a central role in intracellular and [in milligrams per deciliter]2.8)+
extracellular acidbase regulation.55 A normal (glucose [in milligrams per deciliter]18)
anion-gap acidosis occurs when the decrease in
bicarbonate ions corresponds with an increase in or (in millimoles per liter):
chloride ions to retain electroneutrality, which is
also called hyperchloremic metabolic acidosis. (2[Na+]+2[K+])+(blood urea nitrogen)+(glucose).
This type of acidosis occurs from gastrointesti-
nal loss of bicarbonate (e.g., because of diarrhea In patients without diabetes, the glucose con-
or ureteral diversion), from renal loss of bicar- centration is often omitted from this calcula-
bonate that may occur in defective urinary acidi- tion. A urinary osmolal gap below 40 mmol per
fication by the renal tubules (renal tubular acido- liter in normal anion-gap acidosis indicates im-
sis), or in early renal failure when acid excretion pairment in excretion of urinary ammonium.
is impaired.12,56,57 Hospital-acquired hyperchlo- The urinary osmolal gap usually reflects the
remic acidosis is usually caused by the infusion level of ammonium, except in the presence of
of large volumes of normal saline (0.9%).58-67 large quantities of a nondissociated acid, such as
Hyperchloremic acidosis should lead to increased -hydroxybutyric acid in ketoacidosis. The uri-
renal excretion of ammonium, and measurement nary osmolal gap, as compared with the urinary
of urinary ammonium can therefore be used to anion gap, has a better correlation with the
differentiate between renal and extrarenal causes urinary ammonium value.9,67
of normal anion-gap acidosis. However, since
urinary ammonium is seldom measured, the uri- Metabolic Alkalosis
nary anion gap and urinary osmolal gap are of- The normal kidney is highly efficient at excreting
ten used as surrogate measures of excretion of large amounts of bicarbonate and, accordingly,
urinary ammonium.9,67 the generation of metabolic alkalosis (Fig. 2) re-
The urinary anion gap ([Na+]+[K+][Cl]) is quires both an increase in alkali and impairment
usually negative in normal anion-gap acidosis, in renal excretion of bicarbonate.68-71 Loss of
but it will become positive when excretion of gastric fluid and the use of diuretics account for
urinary ammonium (NH4+) (as ammonium chlo- the majority of cases of metabolic alkalosis. By
ride [NH4Cl]) is impaired, as in renal failure, measuring chloride in urine, one can distinguish
distal renal tubular acidosis, or hypoaldosteron- between chloride-responsive and chloride-resis-
ism.9,67 A negative urinary anion gap occurs in tant metabolic alkalosis. If the effective circulat-
normal anion-gap acidosis because of diarrhea ing volume is reduced, the kidneys avidly reab-
and proximal renal tubular acidosis, in which sorb filtered sodium, bicarbonate, and chloride,
the distal acidification is intact.56 The urinary largely through activation of the reninangioten-
anion gap becomes unreliable when polyuria is sinaldosterone system, thus reducing the con-
present, when the urine pH exceeds 6.5,67 or centration of urinary chloride.
when urinary ammonium is excreted with an A (spot sample) urinary chloride concentra-
anion other than chloride (e.g., keto acids, ace- tion of less than 25 mmol per liter suggests

n engl j med 371;15nejm.orgoctober 9, 2014 1439

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

pH >7.42

Metabolic alkalosis Respiratory alkalosis

[HCO3] >26 mmol/liter PaCO2 <38 mm Hg

Secondary (respiratory) response Secondary (metabolic) response

Calculate expected PaCO2 : 0.7([HCO3] 24)+402 mm Hg Examine measured [HCO3]
If measured PaCO2 lower than calculated: additional If there is a change in [HCO3] of
respiratory alkalosis 2 mmol/liter decrease per 10 mm Hg PaCO2 decrease
If measured PaCO2 higher than calculated: additional below 40 mm Hg: acute respiratory alkalosis
respiratory acidosis <2 mmol/liter decrease per 10 mm Hg PaCO2 decrease
below 40 mm Hg: additional metabolic alkalosis
45 mmol/liter decrease per 10 mm Hg PaCO2 decrease
below 40 mm Hg: chronic respiratory alkalosis
>5 mmol/liter decrease per 10 mm Hg PaCO2 decrease
below 40 mm Hg: additional metabolic acidosis

Milk alkali syndrome Aa difference (Aa gradient) in mm Hg

Exogenous alkali or
(hypercalcemia in renal Yes At sea level (ambient air): 150PaO2 1.25PaCO2
severe hypercalemia?

Chloride-responsive Chloride-resistant Aa difference 10 mm Hg Aa difference >10 mm Hg

responsive to NaCl, KCl, or both Urinary Cl >40 mmol/liter (20 mm Hg in elderly) (>20 mm Hg in elderly)
(urinary chloride testing not If not due to continuing Hyperventilation without Hyperventilation with intrinsic
necessary in obvious gastric diuretic use or magnesium intrinsic lung disease lung disease, ventilation
fluid loss or with use of deficiency, analyze other (e.g., fever, pregnancy) perfusion mismatch, or
chloride-wasting diuretics) options both (e.g., pneumonia,
Urinary Cl <25 mmol/liter pulmonary embolism)
(e.g., vomiting)

Urinary K+ <20 mmol/day

Urinary K+ >30 mmol/day
(e.g., laxative abuse)

Low or normal blood pressure High blood pressure

The Gitelman syndrome Real or apparent mineralo-
(low urinary calcium) corticoid excess, often with
The Bartter syndrome hypokalemia
(high urinary calcium)

Figure 2. Assessment of Alkalosis.

Reference values for the alveolararterial (Aa) oxygen tension difference are less than 10 mm Hg in young persons and less than 20 mm Hg
in the elderly. Paco2 denotes partial pressure of arterial carbon dioxide (mm Hg), and Pao2 partial pressure of arterial oxygen (mm Hg). To
convert the values for Paco2, Pao2, and the alveolararterial difference to kilopascals, multiply by 0.1333.

chloride-responsive metabolic alkalosis. Adminis Metabolic alkalosis with a urinary chloride

tration of fluids with sodium chloride (usually concentration of more than 40 mmol per liter is
with potassium chloride) restores effective arte- mainly caused by inappropriate renal excretion
rial volume, replenishes potassium ions, or both of sodium chloride, often reflecting mineralo-
with correction of metabolic alkalosis. corticoid excess or severe hypokalemia (potassi-

1440 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

um concentration <2 mmol per liter). The ad- mmol per liter in a patient with ketoacidosis or if
ministration of sodium chloride does not correct 0.6 AG [HCO3]=05 mmol per liter in a pa-
this type of metabolic alkalosis, which, for that tient with lactic acidosis, simple anion-gap meta-
reason, is called chloride-resistant. Diuretic- bolic acidosis is present. A difference greater
induced metabolic alkalosis is an exception be- than 5 mmol per liter suggests a concomitant
cause the concentration of chloride in urine may metabolic alkalosis, and if the difference is less
increase initially, until the diuretic effect wanes, than 5 mmol per liter, a concomitant normal
after which the concentration will decrease to a anion-gap metabolic acidosis is diagnosed.
level below 25 mmol per liter.68-70 Other impor- In certain cases, normal values for concentra-
tant causes of chloride-resistant metabolic alka- tions of bicarbonate, Paco2, and pH do not en-
losis are the Bartter syndrome, the Gitelman sure the absence of an acidbase disturbance. An
syndrome, extreme hypercalcemia, and severe increase in the anion gap of more than 5 mmol
magnesium deficiency. In contrast to hyperaldo- per liter may then be the only clue to an underly-
steronism, these causes are not associated with ing mixed acidbase disorder.9,71 Because the
sodium retention (Fig. 2). individual anion gap and concentration of bicar-
bonate before the acidbase disorder are usually
not known, and the ranges of normal values for
E va luat ion for the Pr e sence
of Mi x ed Me ta bol ic Acid B a se the anion gap and the concentration of bicar-
Dis t ur b a nce s bonate are wide, the AG[HCO3] remains
an approximation.70,71
The fourth step in the evaluation of acidbase
disturbances is to consider the possibility of a C onsider at ion of the Serum
mixed metabolic acidbase disturbance. In high (or Pl a sm a) Osmol a l G a p
anion-gap metabolic acidosis, the magnitude of
the increase in the anion gap (the delta AG, or The fifth step in the evaluation of an acidbase
AG) is related to the decrease in the bicarbonate disturbance is to note the serum osmolal gap in
ions ([HCO3]). To diagnose a high anion-gap any patient with an unexplained high anion-gap
acidosis with concomitant metabolic alkalosis or acidosis, coma, or suspicion of ingestion of a
normal anion-gap acidosis, the so-called delta- (toxic) alcohol and in hospitalized patients with
delta (-) may be used.70,71 The delta gap is the an increased risk of iatrogenic propylene glycol
comparison between the increase (delta) in the intoxication (e.g., because of high-dose loraze-
anion gap above the upper reference value (e.g., pam administration in sedated patients in an in-
12 mmol per liter) and the change (delta) in the tensive care unit).72-76 Laboratory confirmation
concentration of bicarbonate ions from the low- of toxic alcohol ingestion is generally not rapidly
er reference value of bicarbonate ions (e.g., 24 available, and physicians must infer such a diag-
mmol per liter).9 In ketoacidosis, there is a 1:1 nosis by considering disorders that may necessi-
correlation between the increase in the anion tate immediate treatment. The osmolal gap is the
gap and the decrease in the concentration of bi- difference between measured serum osmolality
carbonate. In lactic acidosis, the decrease in the and calculated serum osmolality. The serum os-
concentration of bicarbonate is 0.6 times the in- molality is calculated as
crease in the anion gap (e.g., if the anion gap
increases by 10 mmol per liter, the concentration 2([Na+] [in millimoles per liter])+
of bicarbonate should decrease by approximately (glucose [in milligrams per deciliter])
6.0 mmol per liter). This difference is probably 18+(BUN [in milligrams per deciliter])2.8.
due to the lower renal clearance of lactate as
compared with keto-anions.71 Hydrogen buffer- If ethanol is involved, the result of this calcula-
ing in cells and bone takes time to reach comple- tion would be added to the amount of ethanol (in
tion. Accordingly, the ratio may be close to 1:1 milligrams per deciliter) divided by 3.7. An os-
with very acute lactic acidosis (e.g., shortly af- molal gap below 10 mOsm per kilogram is con-
ter seizures or in persons who exercise to the sidered to be normal, but the normal range in the
point of exhaustion).71 If AG[HCO3]=05 general population is large (10 to 10 mOsm per

n engl j med 371;15nejm.orgoctober 9, 2014 1441

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

evated in other clinical situations such as lactic

Table 3. Common Medical Conditions Characterized by Respiratory Acidosis
and Alkalosis.* acidosis, alcoholic ketoacidosis, and diabetic
Type of Acidosis Common Medical Conditions
Respiratory acidosis
E va luat ion of the R e spir at or y
Acute C omp onen t of a n Acid B a se
Normal alveolararterial O2 Depression of the central respiratory center Disor der
difference by cerebral disease (encephalitis or
trauma) or drugs (narcotics, barbitu The respiratory component of an acidbase dis-
rates, or benzodiazepines)
order can be determined by differentiating be-
High alveolararterial O2 Airway obstruction related to acute exacer
difference bations of asthma or pneumonia
tween acute and chronic respiratory acidbase
disorders with the use of clinical information
and calculations (Table 1) and the oxygenation
Normal alveolararterial O2 Neuromuscular disease (e.g., myasthenia
difference gravis, amyotrophic lateral sclerosis,
level. Hypoxemia, a major cause of lactic acido-
GuillainBarr syndrome, or muscular sis, may induce respiratory alkalosis. Evaluation
dystrophy), kyphoscoliosis of the partial pressure of arterial oxygen (Pao2)
High alveolararterial O2 Chronic obstructive pulmonary disease relative to ventilation, with the alveolararterial
difference oxygen-tension difference (hereafter called the
Respiratory alkalosis alveolararterial difference) taken into account,
Acute may distinguish pulmonary from extrapulmo-
Normal alveolararterial O2 Pain, anxiety, fever, stroke, meningitis, nary diseases. The difference in the partial oxy-
difference trauma, severe anemia, salicylate toxicity gen pressures between the alveolar and arterial
High alveolararterial O2 Pneumonia, pulmonary edema, pulmonary side of the alveolarcapillary membrane will be
difference embolism, aspiration, congestive heart high if the patient has associated lung disease
failure, sepsis (Table 3).77,78 The Pao2 in the alveolus is not
Chronic equal to that in the pulmonary circulation be-
Normal alveolararterial O2 Pregnancy, hyperthyroidism, hepatic failure cause physiological hypoventilation occurs in vari-
difference ous portions of the lung; therefore, the alveolar
High alveolararterial O2 Pulmonary embolism in pregnancy, liver arterial difference will be about 5 to 10 mm Hg
difference failure with aspiration pneumonia
in healthy young persons and 15 to 20 mm Hg in
* The alveolararterial O2 difference increases with age. For every decade a person healthy elderly persons. The alveolararterial dif-
has lived, the alveolararterial difference is expected to increase by 2 mm Hg; ference is calculated as
alternatively, one can compensate for age using the following formula: (alveo
lararterial O2 difference=[Age
4 +4]).
Minor defects may result in a normal alveolararterial O2 difference. Fio2(barometric pressurewater-vapor pressure)
Pao2(Paco2gas-exchange ratio).

The fraction of inspired oxygen (Fio2) is 0.21

liter).73,74 In ethylene glycol and methanol intoxi- in ambient air, the barometric pressure is 760
cation, the osmolal gap will be high shortly after mm Hg at sea level, and the water-vapor pres-
ingestion, but substantial amounts of acids will sure is 47 mm Hg at 37C. The gas-exchange
not be generated for several hours.72-76 Symptoms ratio, which is approximately 0.8 at steady-state
are considerably delayed by simultaneous ethanol levels, varies according to the relative utilization
ingestion because of competition for the enzyme of carbohydrate, protein, and fat. At sea level
alcohol dehydrogenase.74-76 and a body temperature of 37C, the alveolar
The use of the osmolal gap has some pit- arterial difference can be estimated77,78 as
falls. The wide normal range of the osmolal
gap in the general population renders the test Fio2(76047)Pao2(Paco20.8)
rather insensitive to small but potentially toxic
concentrations of ethylene glycol and metha- or
nol.74 In addition, the osmolal gap lacks speci-
ficity, given that it may also be moderately el- 150Pao21.25 Paco2.

1442 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

lactate gap in ethylene glycol intoxication79 (see

In ter pr e tat ion of Acid B a se
Disor der s in the Cl inic a l the Supplementary Appendix, available with the
C on te x t full text of this article at or an oxygen-
saturation gap in carbon monoxide, methemo-
The final step in evaluating acidbase disorders globinemia, or cyanide intoxication.80
is to determine the cause of the identified pro-
cesses. The evaluation of the laboratory data C onclusions
must fit with the clinical presentation of the pa-
tient (see box). The stepwise approach described Currently, there is no ideal method of assessing
here can be helpful in assessing acidbase disor- acidbase disturbances. The two other widely
ders, but one should always check for other in- practiced methods also have limitations. The
formation to support the diagnosis, such as a physicochemical (strong ion or Stewart 22,57,81)

Three Case Examples

Patient 1, a 22-year-old woman who had been rone-secreting adrenal adenoma. In a patient
injured in an accident, received 6 liters of iso with metabolic alkalosis and hypokalemia, the
tonic saline, after which the level of sodium was clinician should always rule out vomiting and
135 mmol per liter, potassium 3.8 mmol per liter, the use of diuretics before considering a renin
chloride 115 mmol per liter, and bicarbonate 18 aldosterone problem. Vomiting should lead to a
mmol per liter. The arterial blood pH was 7.28, chloride level below 10 mmol per liter in the
and the Paco2 was 39 mm Hg. The urinary sodium urine, whereas an aldosterone-secreting tumor
level was 65 mmol per liter, potassium 15 mmol should lead to a urinary chloride level greater
per liter, and chloride 110 mmol per liter. than 40 mmol per liter.63 The expected Paco2
This patient had a low anion-gap metabolic would be 40+0.7bicarbonate ions=40+0.7
acidosis (2 mmol per liter), but she also had (3224)=45.7 mm Hg, which is only margin
respiratory acidosis, because the expected Paco2 ally higher than the value in the patient.
is lower (1.5bicarbonate+82 mm Hg=
352 mm Hg). If these findings are the result of Patient 3, a previously healthy 22-year-old man,
chest-wall expansion problem such as rib frac developed large volumes of watery diarrhea
tures, the alveolararterial O2 difference could from infectious gastroenteritis. Laboratory tests
be normal, assuming no underlying lung pathol revealed a plasma sodium concentration of
ogy. The majority of patients with a normal anion- 140 mmol per liter, potassium 3.0 mmol per
gap metabolic acidosis have diarrhea and renal liter, chloride 86 mmol per liter, and bicarbon
tubular acidosis. The high chloride content of ate 38 mmol per liter. The arterial blood pH was
saline normalizes the anion gap because of the 7.60, and the Paco2 was 40 mm Hg.
concomitant decrease in the level of bicarbon The patient had metabolic alkalosis. The levels
ate. The low anion gap is probably the result of of pH and bicarbonate increased, but because
a low albumin level because of bleeding and the Paco2 did not increase, the patient also had
dilution. The urinary anion gap ([Na+]+[K+][Cl]) respiratory alkalosis, perhaps because of stress
was negative (30 mmol per liter) because of or fever. The albumin-uncorrected anion gap
the use of saline. It would have been positive in was 16 mmol per liter; a much higher value
a patient with renal tubular acidosis type 1 or 4. might be indicative of additional metabolic aci
dosis. Also, metabolic alkalosis, particularly that
Patient 2, a 50-year-old woman with a recent which is caused by vomiting or diuretic use, can
onset of hypertension, the level of sodium was be associated with an increment in the serum
150 mmol per liter, potassium 2.2 mmol per anion gap of approximately 4 to 6 mmol per liter
liter, chloride 103 mmol per liter, and bicarbon because of an increase of the albumin concen
ate 32 mmol per liter. The arterial blood pH was tration and its release of protons.33 The meta
7.50, and the Paco2 was 43 mm Hg. bolic alkalosis was the result of gastrointestinal
This patient was found to have an aldoste losses.

n engl j med 371;15nejm.orgoctober 9, 2014 1443

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
The n e w e ng l a n d j o u r na l of m e dic i n e

approach is complex and often requires cumber- chine.82 However, mixed acidbase disorders
some calculations that cannot be performed at will not be detected42 by that method without the
the bedside. Many clinicians think that it does use of elaborate base-excess partitioning.34,52,53
not provide a diagnostic or prognostic advan- Therefore, in our view, the physiological ap-
tage33,42,81 and that the large number of param- proach, considered here, remains the simplest,
eters used in calculations will increase the mag- most rigorous, and most serviceable approach to
nitude of variability and error.26 The standard the assessment of acidbase disorders.42
base-excess method accurately quantifies the No potential conflict of interest relevant to this article was
change in metabolic acidbase status in vivo and Disclosure forms provided by the authors are available with
is conveniently provided by the blood-gas ma- the full text of this article at

1. Narins RG, Emmett M. Simple and ods of evaluation of metabolic acid-base 1,400 meters. Am J Respir Crit Care Med
mixed acid-base disorders: a practical disorders. Crit Care Med 2007;35:1264- 1999;160:1525-31.
approach. Medicine (Baltimore) 1980;59: 70. 29. Funk GC, Doberer D, Kneidinger N,
161-87. 17. Jones NL. Respiratory acidosis sans Lindner G, Holzinger U, Schneeweiss B.
2. Morris CG, Low J. Metabolic acidosis acidemia. Can Respir J 2003;10:301-3. Acid-base disturbances in critically ill
in the critically ill: part 1. Classification 18. Krapf R, Beeler I, Hertner D, Hulter patients with cirrhosis. Liver Int 2007;27:
and pathophysiology. Anaesthesia 2008; HN. Chronic respiratory alkalosis the 901-9.
63:294-301. effect of sustained hyperventilation on 30. Zavorsky GS, Lands LC, Schneider W,
3. Rennke HG, Denker BM. Renal patho- renal regulation of acid-base equilibrium. Carli F. Comparison of fingertip to arte-
physiology, the essentials. 3rd ed. Phila- N Engl J Med 1991;324:1394-401. rial blood samples at rest and during exer-
delphia: Lippincott Williams & Wilkins, 19. Ayers P, Warrington L. Diagnosis and cise. Clin J Sport Med 2005;15:263-70.
2010. treatment of simple acid-base disorders. 31. Emmet M. Diagnosis of simple and
4. Guyton AC, Hall JE. Textbook of med- Nutr Clin Pract 2008;23:122-7. mixed disorders. In: Dubose T Jr, Hamm L,
ical physiology. 11th ed. Philadelphia: 20. Kellum JA, Murugan R. Anion gap eds. Acid base and electrolyte disorders:
Saunders Elsevier, 2006. and strong ion gap. In: Ronco C, Bellomo a companion to Brenner & Rectors The
5. Palmer BF. Approach to fluid and R, Kellum JA, eds. Critical care nephrolo- Kidney. Philadelphia: Saunders, 2002:41-53.
electrolyte disorders and acid-base prob- gy. Philadelphia: Elsevier, 2009:611-4. 32. Duewall JL, Fenves AZ, Richey DS,
lems. Prim Care 2008;35:195-213. 21. Berend K. Acid-base pathophysiology Tran LD, Emmett M. 5-Oxoproline (pyro-
6. Henderson LJ. The theory of neutrali- after 130 years: confusing, irrational and glutamic) acidosis associated with chron-
ty regulation in the animal organism. Am controversial. J Nephrol 2013;26:254-65. ic acetaminophen use. Proc (Bayl Univ
J Physiol 1908;21:427-48. 22. Fidkowski C, Helstrom J. Diagnosing Med Cent) 2010;23:19-20.
7. Hasselbalch KA. The calculation of metabolic acidosis in the critically ill: 33. Kraut JA, Madias NE. Serum anion
blood pH via the partition of carbon diox- bridging the anion gap, Stewart, and base gap: its uses and limitations in clinical
ide in plasma and oxygen binding of the excess methods. Can J Anaesth 2009;56: medicine. Clin J Am Soc Nephrol 2007;2:
blood as a function of plasma pH. Bio- 247-56. 162-74.
chem Z 1916;78:112-44. 23. Lolekha PH, Vanavanan S, Lolekha S. 34. Maciel AT, Park M. Differences in acid-
8. Berend K. Bedside rule secondary re- Update on value of the anion gap in clini- base behavior between intensive care unit
sponse in metabolic acid-base disorders is cal diagnosis and laboratory evaluation. survivors and nonsurvivors using both a
unreliable. J Crit Care 2013;28:1103. Clin Chim Acta 2001;307:33-6. physicochemical and a standard base ex-
9. Reddy P, Mooradian AD. Clinical util- 24. Ucgun I, Oztuna F, Dagli CE, Yildirim cess approach: a prospective, observa-
ity of anion gap in deciphering acid-base H, Bal C. Relationship of metabolic alka- tional study. J Crit Care 2009;24:477-83.
disorders. Int J Clin Pract 2009;63:1516-25. losis, azotemia and morbidity in patients 35. Feldman M, Soni N, Dickson B. Influ-
10. Ghosh AK. Diagnosing acid-base dis- with chronic obstructive pulmonary dis- ence of hypoalbuminemia or hyperalbu-
orders. J Assoc Physicians India 2006;54: ease and hypercapnia. Respiration 2008; minemia on the serum anion gap. J Lab
720-4. 76:270-4. Clin Med 2005;146:317-20.
11. Rowe KJ, Arrowsmith JE. Interpreta- 25. Otani N, Ohde S, Mochizuki T, Ishi- 36. Moe OW, Fuster D. Clinical acid-base
tion of measurements of arterial blood matsu S. Reliability of anion gap calculat- pathophysiology: disorders of plasma an-
gases. Surgery 2007;25:375-9. ed from data obtained using a blood gas ion gap. Best Pract Res Clin Endocrinol
12. Finkel KW, Dubose TF. Metabolic analyzer: is the probability of error predict- Metab 2003;17:559-74.
acidosis. In: Dubose T Jr, Hamm L, eds. able? Am J Emerg Med 2010;28:577-81. 37. Kellum JA. Making strong ion differ-
Acid base and electrolyte disorders: a com- 26. Nguyen BV, Vincent JL, Hamm JB, et al. ence the Euro for bedside acid-base
panion to Brenner & Rectors The Kidney. The reproducibility of Stewart parameters analysis. In: Vincent JL, ed. Yearbook of
Philadelphia: Saunders, 2002:55-66. for acid-base diagnosis using two central intensive care and emergency medicine.
13. Martinu T, Menzies D, Dial S. Re- laboratory analyzers. Anesth Analg 2009; Berlin: Springer-Verlag, 2005:675-85.
evaluation of acid-base prediction rules in 109:1517-23. 38. Hatherill M, Waggie Z, Purves L,
patients with chronic respiratory acidosis. 27. Sarrazin F, Tessler MJ, Kardash K, Reynolds L, Argent A. Correction of the
Can Respir J 2003;10:311-5. McNamara E, Holcroft C. Blood gas mea- anion gap for albumin in order to detect
14. Kellum JA. Determinants of plasma surements using the Bayer Rapid Point occult tissue anions in shock. Arch Dis
acid-base balance. Crit Care Clin 2005;21: 405: are we basing our decisions on ac- Child 2002;87:526-9.
329-46. curate data? J Clin Monit Comput 2007;21: 39. Chawla LS, Shih S, Davison D, Junker
15. Jones BJ, Twomey PJ. The anion gap 253-6. C, Seneff MG. Anion gap, anion gap cor-
revisited. Int J Clin Pract 2009;63:1409-12. 28. Crapo RO, Jensen RL, Hegewald M, rected for albumin, base deficit and un-
16. Dubin A, Menises MM, Masevicius FD, Tashkin DP. Arterial blood gas reference measured anions in critically ill patients:
et al. Comparison of three different meth- values for sea level and an altitude of implications on the assessment of meta-

1444 n engl j med 371;15nejm.orgoctober 9, 2014

The New England Journal of Medicine

Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.
Physiological Assessment of Acid Base Disturbances

bolic acidosis and the diagnosis of hyper- Kakepoto G, Fatmi Z, Ghani F. Anion gap Kellum JA. Diagnosis and therapy of met-
lactatemia. BMC Emerg Med 2008;8:18. among patients of multiple myeloma and abolic alkalosis. In: Ronco C, Bellomo R,
40. Berkman M, Ufberg J, Nathanson LA, normal individuals. Clin Biochem 2007; Kellum JA, eds. Critical care nephrology,
Shapiro NI. Anion gap as a screening tool 40:226-9. Philadelphia: Elsevier, 2009:621-4.
for elevated lactate in patients with an 55. Durward A, Skellett S, Mayer A, Taylor 70. Adrogu HJ. Mixed acid-base distur-
increased risk of developing sepsis in the D, Tibby SM, Murdoch IA. The value of bances. J Nephrol 2006;19:Suppl 9:S97-
Emergency Department. J Emerg Med the chloride: sodium ratio in differentiat- S103.
2009;36:391-4. ing the aetiology of metabolic acidosis. 71. Rastegar A. Use of the DeltaAG/Del-
41. Mirza N, Marson AG, Pirmohamed M. Intensive Care Med 2001;27:828-35. taHCO3- ratio in the diagnosis of mixed
Effect of topiramate on acid-base balance: 56. Katzir Z, Dinour D, Reznik-Wolf H, acid-base disorders. J Am Soc Nephrol
extent, mechanism and effects. Br J Clin Nissenkorn A, Holtzman E. Familial pure 2007;18:2429-31.
Pharmacol 2009;68:655-61. proximal renal tubular acidosis a clini- 72. Kraut JA, Kurtz I. Toxic alcohol inges-
42. Adrogu HJ, Gennari FJ, Galla JH, cal and genetic study. Nephrol Dial Trans- tions: clinical features, diagnosis, and
Madias NE. Assessing acid-base disorders. plant 2008;23:1211-5. management. Clin J Am Soc Nephrol 2008;
Kidney Int 2009;76:1239-47. 57. Corey HE, Vallo A, Rodrguez-Soriano J. 3:208-25.
43. Farwell WR, Taylor EN. Serum anion An analysis of renal tubular acidosis by 73. Whittington JE, Laulu SL, Hunsaker
gap, bicarbonate and biomarkers of in- the Stewart method. Pediatr Nephrol 2006; JJ, Roberts WL. The osmolal gap: what
flammation in healthy individuals in a 21:206-11. has changed? Clin Chem 2010;56:1353-5.
national survey. CMAJ 2010;182:137-41. 58. Bull SV, Douglas IS, Foster M, Albert 74. Lynd LD, Richardson KJ, Purssell RA,
44. Gunnerson KJ, Saul M, He S, Kellum RK. Mandatory protocol for treating adult et al. An evaluation of the osmole gap as a
JA. Lactate versus non-lactate metabolic patients with diabetic ketoacidosis de- screening test for toxic alcohol poisoning.
acidosis: a retrospective outcome evalua- creases intensive care unit and hospital BMC Emerg Med 2008;8:5.
tion of critically ill patients. Crit Care lengths of stay: results of a nonrandom- 75. Jammalamadaka D, Raissi S. Ethylene
2006;10:R22. ized trial. Crit Care Med 2007;35:41-6. glycol, methanol and isopropyl alcohol in-
45. Lolekha PH, Vanavanan S, Teerakarn- 59. Corey HE. The anion gap (AG): stud- toxication. Am J Med Sci 2010;339:276-81.
jana N, Chaichanajarernkul U. Reference ies in the nephrotic syndrome and dia- 76. Horinek EL, Kiser TH, Fish DN, Mac
ranges of electrolyte and anion gap on the betic ketoacidosis (DKA). J Lab Clin Med Laren R. Propylene glycol accumulation in
Beckman E4A, Beckman Synchron CX5, 2006;147:121-5. critically ill patients receiving continuous
Nova CRT, and Nova Stat Profile Ultra. 60. Mrozik LT, Yung M. Hyperchloraemic intravenous lorazepam infusions. Ann
Clin Chim Acta 2001;307:87-93. metabolic acidosis slows recovery in chil- Pharmacother 2009;43:1964-71.
46. Mehta AN, Emmett JB, Emmett M. dren with diabetic ketoacidosis: a retro- 77. Moammar MQ, Azam HM, Blamoun
GOLD MARK: an anion gap mnemonic spective audit. Aust Crit Care 2009;22: AI, et al. Alveolar-arterial oxygen gradi-
for the 21st century. Lancet 2008;372:892. 172-7. ent, pneumonia severity index and out-
47. Handy J. Lactate the bad boy of me- 61. Taylor D, Durward A, Tibby SM, et al. comes in patients hospitalized with com-
tabolism, or simply misunderstood? Curr The influence of hyperchloraemia on acid munity acquired pneumonia. Clin Exp
Anaesth Crit Care 2006;17:71-6. base interpretation in diabetic ketoacido- Pharmacol Physiol 2008;35:1032-7.
48. Leverve XM. Lactate in the intensive sis. Intensive Care Med 2006;32:295-301. 78. Jones JS, VanDeelen N, White L,
care unit: pyromaniac, sentinel or fire- 62. Story DA, Morimatsu H, Bellomo R. Dougherty J. Alveolararterial oxygen
man? Crit Care 2005;9:622-3. Hyperchloremic acidosis in the critically gradients in elderly patients with suspect-
49. Chang YM, Chiew YW, Yang CS. The ill: one of the strong-ion acidoses? Anesth ed pulmonary embolism. Ann Emerg Med
case mid R: a woman with severe meta- Analg 2006;103:144-8. 1993;22:1177-81.
bolic acidosis. Kidney Int 2010;77:261-2. 63. Berend K, van Hulsteijn LH, Gans RO. 79. Meng QH, Adeli K, Zello GA, Porter
50. Noritomi DT, Soriano FG, Kellum JA, Chloride: the queen of electrolytes? Eur J WH, Krahn J. Elevated lactate in ethylene
et al. Metabolic acidosis in patients with Intern Med 2012;23:203-11. glycol poisoning: true or false? Clin Chim
severe sepsis and septic shock: a longitu- 64. Handy JM, Soni N. Physiological ef- Acta 2010;411:601-4.
dinal quantitative study. Crit Care Med fects of hyperchloraemia and acidosis. Br 80. Mokhlesi B, Leiken JB, Murray P, Cor-
2009;37:2733-9. J Anaesth 2008;101:141-50. bridge TC. Adult toxicology in critical
51. Hassan H, Joh JH, Bacon BR, Bastani 65. Gennari FJ, Weise WJ. Acid-base dis- care: part I: general approach to the in-
B. Evaluation of serum anion gap in pa- turbances in gastrointestinal disease. Clin toxicated patient. Chest 2003;123:577-92.
tients with liver cirrhosis of diverse etiol- J Am Soc Nephrol 2008;3:1861-8. 81. Kurtz I, Kraut J, Ornekian V, Nguyen
ogies. Mt Sinai J Med 2004;71:281-4. 66. Gattinoni L, Carlesso E, Maiocchi G, MK. Acid-base analysis: a critique of the
52. Funk GC, Doberer D, Osterreicher C, Polli F, Cadringher P. Dilutional acidosis: Stewart and bicarbonate-centered ap-
Peck-Radosavljevic M, Schmid M, Schnee- where do the protons come from? Inten- proaches. Am J Physiol Renal Physiol
weiss B. Equilibrium of acidifying and sive Care Med 2009;35:2033-43. 2008;294:F1009-F1031.
alkalinizing metabolic acid-base disorders 67. Rodrguez Soriano J. Renal tubular 82. Tuhay G, Pein MC, Masevicius FD,
in cirrhosis. Liver Int 2005;25:505-12. acidosis: the clinical entity. J Am Soc Kutscherauer DO, Dubin A. Severe hyper-
53. Ahya SN, Jos Soler M, Levitsky J, Nephrol 2002;13:2160-70. lactatemia with normal base excess:
Batlle D. Acid-base and potassium disor- 68. Laski ME, Sabatini S. Metabolic alka- a quantitative analysis using conventional
ders in liver disease. Semin Nephrol 2006; losis, bedside and bench. Semin Nephrol and Stewart approaches. Crit Care 2008;
26:466-70. 2006;26:404-21. 12:R66.
54. Mansoor S, Siddiqui I, Adil S, Nabi 69. Heffner AC, Murugan R, Madden N, Copyright 2014 Massachusetts Medical Society.

images in clinical medicine

The Journal welcomes consideration of new submissions for Images in Clinical
Medicine. Instructions for authors and procedures for submissions can be found
on the Journals website at At the discretion of the editor, images that
are accepted for publication may appear in the print version of the Journal,
the electronic version, or both.

n engl j med 371;15nejm.orgoctober 9, 2014 1445

The New England Journal of Medicine
Downloaded from at USP on October 8, 2014. For personal use only. No other uses without permission.
Copyright 2014 Massachusetts Medical Society. All rights reserved.