Prim Care Clin Office Pract

35 (2008) 195213

Approach to Fluid and Electrolyte

Disorders and Acid-Base Problems
Bi F. Palmer, MD
Department of Internal Medicine, Division of Nephrology, University of Texas Southwestern
Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390, USA

Employing a systematic approach to the interpretation of serum chemis-

tries is the most eective way to ensure abnormalities are detected and cor-
rectly interpreted. This article reviews a series of steps that can be used in
both the outpatient and inpatient settings. These steps will help to ensure
the clinician not only identies overt abnormalities but also subtle distur-
bances that may lay hidden in a routine set of serum chemistry values.

Steps to interpreting a set of chemistry values

 Step 1: First check serum sodium [Na] and work up tonicity disorders if
abnormal
 Step 2: Check serum chloride [Cl]
 Step 3: Calculate anion gap
 Step 4: If anion gap [, determine whether measured serum bicarbonate
[HCO3] is equal to predicted [HCO3]
 Step 5: Analyze arterial blood gas
 Step 6: Check potassium

First step: examine the serum sodium concentration

Hyponatremia
Is the hyponatremia representative of a hypoosmolar state?
There are two general causes of hyponatremia in which it is not associ-
ated with a hypoosmolar state. The rst of these is pseudohyponatremia,
which involves an abnormal measurement of the serum sodium (Na). This

E-mail address: [email protected]

0095-4543/08/$ - see front matter 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.pop.2008.01.004 primarycare.theclinics.com
196 PALMER

occurs in patients with hyperglobulinemia or hypertriglyceridemia, in whom

plasma water relative to plasma solids is decreased in blood, leading to less
sodium in a given volume of blood. In general, hyperglobulinemia sucient
to cause pseudohyponatremia is rare and occurs only in Waldenstroms
macroglobulinemia. Triglycerides must be in the thousands to cause this
condition and are most commonly seen in diabetics. In general, this problem
is becoming less prevalent, as many laboratories are using sodium electrodes
without diluting the blood such that the plasma sodium measurement be-
comes independent of plasma water and solid contents.
The other cause of hyponatremia in the absence of a hypoosmolar state
involves true hyponatremia, but with elevations in the concentration of an-
other osmole. Clinical examples include hyperglycemia as seen in uncon-
trolled diabetes or, rarely, hypertonic infusion of mannitol used in the
treatment of cerebral edema. The increases in plasma glucose raise serum
osmolality, which pulls water out of cells and dilutes the serum Na. For
every 100 mg/dL rise in glucose or mannitol, the serum Na will quickly
fall by 1.6 mEq/L. The increased tonicity will also stimulate thirst and argi-
nine vasopressin (AVP) secretion, both of which contribute to further water
retention. As the plasma osmolality returns toward normal, the decline in
serum Na will be 2.8 mEq/L for every 100 mg/dL rise in glucose. The net
result is a normal plasma osmolality but a low serum sodium.

Is the kidneys ability to dilute the urine intact?

The presence of hypotonic hyponatremia implies that water intake ex-
ceeds the ability of the kidney to excrete water. In unusual circumstances,
this can occur when the kidneys ability to excrete free water is intact. How-
ever, because a normal kidney can excrete 20 L to 30 L of water per day, the
presence of hyponatremia with normal renal water excretion implies that the
patient is drinking more than 20 L to 30 L of water per day. This condition
is referred to as primary polydipsia. These patients should have a urine
osmolality less than 100 mOsm/L. While primary polydipsia is a common
condition that leads to polyuria and polydipsia, it is uncommon as a sole
cause of hyponatremia.
In the absence of primary polydipsia, hyponatremia is associated with
decreased renal water excretion and urine that is inappropriately concen-
trated. It is important to note that in the presence of hyponatremia, urine
should be maximally dilute and a urine osmolality higher than this (greater
than 100 mOsm/L) is inappropriate. An inappropriately concentrated urine
implies a defect in renal water excretion (Fig. 1).
Excretion of water by the kidney is dependent on three factors. First,
there must be adequate delivery of ltrate to the tip of the loop of Henle.
Second, solute absorption in the ascending limb and the distal nephron
must be normal so that the tubular uid will be diluted. Lastly, AVP levels
must be low in the plasma. Of these three requirements for water excretion,
the one which is probably most important in the genesis of hyponatremia is
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 197

Serum Osmolality

Decreased Normal
Pseudohyponatremia
Urine osmolality

Appropriately low Inappropriately high (>200 mOsm/L)

<100 mOsm/L
EABV
Primary polydipsia

Low Normal
SIADH
ECF volume
Glucocorticoid deficiencyHy
pothyroidism
Low High Drugs

Diarrhea Edematous states

Diuretics
Mineralocorticoid deficiency

Fig. 1. Approach to the hyponatremic patient. Abbreviations: EABV, eective arterial blood
volume; ECF, extracellular uid; SIADH, syndrome of inappropriate secretion of antidiuretic
hormone.

the failure to maximally suppress AVP levels. In many conditions, decreased

delivery of ltrate to the tip of the loop of Henle also contributes. Defective
solute absorption in the ascending limb and distal nephron is probably only
relevant to hyponatremia seen with chloriuretic diuretics.

Assess the patients volume status

In patients with hypotonic hyponatremia with an inappropriately con-
centrated urine, one needs to dene whether eective arterial volume is de-
creased. Most of the causes of hyponatremia result from a decrease in
eective arterial volume, which causes baroreceptor stimulation of AVP se-
cretion and leads to decreased distal delivery of ltrate to the tip of the loop
of Henle. If eective arterial volume is low, extracellular uid volume can be
low in the volume-depleted patient (hypovolemic hyponatremia) or can be
high in the edematous patient (hypervolemic hyponatremia). If eective ar-
terial volume is normal, one is dealing with the euvolemic causes of hypona-
tremia (isovolemic hyponatremia).
The clinical determination of eective arterial volume is usually straight-
forward. On physical examination, the best index of eective arterial volume
is the pulse and blood pressure. Urinary electrolytes are also extremely use-
ful in the assessment of eective arterial volume. Patients with a low eec-
tive arterial volume will tend to have low urinary sodium, low urinary
chloride, and low fractional excretions of sodium and chloride in the urine.
Patients with euvolemic hyponatremia, however, will be in balance and will
excrete sodium and chloride at rates that reect dietary intake of sodium
198 PALMER

and chloride. Thus, generally they have urinary sodium and chlorides
greater than 20 mEq/L and fractional excretions of these electrolytes more
than 1%.
Plasma composition can also be used to assess eective arterial volume.
The blood urea nitrogen (BUN) is particularly sensitive to eective arterial
volume. In patients with normal serum creatinines, a high BUN suggests
a low eective arterial volume and a low BUN suggests a high eective ar-
terial volume. The plasma uric acid can also be used as a sensitive index of
eective arterial volume. In comparing patients with SIADH and other
causes of hyponatremia, patients with low eective arterial volume tend
to have an elevated serum uric acid. In patients with SIADH, serum urate
is not only not elevated, but is actually depressed. This is because of the
fact that these patients are volume expanded, although it is clinically dicult
to detect the degree of volume expansion.

Hypernatremia
Why is the patient not drinking?
The initial approach to any patient with hypernatremia is to determine
why there has been inadequate intake of water. Hypernatremia is rare in
conscious patients who have free access to water because of the extreme sen-
sitivity of the thirst mechanism. Inadequate water intake may result from
conditions in which there is a specic lesion of the thirst center. More com-
monly, there is an alteration in the level of consciousness so that patients
become unaware of thirst or cannot adequately communicate the need for
water. In some instances, thirst is adequately sensed but water is unavailable
or there is restricted access to water because patients are restrained.
Reduced sensation of thirst also occurs in otherwise normal individuals as
a feature of increasing age, rendering the elderly particularly susceptible
to the development of hypernatremia.

Has there been accelerated water loss or increased sodium gain?

The next step in the evaluation of hypernatremia is to search for the pres-
ence of accelerated water loss or increased sodium gain, both of which will
increase the likelihood of a patient developing hypernatremia. This best can
be accomplished by clinically assessing the patients extracellular uid
status.

Hypovolemic hypernatremia. Hypernatremia in the setting of hypovolemia

results from uid losses in which the sodium concentration is less than
the plasma concentration. Extrarenal losses of salt and water from the
gastrointestinal tract or from profuse sweating or renal losses because
of osmotic diuresis are the major causes. Diuretics can also predispose
to the development of hypernatremia, because these agents are associated
with renal salt and water loss but water loss to a greater extent. It should
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 199

be emphasized, however, that hypernatremia will only develop if there is

an associated impairment in water intake. Urine sodium concentration
should be low with extrarenal uid losses, while the concentration is typ-
ically high with an osmotic diuresis or during the administration of
a diuretic.

Hypervolemic hypernatremia. Hypernatremia in the setting of hypervolemia

can be caused by iatrogenic administration of hypertonic sodium chloride
(NaCl) or hypertonic sodium bicarbonate or mineralocorticoid excess. Ad-
ministration of hypertonic uids is usually evident from the clinical setting
and is associated with a high urine sodium concentration. Mineralocorticoid
excess is suggested by the presence of hypertension and hypokalemic meta-
bolic alkalosis. Urine sodium concentration will vary according to dietary
intake.

Isovolemic hypernatremia. Pure water loss, whether from mucocutaneous

routes or from the kidneys, causes isovolemic hypernatremia. Because two
thirds of pure water loss is sustained from within cells, patients will not be-
come clinically volume-depleted unless the water decit becomes substantial.
Insensible losses from the respiratory tract or skin results in a concentrated
urine. Inappropriate water loss by the kidney, whether from central or neph-
rogenic diabetes insipidus, results in a dilute urine. Although renal water
loss can lead to hypernatremia in patients with impaired thirst or access
to water, most patients with diabetes insipidus have neither of these defects,
and typically a patient presents with polyuria and polydipsia and a normal
serum sodium concentration. As a result, the initial clue to the presence of
diabetes insipidus usually comes not from the detection of hypernatremia
but rather during the evaluation of the polyuric patient.

Step 2: Examine the serum chloride concentration

One should always examine the serum chloride concentration with re-
spect to the serum sodium concentration. As the serum sodium concentra-
tion either increases or decreases because of disorders in tonicity, the
serum chloride concentration will change in parallel and to the same extent.
Whenever the serum chloride concentration moves in a direction opposite or
changes disproportionately to the change in serum sodium concentration,
an acid-base disorder is suggested.

The chloride concentration is decreased with respect to the serum sodium

concentration
Whenever the chloride concentration is decreased with respect to the se-
rum sodium concentration, one should suspect the presence of either an
acute or chronic respiratory acidosis or a metabolic alkalosis. The fall in
200 PALMER

chloride concentration will be accompanied by an increase in the serum

bicarbonate concentration in these conditions.

Acute and chronic respiratory acidosis

Acute hypercapnia is associated with several eects that lead to an imme-
diate small decrease in the chloride concentration and rise in the serum
bicarbonate (HCO3) concentration. First, the decrease in pH that accom-
panies acute respiratory acidosis increases hydrogen (H) binding to albu-
min. This eect will cause a fall in the anion gap and a slight rise in the
serum HCO3 concentration. Second, a small amount of H enters parenchy-
mal cells in exchange for Na and potassium (K), also contributing to
a small increase in extracellular HCO3 concentration. Third, the high carbon
dioxide (CO2) tension is immediately transmitted into red blood cells, where
in the presence of carbonic anhydrase (H2CO3) is generated. This acid dis-
associates and the H binds to hemoglobin, leaving HCO3 in the cytoplasm
of the red cell. Hemoglobin can bind a considerable amount of H because
of the rich histidine content of the molecule. As the HCO3 concentration
rises, it exits from the cell in exchange for plasma chloride (Cl), a process
termed the red cell HCO3-Cl shift. The net eect of these changes is that in
acute respiratory acidosis, the plasma HCO3 concentration increases by
1 mEq/L for each 10 mm Hg elevation in partial pressure of carbon dioxide
(PaCO2) (Box 1).

Box 1. Compensation in acid-base disorders

 Acute respiratory acidosis
For every 10 mm Hg, rise in PaCO2 the HCO3 increases by
1 mEq/L
 Chronic respiratory acidosis
For every 10 mm Hg, rise in PaCO2 the HCO3 increases by
3.5 mEq/L
 Acute respiratory alkalosis
For every 10 mm Hg, fall in PaCO2 the HCO3 decreases
by 2 mEq/L
 Chronic respiratory alkalosis
For every 10 mm Hg, decrease in PaCO2 the HCO3 decreases
by 5 mEq/L
 Metabolic acidosis
1.2 mm Hg decrease in PaCO2 for each 1 mEq/L fall in HCO3
PaCO2 = HCO3 + 15
PaCO2 = last two digits of pH
 Metabolic alkalosis
PaCO2 increases by 0.7 for each mEq/L HCO3
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 201

In chronic respiratory acidosis, the increase in plasma HCO3 concentra-

tion is higher because of compensatory renal mechanisms. The chronic
elevation in CO2 leads to intracellular acidosis in proximal tubular cells,
increasing H secretion and resulting in accelerated HCO3 reabsorption.
The retention of NaHCO3 leads to slight expansion of the extracellular uid
compartment and causes increased renal excretion of NaCl, so as to return
volume back to normal. The net eect is an increase in serum HCO3 and
decreased Cl concentration. In chronic respiratory acidosis there is
a 3.5 mEq/L increase in HCO3 for each 10 mm Hg elevation in PaCO2.
Higher or lower plasma HCO3 concentrations suggest the presence of mixed
respiratory and metabolic acid-base disorders.

The chloride concentration is increased with respect to the serum sodium

concentration
Whenever the chloride concentration is increased with respect to the
serum sodium concentration, one should suspect the presence of either
an acute or chronic respiratory alkalosis or a normal gap hyperchlore-
mic metabolic acidosis. The rise in chloride concentration will be accom-
panied by a fall in the serum bicarbonate concentration in these
conditions.

Acute and chronic respiratory alkalosis

An acute fall in PaCO2 causes plasma and red blood cell CO2 tensions to
fall. In response, albumin and other non-HCO3 buers release H to de-
crease the plasma HCO3 concentration. Within red cells, the fall in
PaCO2 causes hemoglobin to release H and red blood cell HCO3 concen-
tration also falls. Plasma HCO3 will enter the red cell in exchange for Cl.
This HCO3-Cl shift accounts for the small initial compensatory response in
acute respiratory alkalosis in which the HCO3 concentration falls by 2 mEq/
L for every 10-mm Hg decrease in PaCO2.
In chronic respiratory alkalosis the renal HCO3 reabsorptive capacity
decreases and there is a transient NaHCO3 diuresis. This process takes
2 to 3 days to become fully manifest. Once a new steady state is achieved,
the HCO3 concentration will have decreased by 5 mEq/L for each 10-mm
Hg fall in PaCO2. A higher or lower value for the plasma HCO3 concentra-
tion suggests the presence of an additional metabolic disorder.
While the urinary loss of NaHCO3 does not increase total body Cl, it
causes a decrease in total body Na and total body HCO3, leading to con-
traction of extracellular uid volume and a hypotonic state. The response to
these changes is increased renal reabsorption of NaCl and suppression of
antidiuretic hormone secretion. The increase in renal water excretion returns
the Na concentration to normal and leads to a high Cl concentration. The
nal result is a decrease in the HCO3 concentration and an increase in the
Cl concentration.
202 PALMER

Another characteristic nding in chronic respiratory alkalosis is an in-

crease of 3 mEq/L to 5 mEq/L in the serum anion gap. The increase in
gap is the result of an increase in the number of anionic sites on albumin
as protons disassociate in response to the increased blood pH. The circulat-
ing levels of lactate are also increased in this setting. Lactate production is
increased because of a stimulatory eect of high pH on phosphofructoki-
nase, the rate-limiting step in the glycolytic pathway.

Step 3: Calculate the anion gap

Calculation of the anion gap should be a routine part of the examination
of every set of electrolytes, no matter how normal the individual values may
appear. The anion gap is equal to the dierence between the plasma concen-
trations of the major cation (Na) and the major measured anions (Cl
HCO3).




Anion gap Na  Cl  HCO
3

The normal value of the anion gap is approximately 12 plus or minus 2.

Most of the unmeasured anions consists of albumin, and therefore the normal
anion gap changes in the setting of hypoalbuminemia (normal anion gap is
approximately three times the serum albumin in g/dL). Because the total
number of cations must equal the total number of anions, a fall in the serum
HCO3 concentration must be oset by a rise in the concentration of other
anions. If the anion accompanying excess H is Cl, then the fall in the serum
HCO3 concentration is matched by an equal rise in the serum Cl concen-
tration. The acidosis is classied as a normal gap or hyperchloremic meta-
bolic acidosis. By contrast, if excess H is accompanied by an anion other
than Cl, then the fall in HCO3 is balanced by a rise in the concentration
of the unmeasured anion. The Cl concentration remains the same. In this
setting, the acidosis is said to be a high anion gap metabolic acidosis. An in-
crease in the anion gap can occur whenever there is an increase in unmeasured
anions (increased valency of albumin, hyperphosphatemia) or a decrease in
the unmeasured cations (hypocalcemia, hypomagnesemia). However, the
most common cause of an increase in the anion gap is the generation of a met-
abolic acidosis by addition of a non-Cl acid.

Classication of metabolic acidosis

Hyperchloremic normal gap acidosis
Renal origin
Proximal renal tubular acidosis (Type II RTA)
Hypokalemic distal renal tubular acidosis (Type 1 RTA)
Hyperkalemic distal renal tubular acidosis (Type IV RTA)
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 203

Renal tubular acidosis of renal insuciency (glomeruler ltration

rate or GFR usually O 15 mL/min20 mL/min)
Extrarenal origin
Diarrhea
Gastrointestinal ureteral connections
External loss of pancreatic or biliary secretions
Anion gap metabolic acidosis
Renal origin
Uremic acidosis (GFR usually ! 15 mL/min20 mL/min)
Extrarenal origin
Lactic acidosis
Diabetic ketoacidosis
Starvation ketoacidosis
Alcoholic ketoacidosis
Poisoning (ethylene glycol, methanol, salicylate)
Pyroglutamic acidosis
The patient with metabolic acidosis and an increased anion gap should be
worked up aggressively because of the numerous life-threatening conditions
that can cause this disorder. Measurement of the BUN or creatinine will dis-
close the presence of an acidosis of renal insuciency. Blood lactate levels
and ketoacids can be measured in the plasma and urine using the nitroprus-
side test. The presence of a signicant osmolar gap, calcium oxalate crystals
in the urine, abnormal vision, coexistent respiratory alkalosis, or tinnitus
may suggest an overdose. The clinical setting is often helpful. The patient
may be in shock, the patient may be a known diabetic, or the patient may
have a known history of suicide attempts.

Step 4: Is the measured HCO3 concentration equal to the predicted

HCO3?
In the setting of an increased anion gap, one needs to determine whether
the measured serum HCO3 is equal to the predicted serum HCO3. In gen-
eral, the serum HCO3 concentration will fall by an amount equal to the in-
crease in the anion gap. Sometimes an increased anion gap may be the only
clue to a metabolic acidosis. For instance, consider a patient with a measured
plasma HCO3 concentration of 22 mEq/L in the setting of an anion gap of
22. In this instance the anion gap has increased by 10, assuming the normal
anion gap is 12. One would predict the serum HCO3 should be approxi-
mately 14 mEq/L, assuming a normal value of 24 mEq/L (24  10 14).
To explain the near normal serum HCO3 in this setting, one concludes
a metabolic alkalosis is also present. Thus, the patient has two acid-base
processes: an anion gap metabolic acidosis and a metabolic alkalosis. If
one did not calculate the anion gap one might mistakenly call this patients
acid-base status normal.
204 PALMER

Consider a patient with a measured serum HCO3 concentration of

6 mEq/L in the setting of an anion gap of 22. Once again, the predicted
HCO3 concentration based on an increase in anion gap of 10 would be
14 mEq/L. To explain the lower measured value, one concludes there is
also a normal gap hyperchloremic metabolic acidosis. Thus, this patient
has two acid-base disturbances: an anion gap metabolic acidosis and a nor-
mal gap hyperchloremic metabolic acidosis.

Step 5: Interpret the arterial blood gas

Before one can approach a patient with an acid-base disorder, it is impor-
tant to distinguish the suxes -emia and -osis. The sux, -emia re-
fers to the concentration in the blood. Thus, acidemia means excess acid in
the blood (low blood pH), and alkalemia means excess alkali in the blood
(high blood pH). The sux -osis refers to a process. Thus, acidosis refers
to a process that adds acid to the blood, while alkalosis refers to a process
that adds alkali to the blood. A patient could have a metabolic acidosis and
a respiratory alkalosis, and these could result in a normal blood pH, a high
blood pH (alkalemia) or a low blood pH (acidemia).

Determine primary disorders

Once a patient is found to have an acid-base abnormality, the next step is
to determine which abnormalities are primary disorders and which involve
normal compensations. For instance, increased net acid production leads to
a drop in serum HCO3 concentration in a primary metabolic acidosis. The
drop in PaCO2, which occurs in response to this, is a normal respiratory
compensation and is not a primary disorder. A primary metabolic acid-
base disturbance refers to a primary increase or decrease in HCO3 concen-
tration because of addition or loss of nonvolatile acids or alkali from the
extracellular uid. A primary respiratory acid-base disturbance refers to
a primary change in PaCO2, reecting a primary increase or decrease in al-
veolar ventilation relative to existing CO2 production. In both of these cases,
the word primary implies that the change is not because of a change in
blood pH, and is thus not a secondary compensation. To determine which
process is primary and which is secondary, one generally follows a three-
step process, addressing three questions.

Step 1: What is blood pH?

Compensations generally do not return blood pH back to 7.4, such that
a pH of 7.4 in the presence of an abnormal HCO3 concentration and
PaCO2 generally implies two simultaneous primary acid-base disorders:
a metabolic and a respiratory one. If blood pH is low, then at least one pri-
mary abnormality is an acidosis and if blood pH is high at least one primary
abnormality is an alkalosis.
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 205

Step 2: Has PaCO2 or HCO3 concentration moved in the direction

of the pH?
Taking the example of an acidemia, one would then look at the HCO3
concentration and the PaCO2 to see which has moved in the acid direction.
A drop in the HCO3 concentration or an increase in the PaCO2 would
both be a change in the acid direction. If either has moved in the acid direc-
tion, it must represent a primary disorder. If they have both moved in the acid
direction, a mixed metabolic respiratory acidosis is present. If only one of
them has moved in the acid direction, this is a cause of a primary acidosis.

Step 3: Are there one or more abnormalities?

The last question is whether the response of the other variable is appro-
priate (in which case it would simply be a normal compensation) or whether
the response of the other variable is outside the acceptable limits implying,
once again, two abnormalities or a mixed acid-base disorder. To answer the
last question requires knowledge of the appropriate compensations in re-
sponse to primary metabolic or respiratory acid-base changes. For this,
one may merely refer to an acid-base map. The map plots the combina-
tions of HCO3 concentration, PaCO2, and pH that are seen in the various
disorders, with 95% condence limits. If one does not wish to carry such
a map around with them, a few general rules can be remembered. As stated
above, as a general rule pH should be returned toward and not to or past
normal. In addition, there are some numbers that the physician can carry
with him or memorize, which describe an average compensation. These
are listed in Box 1.
Using these equations, one can calculate the appropriate compensatory
response. If the response is far dierent from this, a second primary process
(ie, a mixed acid-base disorder) is implied. For instance, if a patient has a pri-
mary metabolic acidosis with a decrease in his plasma HCO3 concentration
to 14 mEq/L (a decrease of 10 mEq/L), the appropriate pulmonary response
for a respiratory compensation is a decrease in the PaCO2 to approximately
29 mm Hg. If the PaCO2 only falls to 35, this would represent an inadequate
respiratory response and the patient would be said to have a combined met-
abolic and respiratory acidosis. Conversely, if the PaCO2 is 18 (a greater
than normal response), then the patient would be said to have a combined
metabolic acidosis and respiratory alkalosis.

Interpret the acid-base disorders in the context of the clinical

presentation
The nal step in evaluating acid-base disorders is to determine the etiol-
ogy of the identied processes based on the clinical presentation of the
patient. The evaluation of the laboratory data must t with the clinical pre-
sentation of the patient. Consider the following set of laboratory values: pH
7.47, PaCO2 20, HCO3 14 mEq/L. These values are typical of a normal
206 PALMER

pregnancy at 34 weeks gestation. The high progesterone levels during preg-

nancy exert a stimulatory eect on the respiratory center, causing the devel-
opment of a chronic respiratory alkalosis. The decrease in HCO3 of 10 mEq/
L (from 24 to 14) is the expected compensatory change for a chronic de-
crease in PaCO2 of 20 mm Hg.
However, the interpretation of these same values would dier if they were
obtained from a 23-year-old man who 6 hours earlier ingested a large quan-
tity of aspirin. First, the alkaline pH and low PaCO2 are indicative of a re-
spiratory alkalosis. However, based on the clinical history, the process is
acute. In the setting of acute respiratory alkalosis the expected compensa-
tory change in serum HCO3 is a decline to 20 mEq/L (2 mEq/L for each
10-mm Hg decline in PaCO2). The measured value of 14 mEq/L suggests
an additional process is present. In this case one concludes there is a meta-
bolic acidosis. Indeed, a combination of respiratory alkalosis and anion gap
metabolic acidosis is typical of adults who present with aspirin intoxication.
Consider a second set of laboratory data: pH 7.53, PaCO2 51, HCO3 40.
The pH is alkaline and the serum HCO3 has moved in the alkaline direction,
making metabolic alkalosis the primary disturbance. The increase in PaCO2
to 53 mm Hg is an appropriate compensatory response for a simple meta-
bolic alkalosis (0.7  the increase in HCO3 above the normal value of
24 mEq/L). These values are consistent with a 56-year-old man who presents
with nausea and vomiting because of a gastric outlet obstruction.
However, one would need to alter the interpretation of the values if ob-
tained from a 63-year-old woman with severe chronic obstructive lung disease
and cor pulmonale recently started on loop diuretics. In this setting, the ex-
pected acid base disturbance is a chronic respiratory acidosis. If the PaCO2
is chronically 51 mm Hg, the expected renal compensation is for the serum
HCO3 concentration to be increased to approximately 28 mm Hg. In addition,
the arterial pH should be less than 7.4. Given the alkaline pH, the higher than
expected serum HCO3 concentration of 40, and clinical history, one concludes
the patient has chronic respiratory acidosis with diuretic induced metabolic
alkalosis. The latter is now the dominant acid base disturbance.

Step 6: Examine serum potassium concentration

Hypokalemia
Is the hypokalemia due to a cell shift?
In the absence of physical and historical evidence of gastrointestinal or
renal potassium (K) losses, either a redistribution of K at the cellular level
or laboratory error will account for a low serum K. The regulation of K
distribution between the intracellular and extracellular space is referred to
as internal K balance. While the kidney is ultimately responsible for main-
tenance of total body K, factors that modulate internal balance are impor-
tant in the disposal of acute K loads. Cell shifts are extremely important, in
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 207

that only 2% of total body potassium is located in the extracellular uid. A

large potassium meal could potentially double extracellular K, were it not
for the rapid shift of the potassium load into cells. The kidney cannot ex-
crete potassium rapidly enough in this setting to prevent life-threatening hy-
perkalemia. Thus, it is important that this excess K be rapidly shifted and
stored in cells until the kidney has successfully excreted the K load. The ma-
jor regulators of K shift into cells are insulin and catecholamines, with
a lesser eect mediated by metabolic and respiratory alkalosis.

What is the cause of decreased total body potassium?

In the absence of cellular redistribution, a low serum K can result from
inadequate dietary intake, extra-renal losses as in gastrointestinal or skin,
or renal losses. It should be emphasized that there are overlaps among these
groups. The urinary K concentration serves as a useful guide in discerning
between these possibilities. A urine K concentration of less than 20 mEq/
L is suggestive of extra-renal losses, whereas a urine concentration of greater
than 20 mEq/L suggests renal K losses.

Inadequate dietary intake. Inadequate dietary intake is is an unusual cause

of hypokalemia. However, if patients go extended lengths of time without
potassium ingestion, hypokalemia will develop. Clinical situations associ-
ated with extreme K-decient diets include anorexia nervosa, crash diets, al-
coholism, and intestinal malabsorption. Increased renal K excretion owing
to magnesium deciency (which is often present in these clinical situations)
may contribute to the observed hypokalemia.

Extra-renal potassium losses. Sweat, with its low concentration of potas-

sium, is an unusual cause of K depletion. However, during physical training
sweat losses can become substantial and K depletion may result. Gastroin-
testinal syndromes are the most common clinical disorders of extra-renal
K losses. Diarrhea truly leads to fecal K wastage and is associated with
a normal anion gap acidosis. Although usually associated with a low urinary
K concentration, the acidosis per se can lead to some degree of renal K wast-
ing through increased distal delivery of sodium. As discussed below, the
acidosis will result in K redistribution out of cells, leading to a degree of
hypokalemia that is not as severe as the degree of K depletion.

Renal potassium losses. Potassium is freely ltered by the glomerulus but is

extensively reabsorbed by the proximal tubule and loop of Henle, so that
approximately 10% of the ltered load reaches the distal nephron. The dis-
tal nephron secretes K into the tubular uid, which will be excreted. Under
most physiologic and pathologic conditions, K delivery to the distal neph-
ron remains small and does not vary, but rather the rate of secretion by
the distal nephron varies. Thus, the rate of K secretion in the distal nephron
will determine the rate of K excretion. Two of the most important
208 PALMER

physiologic determinants of renal K excretion are mineralocorticoid secre-

tion and distal sodium delivery.
Aldosterone is a major determinant of K secretion. Aldosterone-induced
stimulation of sodium reabsorption in the collecting tubule makes the lumi-
nal potential more negative, which stimulates K secretion. In addition, min-
eralocorticoids directly stimulate K secretion in the distal nephron.
Increased distal delivery of Na stimulates distal Na absorption, which
makes the luminal potential more negative and thus increases K secretion.
In addition, increased luminal ow rate lowers luminal K concentration,
which secondarily stimulates K secretion.
The dependence of K secretion on distal delivery and aldosterone levels
helps to make K excretion independent of volume status. When patients
are volume overloaded, distal delivery is increased but aldosterone is sup-
pressed. When patients are volume depleted, aldosterone is increased (sec-
ondary hyperaldosteronism), but distal delivery is decreased. In both of
the above states K homeostasis is maintained. Disruption of this balance ex-
plains many of the renal forms of hypokalemia. The approach to the patient
with hypokalemia is given in Fig. 2.

Primary mineralocorticoid excess

In primary mineralocorticoid excess, mineralocorticoid levels are in-
creased. The increase in mineralocorticoid is referred to as primary because
it is not in response to a change in volume status or a result of hyperkalemia.
An example of this would be an aldosterone-secreting tumor. Increased

Urinary K+

<20 mEq/day >20 mEq/day

Diarrhea
Blood pressure, effective blood volume

High Low-normal

Renin (R), aldosterone (A) Serum [HCO3]

Renin, A
Renal artery stenosis High
R, A
Adrenal adenoma Urine [Cl]
R, A
Cushings Low High
Liddle syndrome
NRA Diuretics
Mg++ deficiency
Bartter syndrome
Gitelman syndrome

Fig. 2. Approach to the hypokalemic patient. NRA, non-reabsorbable anion.

 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 209

mineralocorticoid activity causes renal Na retention, which leads to vol-

ume expansion, which then leads to increased distal delivery. The net result
is increased mineralocorticoid activity and increased distal delivery. The
combination of these causes renal K loss and hypokalemia. An increase
in mineralocorticoids in the absence of volume contraction leads to a high
incidence of hypertension, but edema is unusual. A so-called escape oc-
curs before sucient salt is retained for edema formation.

Primary increase in distal delivery

A primary increase in distal delivery of Na, as would be seen with a di-
uretic, leads to an increase in urinary Na excretion and volume depletion.
This leads to secondary increases in aldosterone secretion. The combination
of increased mineralocorticoid activity and increased distal delivery again
leads to renal K loss and hypokalemia. This group of disorders is dierent
from the previous group in that elevated aldosterone levels are appropriate
for the mildly decreased extracellular uid volume.
The most common cause of renal K wasting is diuretic ingestion. As is
obvious from the above discussion, any diuretic which acts proximal to
the cortical collecting tubule will increase renal K excretion. This includes
diuretics that act in the proximal tubule, (such as acetazolamide), diuretics
that act in the loop of Henle (such as furosemide and ethacrynic acid),
and diuretics that work in the early distal tubule (such as the thiazides). Os-
motic diuresis will also lead to K wasting by this mechanism. Osmotic diure-
sis occurs in poorly controlled diabetes mellitus when serum glucose rises to
higher levels than the proximal tubule can absorb. In addition, osmotic di-
uretics are used therapeutically in some conditions.
Bartter syndrome is a rare condition in which there appears to be a pri-
mary defect in the loop of Henle salt absorption. This leads to chronic
increases in distal delivery and a chronic state of mild volume contraction.
Patients are hypokalemic, with very high renin and aldosterone levels.
In certain conditions, Na is delivered distally with a nonreabsorbable an-
ion. This will increase the lumen negative potential dierence and lead to K
wasting. All penicillins are nonreabsorbable anions, but only a few are given
in sucient quantities to lead to signicant K wasting; carbenicillin is an
example of one of these. Ketoacids will act as nonreabsorbable anions in
patients with ketoacidosis because of diabetes, alcoholism, or starvation.
Bicarbonate can function as a nonreabsorbable anion if it is delivered to the
distal nephron in greater amounts than can be reabsorbed proximally. This oc-
curs in active vomiting, proximal renal tubular acidosis, and with
acetazolamide administration.

Hyperkalemia
Like the hypokalemic disorders, a high serum K can occur in the setting
of normal or altered body stores of K. The body has a marked ability to
210 PALMER

protect against hyperkalemia. This includes regulatory mechanisms that will

excrete excess K quickly and mechanisms that will redistribute excess K into
cells until it is excreted. All causes of hyperkalemia, therefore, involve abnor-
malities in these mechanisms. The causes of hyperkalemia are listed in Box 2.

Does the patient have pseudohyperkalemia?

In approaching a patient with a high measured serum K concentration, it
is important to remember that not all of these patients have true hyperka-
lemia. Because cell K concentrations are large and plasma K concentra-
tions are small, small leaks of K out of blood cells can have large eects
on measured serum K. Normally, when blood is allowed to clot before cen-
trifugation, enough K is released from platelets to raise serum K by approx-
imately 0.5 mEq/L. This is accounted for within normal limits. However,
excessive errors can occur in the presence of marked thrombocytosis,
marked leukocytosis, or hemolysis on obtaining blood samples. These con-
ditions are referred to as pseudohyperkalemia.

Is the hyperkalemia the result of a cellular shift?

Cellular redistribution is a more important cause of hyperkalemia than
hypokalemia. One should realize that as little as a 2% shift of intracellular
K to the extracellular uid will result in a serum K of 8 mEq/L. Metabolic
acidosis promotes K exit from cells dependent upon the type of acid present.
Mineral acidosis (NH4Cl or HCl), by virtue of the relative impermeability of
the chloride anion, results in the greatest eux of K from cells, while or-
ganic acidosis (ie, lactic, b-hydroxybutyric, or methylmalonic acid) result
in no signicant eux of K. Acute respiratory acidosis also results in a small
shift of K out of cells.

Box 2. Causes of hyperkalemia

 Pseudohyperkalemia
 Cellular redistribution
Mineral acidosis
Cell shrinkage (hypertonicity)
Deficiency of insulin
b-Blockers
Hyperkalemic periodic paralysis
Cell injury
 Excess intake (very rare)
 Decreased renal excretion
Decreased distal delivery of Na (oliguric renal failure)
Mineralocorticoid deficiency
Defect of cortical collecting tubule
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 211

Beta-adrenergic blocking agents can interfere with the disposal of acute

K loads. Other drugs that can result in hyperkalemia include the depolariz-
ing muscle relaxant succinylcholine and severe digitalis poisoning. Hyperka-
lemic periodic paralysis is a rare autosomal dominant disorder characterized
by repeated bouts of paralysis associated with hyperkalemia.
As 98% of body potassium is located within cells, cell death can result in
substantial endogenous loads of K. Muscle breakdown from crush injury or
rhabdomyolysis can be associated with a substantial increase in serum K
concentration. Cell death as seen in tumor lysis syndromes can also be as
source of substantial K loads. These syndromes are often associated with
compromised renal function.
Although redistribution of K can result in hyperkalemia, the rise in K is
generally mild and nonsustained. Prolonged and severe hyperkalemia im-
plies the presence of concomitant decreases in renal K excretion. After ex-
cluding pseudohyperkalemia and a cell shift, one has to consider
a disorder in renal potassium excretion.

Why does the patient have a disturbance in renal potassium excretion?

Decreased renal excretion of K can be caused by one or more of three
abnormalities: decreased distal delivery of Na, mineralocorticoid deciency,
and abnormal cortical collecting tubule function.

Decreased distal delivery of sodium. As discussed previously, most of ltered

K is reabsorbed before the distal tubule. K excretion is then determined by
the rate at which K is secreted in the distal nephron. Acute decreases in
GFR, as occur in acute renal failure, would therefore not be expected to
have a marked eect on K excretion. Acute decreases in GFR may, how-
ever, lead to marked decreases in distal delivery of salt and water, which
may secondarily decrease distal K secretion. Thus, when acute renal failure
is oliguric, hyperkalemia is a frequent problem; when acute renal failure is
nonoliguric, distal delivery is usually sucient and hyperkalemia is unusual.
Chronic renal failure is more complicated than acute renal failure. In ad-
dition to the decreased GFR and secondary decrease in distal delivery, there
is nephron dropout and less collecting tubule mass to secrete K. However,
this is counterbalanced by a K adaptation, in which the remaining nephrons
develop an increased ability to excrete K. In addition, these patients possess
two other defenses against hyperkalemia. First, in response to a K load they
will redistribute the K into cells faster than normal people. Second, they
have a markedly increased rate of K excretion in their stool. Thus, although
patients with chronic renal failure do not excrete a K load as fast as normal
people, hyperkalemia is unusual until chronic renal failure has progressed to
GFRs less than 5 mL per minute. The occurrence of hyperkalemia with
a GFR of greater than 10 mL per minute should raise the question of de-
creased mineralocorticoid activity or a specic lesion of the cortical collect-
ing tubule.
212 PALMER

Decreased mineralocorticoid activity. Aldosterone deciency can occur

alone or in combination with decreased cortisol levels. Addisons disease
is the deciency of aldosterone and cortisol because of destruction of the ad-
renal glands. Certain enzyme defects can result in either isolated deciency
of aldosterone or adrenogenital syndromes associated with decreased miner-
alocorticoid activity. Heparin administration is associated with decreased
adrenal secretion of aldosterone.
Angiotensin converting enzyme inhibitors will lead to hyperkalemia by
decreasing angiotensin II levels, a critical mediator of aldosterone secretion.
Renin levels are high and aldosterone levels are low in all of these condi-
tions. The syndrome of hyporeninemic hypoaldosteronism accounts for
the majority of unexplained hyperkalemia in patients where the GFR and
K intake would not be expected to result in hyperkalemia. Diabetic ne-
phropathy and interstitial renal disease are the most common clinical enti-
ties associated with this syndrome. Other causes of renal disease
associated with hyporeninemic hypoaldosteronism include analgesic ne-
phropathy, urinary tract obstruction, sickle cell disease, systemic lupus er-
ythematosus, and amyloidosis. Nonsteroidal anti-inammatory drugs are
associated with decreased renin secretion. Additionally, these agents can
cause hyperkalemia by decreasing GFR and reducing distal delivery of so-
dium. Cyclosporine administration is associated with the development of
hyperkalemia in renal transplant patients. Although cyclosporine has a di-
rect eect on the renal tubule, many patients taking this drug have low renin
and aldosterone levels. These patients additionally may have a primary tu-
bular defect at the level of the tubule independent of cyclosporine. Beta re-
ceptor blockade will also result in a hyporeninemic state.

Distal tubular defect. Certain interstitial renal diseases can aect the distal
nephron specically and lead to hyperkalemia in the presence of mild de-
creases in GFR and normal aldosterone levels. Many of these diseases are
the same ones that can cause hyporeninemic hypoaldosteronism, and fre-
quently the impaired renin release and defect in tubular secretion coexist.
Examples include renal transplant patients, systemic lupus erythematosus,
amyloidosis, urinary obstruction, and sickle cell disease.
The K sparing diuretics impair the ability of the cortical collecting tubule
to secrete K. Amiloride and triamterene inhibit Na reabsorption, which
abolishes the lumen negative potential and therefore inhibits K secretion.
Other compounds that block the Na channel and that have been clinically
associated with hyperkalemia include trimethoprim and pentamidine. Spiro-
nolactone competes with aldosterone and thus blocks the mineralocorticoid
eect. Although the potassium sparing diuretics are useful in patients with
a hypokalemic tendency, they weaken an important defense mechanism
against hyperkalemia. They should therefore be avoided in patients with
other defects that predispose to hyperkalemia, such as diabetes mellitus
and chronic renal insuciency.
 APPROACH TO FLUID AND ELECTROLYTE DISORDERS 213

Further readings
Naderi A, Palmer BF. Respiratory acid-base disorders. In: Singh AK, Sayegh M, Suki W, et al,
editors. Suki and Massrys therapy of renal diseases and related disorders. 2008, in press.
Palmer BF, Alpern RJ. Metabolic acidosis. In: Johnson RJ, Feehally J, Floege J, editors. Com-
prehensive clinical nephrology. 3rd edition. Philadelphia: Elsevier Inc; 2007. p. 14757.
Palmer BF, Alpern RJ. Metabolic alkalosis. J Am Soc Nephrol 1997;8:14629.
Palmer BF, Alpern RJ. Normal acid-base balance. In: Johnson RJ, Feehally J, Floege J, editors.
Comprehensive clinical nephrology. 3rd edition. Philadelphia: Elsevier Inc; 2007. p. 1416.
Palmer BF. Hyponatremia. In: Rakel RE, Bope E, editors. Conns current therapy. Philadelphia:
Saunders Elsevier; 2007. p. 6936.
Palmer BF. Managing hyperkalemia caused by inhibitors of the renin-angiotensin-aldosterone
system. N Engl J Med 2004;351:58592.
Sterns R, Palmer BF. Fluid, electrolyte and acid-base disturbances. NephSap 2007;6:21080.