Q J Med 2005; 98:305316

doi:10.1093/qjmed/hci046

Advance Access publication 10 March 2005

Masterclasses in medicine
An approach to the patient with severe hypokalaemia:
the potassium quiz
J.H.M. GROENEVELD1, Y.W.J. SIJPKENS1, S.-H. LIN2, M.R. DAVIDS3
and M.L. HALPERIN4
From the 1Department of Nephrology, Leiden University Medical Center, The Netherlands,
2
Renal division, Tri-Service General Hospital, National Defense Medical Center, Taipei,
Taiwan (ROC), 3Nephrology Unit and Department of Internal Medicine,
Stellenbosch University, Cape Town, South Africa, and 4Division of Nephrology,
St. Michaels Hospital, University of Toronto, Toronto, Canada

The objective of this teaching session with Professor

McCance is to develop an approach to the management of patients with a very low plasma potassium
(K) concentration (PK). The session begins with a
quiz based on six recent medical consultations for
a PK 5 2 mmol/l. Professor McCance outlined how
he would proceed with his diagnosis and therapy,
using the synopsis that described each patient.
This approach was then applied to a new patient,

a 69-year-old woman who had a large volume of

dependant oedema and developed a severe degree
of weakness and hypokalaemia during more aggressive diuretic therapy that included a K-sparing
diuretic. The initial challenge for Professor
McCance was to deduce why the K-sparing
diuretic was not effective in this patient. He also
needed to explain why the PK was so low on
admission.

Introduction
In this clinical teaching exercise, the central figure is
Professor McCance, an imaginary consultant who
practiced medicine
70 years ago. The overall
objective is to demonstrate how applying principles
of integrative physiology at the bedside can be
extremely helpful in revealing the pathophysiology
of disease, making more accurate clinical diagnoses, and to plan optimal therapy.

The consultation
When the medical registrar received the first referral
of the day, she was surprised that it was another

request to evaluate a patient who had a very low

plasma potassium (K) concentration (PK). There
seems to be an epidemic of hypokalaemia, she
mused, because their team had seen six patients with
a similar reason for a medical consult that year.
In the present consultation, the patient is a
69-year-old woman with a medical history of
Parkinsons disease, hypertension and a large
volume of dependent oedema. She had normal
plasma electrolyte concentrations before her
diuretic therapy was doubled to thiazide 50 mg/
day and triamterene 25 mg/day. With this change,
she developed severe hypokalaemia. None of her

Address correspondence to Professor M.L. Halperin, University of Toronto, St Michaels Hospital Annex, Lab #1,
Research Wing, 38 Shuter Street, Toronto, Ontario, M5B 1A6, Canada. email: mitchell.halperin@utoronto.ca
! The Author 2005. Published by Oxford University Press on behalf of the Association of Physicians.
All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

Downloaded from by guest on September 27, 2015

Summary

306

J.H.M. Groeneveld et al.

other medications was likely to influence her PK.

In the first two weeks after this increase in diuretic
therapy, there was a marked decrease of oedema
fluid, but she felt very weak and had a decrease in
appetite. On examination in the emergency department, her blood pressure was only 110/60 mmHg,
and she had obvious muscle wasting. Laboratory
testing revealed hypokalaemia (PK 1.7 mmol/l),
metabolic alkalosis (pH 7.54, plasma bicarbonate
(HCO
3 ) concentration (PHCO3) 43 mmol/l), and an
urine K concentration (UK) of 36 mmol/l (Table 1).
Knowing that they would meet Professor
McCance on rounds the next morning, the registrar

asked one of the interns to prepare a synopsis of

the six recent cases with severe hypokalaemia
and present it to their Professor as a diagnostic
challenge. In their minds, he was the ultimate
clinical detective and would no doubt relish the
opportunity to grapple with these interesting cases.
As to their current case, they had one major
question for their Professor, Why might renal
K excretion be so high while on a K-sparing
diuretic?

Table 1 Laboratory data in our patient

The medical team was eager to discuss the present

consultation as well as learn how their Professor
would respond to the K-quiz (Table 2). Would he
have any difficulty matching the case vignettes with
their list of possible diagnoses? they wondered.
They would soon discover that generic approaches,
while useful as a guide, could not replace a
deductive and thorough interpretation based on
integrative physiology for each individual case.
Professor McCance stood at the blackboard with
chalk in handthe focus was on the K-quiz.
Because he loved problem-solving, he immediately
accepted the challenge. He knew that while
simply matching the cases with the diagnoses

Unit

Admission

Day 9

Na
K
Cl
HCO
3
pH
PaCO2
Albumin
Magnesium
Creatinine
Urea

mmol/l
mmol/l
mmol/l
mmol/l

148
1.7

43
7.54
50
33 (3.3)
0.79 (1.9)
84 (1.0)
8.6 (24)

146
3.2
104
30
7.44
38
35 (3.5)

58 (0.7)

mmHg
g/l (g/dl)
mmol/l (mg/dl)
mmol/l (mg/dl)
mmol/l (mg/dl)

Table 2 Synopsis of data for the potassium quiz

Patient . . .

Age
Gender
BP (mmHg)
Heart rate (bpm)

28
F
160/80
124

26
M
130/70
72

28
F
110/80
72

58
F
120/70
80

21
M
110/70
64

78
M
180/100
74

Plasma
K (mmol/l)
HCO
3 (mmol/l)
Mg2 (mmol/l)
Inorganic P (mmol/l)

1.8
24
0.7
0.6

1.8
24
0.8
0.9

1.8
15
0.9
0.8

1.8
31
0.7
0.8

1.8
31
0.5
0.8

1.8
31
0.8
0.9

Urine
Na (mmol/l)
K (mmol/l)
Cl (mmol/l)
Creatinine (mmol/l)
Osmolality (mOsm/kg)
UCa /UCreat (mmol/mmol)
UMg /UCreat (mmol/mmol)
UPi /UCreat (mmol/mmol)
UOsm /UCreat (mOsm/mmol)

102
10
98
10
680
0.63
0.2
0.1
68

84
8
90
11
586
0.18
0.25
0.9
53

44
11
46
3
201
0.39
0.24
3.3
67

72
22
78
4
314
0.56
0.81
1.3
78

146
12
188
4.3
660
0.06
0.66
1.5
77

120
14
117
5
348
0.26
0.4
1.6
67

List of diagnoses: (a) Bartter-like syndrome; (b) Chronic liquorice ingestion; (c) Gitelmans syndrome (GS); (d) Renal tubular
acidosis (RTA); (e) Sporadic periodic paralysis (SPP); (f) Thyrotoxic periodic paralysis (TPP).

Downloaded from by guest on September 27, 2015

Plasma

On rounds the next day

with professor McCance

Masterclasses

307

Flow Chart 1. Initial steps in the patient with a very low PK.

Question 1. What are the major dangers

faced by a patient with an extremely
low PK?
Physiology principle 1. There are two factors that
can influence the movement of K across cell
membranes, the driving force which is the electrical
voltage across cell membranes, and the presence of
open K channels.1 The ratio of K concentrations
in the ICF and ECF compartments reflects this
electrical driving force.
Return to the bedside. The major problem for a
patient with an abnormal trans-membrane voltage
is when it affects the heart because life-threatening
cardiac arrhythmias can develop. This possible
threat is best evaluated by examining the EKG.

A second issue for an extreme degree of hypokalaemia is weakness of the respiratory muscles,
especially if there is a need for increased ventilation
(e.g. metabolic acidosis).

A. Patients with a low rate

of excretion of K
At this point, the intern stated that there was no
emergency demanding urgent therapy in any of their
cases because only U waves were seen on the
EKG, there was no apparent respiratory muscle
weakness, and the arterial PCO2 was not unduly
high. Therefore Professor McCance proceeded with
diagnostic issues, because this would reveal the
goals for therapy. For example, if a shift of K into
cells were the most important cause for hypokalaemia, the goal of therapy would be to reverse this
shift without causing a large negative balance
for K. In contrast, if the low PK were due primarily
to a large total body deficit of K, the major goals for
therapy would be to decrease K loss and to replace
the deficit of K, typically with a large input of
potassium chloride (KCl). Because the PK is so low,
he asked, Which patients are more likely to have a
major shift of K into cells?

Question 2. Which patients are

more likely to have a major shift
of K into cells?
Physiology principle 1, restated. The driving force to
cause a K shift into cells is a more negative voltage

Downloaded from by guest on September 27, 2015

might be entertaining, it would be of little benefit

to his younger colleagues, and he was never one to
miss a good teaching opportunity. Rather, illustrating a physiology-based approach would be a better
way to proceed. Changing the rules of the game
somewhat, he announced that he would solve
the puzzle while developing a diagnostic algorithm,
and invited them to assist in this task.
The initial difficulty in designing a flow chart is
to define a rational starting point. The first question should address potential emergencies for the
patient, said Professor McCance (Flow Chart 1).
Accordingly, he began by asking, What are the
major dangers faced by a patient with an extremely
low PK?

308

J.H.M. Groeneveld et al.

Figure 1. Creation of a more negative voltage in cells. The circle depicts the cell membrane. The Na-K-ATPase pumps
positive voltage out of cells, causing a large inside negative voltage (60 to 90 mV). This ion pump is activated by
b2-adrenergics. Insulin, by activating the Na/H exchanger (NHE), causes the electroneutral entry of Na into cells, and
thereby more positive voltage exit from cells via the Na-K-ATPase. K exits cells through ion channels that are in a
sufficiently open configuration to approach, but not reach, the electrochemical equilibrium for K.

Question 3. What features on clinical

presentation suggest that the sole
basis for a very low PK might be a
shift of K into cells?
Physiology principle 2. There is a very large quantity
of K in cells (
50 mmol/kg body weight), and the
rate of K excretion can decline to 0.2 mmol/kg/day
when the PK begins to fall. Therefore it will usually
take many months of poor K intake to be the sole

cause of a very low PK. In addition, if the PK falls

in a much shorter time, look for a cause other than
just a low K intake.
Return to the bedside. While not always reliable,
the first clue to suggest that the basis for a severe
degree of hypokalaemia is a shift of K into cells is
its timingdid it occur in a matter of hours, rather
than days, weeks or months? If the answer is yes,
suspect that there is an important component of
K shift into cells. This impression could be supported if the patient was also suffering from acute
paralysis.
When acute hypokalaemic periodic paralysis
(HPP) is suspected, there are a number of other
supporting facts to help in this regard.6 It is common
to find provoking factors such as a high carbohydrate meal (high insulin levels activate NHE and the
Na-K-ATPase) and vigorous exercise (b-adrenergic
agonists activate Na-K-ATPase). It is particularly
important to look for clinical evidence of hyperthyroidism, including a wide pulse pressure and
tachycardia. The helpful laboratory data suggestive
of HPP are a low rate of excretion of K and the
absence of an acid-base disorder.6 More detailed
information is provided in Table 3.
Because a low rate of excretion of K is so
important in this differential diagnosis and it is
readily obtained from simple laboratory tests, the
initial step McCance used to construct Flow Chart 2
was to divide patients into two major pathophysiological groups, based on their current rate of
excretion of K. It is not necessary or desirable to
wait for a timed urine collection for this purpose;
the same information can be obtained by comparing
the UK to the concentration of another urinary
constituent that is excreted at a constant rate,
such as creatinine (UCreatinine).7 The index I use to
assess the K excretion rate is the UK/UCreatinine

Downloaded from by guest on September 27, 2015

in cells; this is created by the net export of positively

charged ions (Figure 1). If a more negative intracellular fluid (ICF) voltage were the only defect, one
should have hypokalaemia with few other biochemical abnormalities.
The nephrology consultant explained that the
Na-K-ATPase pumps positive voltage out of cells;
it extrudes three sodium ions (Na) for every
two K ions that enter cells2,3 (Figure 1). The
main hormones that increase the activity of this
Na-K-ATPase are b2-adrenergic agonists and thyroid hormone.4 There is a second way to pump
more Na out of cells without activating the
Na-K-ATPase per se: increase in the concentration
of Na in cells. The mechanism begins when Na
ions enter cells in an electroneutral fashion; this is
achieved by activating the Na/H exchanger
(NHE) in cell membranes, a transporter that is
normally inactive in cell membranes unless insulin
levels rise appreciably (Figure 1).5 This helps to
prevent hyperkalaemia when K is ingested in
food that also contains sugar.
With that important background information,
the team had no problems responding to McCances
next question: What features on clinical presentation suggest that the sole basis for a very low PK
might be a shift of K into cells?

Masterclasses
ratio, said Professor McCance. Patients with a very
low K excretion rate (
1015 mmol/day8) and a
creatinine excretion rate of 1015 mmol/day should
have a UK/UCreatinine that is 51.5 in mmol terms,
Table 3 Features in patients with hypokalaemic periodic
paralysis

Patient
Race
Age (years)
Gender
History
Provoked by high
carbohydrate
Provoked by
vigorous exercise
Symptoms of
hyperthyroidism
Family history of
paralysis

Laboratory data
Low K excretion
Acid-base disorder
Hypophosphataemia
Hypophosphaturia
High urine Ca
EKG
Specific therapy

FPP or SPP

Asian
2040
Male4female

All
520
Male4female

Often

Often

Often

Often

Often (50%)

No

No

Often

Often
Often

Rare
Rare

Yes
No
Often
Often
Yes
LVH,
tachycardia
b-blocker

Yes
No
Less often
Less often
No
Normal
Acetazolamide

Flow Chart 2. Approach to the patient with low K excretion.

or 15 in UK /g UCreatinine for those who use decadent

units for creatinine, he said with a smile. There is a
caveat when using this UK /UCreatinine ratio: because
creatinine is derived from skeletal muscle, one must
make an adjustment in patients who have a very low
or very high muscle mass.9 Before leaving this topic,
a member of the medical team asked, Might a
patient develop a very low PK without having a shift
of K into cells or a high UK /UCreatinine?

Question 4. Might a patient develop

a very low PK without having a shift
of K into cells or a high current
UK /UCreatinine?
Physiology principle 3. Be certain that the data
you examine represent steady state values, so that
they can be interpreted correctly in a chronic
condition.
Return to the bedside. If I do not have to consider
a shift of K into cells, let me turn my attention to
settings where the excretion of K may be intermittent and driven at times by non-physiological
stimuli, said Professor McCance. If a patient were
to take a drug that augmented the excretion of K,
this rate of excretion would be high when the drug
acted and low when the drug was not being used.
The best example of a former K excretion would
be the use of diuretics in the past, but not currently.
Now the UK / UCreatinine should be low because the
patient did not take a diuretic recently.
Return to the K-quiz. Using the above principles,
Professor McCance deduced that patient 1 had
thyrotoxic periodic paralysis (TPP), whereas

Downloaded from by guest on September 27, 2015

Physical examination
Wide pulse pressure
Tachycardia

TPP

309

310

J.H.M. Groeneveld et al.

patient 2 had familial periodic paralysis (FPP) or

spontaneous periodic paralysis (SPP).
One of the registrars pointed out that it is important to make a diagnosis of TPP because there is
now a specific and effective therapy for this type
of HPP. Administering a large dose of a non-specific
b-adrenergic receptor blocker (propranalol 3 mg/kg)
quickly reverses both the paralysis and the hypokalaemia.10

B. Patients with a high rate

of excretion of K
All the remaining cases in Table 2 had a high
UK/UCreatinine and an acid-base disorder. Professor
McCance therefore turned to Flow Chart 3 and
asked: How should we subdivide the patients with
a very low PK and a high UK /UCreatinine?

Physiology principle 4. The rate of excretion of

K is dependent on two factors: the flow rate and
the K concentration in the nephron segment that
controls the secretion of K, the cortical collecting duct ([K]CCD, equation 1).11 The flow rate is
dependent on the osmole delivery rate12 and
the [K]CCD is largely determined by the negative voltage in the lumen of the terminal CCD.
This voltage is reflected by the transtubular K

Flow Chart 3. Approach to the patient with high K excretion.

K excretion KCCD  Flow rateCCD

In-depth look at the K secretory process in the

CCD. To have a high rate of K excretion, one
usually needs a high [K]CCD. This occurs when a
lumen-negative voltage is generated by reabsorbing
Na ions at a faster rate than Cl ions (Figure 2, top
left side). This can occur if there is delivery of Na
with an anion other than Cl to the CCD, together
with a stimulus to reabsorb Na (high aldosterone
level), in conditions where ENaC is active due to a
molecular lesion, or in situations where cortisol acts
like a mineralocorticoid. Measuring the activity of
renin and the level of aldosterone in plasma can
help in this differential diagnosis (Figure 3).
The nephrology consultant added one more
observation. The presence of HCO
3 ions in the
lumen of the CCD can cause a faster rate of Na
than Cl ion reabsorption, because HCO
3 ions
may slow the rate of Cl ion reabsorption in the
CCD (Figure 2, bottom right). Professor McCance
now applied this physiology of K excretion to
the patients with an acid-base disorder: first those
patients with metabolic acidosis and then those with
metabolic alkalosis.

(i) Group with metabolic acidosis:

Physiology principle 5. Patients with hyperchloraemic metabolic acidosis and a high UK/UCreatinine
can be separated into two groups based on their rate

Downloaded from by guest on September 27, 2015

Question 5. What should the initial

step be to subdivide the patients with a
very low PK and a high UK /UCreatinine?

concentration ratio (TTKG).13

Masterclasses

311

High ECFV, Metabolic Alkalosis

What are P Renin and
PAldosterone ?

Both high
Malignant High BP
Renovascular High BP
Renin-secreting tumor

High
Ectopic ACTH
Cushing's syndrome
Exogenous cortisol

Both low
P

Cortisol?

Low
11 -hydroxylase
17 -hydroxylase

Low P Renin
High PAldosterone
Hyperadrenal state
Exogenous aldosterone

Normal
More open ENaC
AME-like
Other

Figure 3. Plasma renin and aldosterone in the syndromes of mineralocorticoid excess. For details, see text. AME, apparent
mineralocorticoid excess syndrome; BP, blood pressure; ENaC, epithelial Na channel.

Downloaded from by guest on September 27, 2015

Figure 2. Components of K excretion in the terminal CCD. The barrel shaped structures represent the terminal CCD.
Reabsorbing Na faster than Cl in the CCD creates the lumen negative voltage that drives the net secretion of K.
This occurs when aldosterone (or cortisol) activates ENaC (upper right). Lower left, Na can be reabsorbed faster than Cl
if there is little Cl delivery and Na is accompanied by a non-Cl anion. Lower right, luminal HCO
3 can slow the
reabsorption of Cl. ENaC, epithelial Na channel.

312

J.H.M. Groeneveld et al.

(ii) Group with metabolic alkalosis

of ammonium (NH
4 ) excretion. Those with a high
rate of NH
4 excretion have a non-renal, direct loss
of NaHCO3 (e.g. diarrhoea14) or an indirect loss of
NaHCO3 where there is production of an acid such
as hippuric acid, followed by the excretion of some
hippurate anions with Na or K instead of NH
4,
Figure 4).15 The group of patients with a low rate of
excretion of NH
4 has a renal defect that leads to a
low rate of excretion of NH
4 (called distal renal
tubular acidosis or RTA).16 Professor McCance
added one more clinical pearl. NH
4 in the urine
can be detected by its charge (urine net charge17) or
as a urinary particle (urine osmolal gap18).
Returning to the K quiz, Professor McCance
pointed out that patient 3 had a UK/UCreatinine 41.5,
and a urine that appeared to have a low NH
4
concentration, as judged from the urine electrolyte
concentrations (Na K 4 Cl). Because Uosm
was not appreciably higher than 2(UNa UK)
urine urea (not shown, but inferred), her urine flow
rate was not high, so the NH
4 excretion rate should
not be appropriately elevated.
The nephrology consultant added that the two
lesions to suspect are distal RTA due to impaired H
secretion in the distal nephron,19 or the secretion
of HCO
3 in this nephron segment as occurs in
certain patients with an abnormal Cl/HCO3 anion
exchanger.20 None of the patients had evidence suggesting a high rate of NH
4 excretion (e.g. patients
with diarrhoea or others who sniff glue producing
hippuric acid15).

Downloaded from by guest on September 27, 2015

Figure 4. Indirect loss of NaHCO3 in glue sniffing. The

metabolism of toluene occurs in the liver, where benzoic
acid is produced via alcohol and aldehyde dehydrogenases. Hippuric acid is produced due to conjugation
with glycine (all represented as site 1 for simplicity).
The H are titrated by HCO
3 for the most part (site 2).
The hippurate anion is secreted by the PCT and excreted
in the urine, initially with NH
4 (site 4) and then with
Na and K when the capacity to excrete NH
4 is
exceeded (site 5). The excretion of hippurate anions
with Na and/or K (and not NH
4 ) is a major factor in the
metabolic acidosis.

The first question asked by Professor McCance

concerned the blood pressure of these patients. His
rationale was that a high blood pressure along with
a high UK/UCreatinine and TTKG would suggest that
its basis was faster reabsorption of Na via ENaC in
the CCD if the patient was not receiving diuretics.
The nephrology consultant added the following
comments. The list of causes include: an inborn
error of metabolism that results in higher ENaC
levels in the CCD; higher aldosterone levels; or
settings where compounds other than aldosterone
activate the mineralocorticoid receptor in principal
cells of the CCD. Measuring the plasma levels
of aldosterone, the activity of renin, and cortisol
levels in plasma can help subdivide these patients
(Figure 3). Patient 6 could have any of these lesions.
The medical team supplied additional information at
this point. Because this patient was elderly (age 78),
had absent aldosterone and normal cortisol levels in
plasma and admitted to using a large amount of
liquorice to sweeten his tea, the diagnosis was
liquorice abuse.21
The final diagnostic category with metabolic
alkalosis is the group with a very low PK, a high K
excretion rate, and the absence of hypertension.
The major subgroups include the use of diuretics,
vomiting or conditions that lead to a slower
reabsorption of Na and Cl in the loop of Henle
or the distal convoluted tubule (DCT) and thereby
an enhanced delivery of Na and Cl to the CCD
while aldosterone levels are high (lower right
portion of Figure 2). In all the patients reported in
Table 2, the urine was negative for diuretics when
the UNa and UCl were high. In addition, vomiting,
diarrhoea and laxative abuse were not present.
A word of caution is necessary at this point,
warned the nephrology consultant. Patients who
take diuretics or those who have conditions that
mimic diuretic actions such as inborn errors that
compromise Na and Cl reabsorption in the loop
of Henle or the DCT (like Bartters or Gitelmans
syndromes) rarely have such low PK values. When
present, suspect that there are other factors that
either cause a low K intake, a shift of K into cells
and/or K loss by renal or non-renal routes. He
then provided a few pointers to aid in identifying the
defective nephron segment.
Loop of Henle: Because this nephron segment
is responsible for concentrating the urine and it is an
important site for Ca2 reabsorption, expect a low
maximum Uosm and a Ca excretion rate that is
not low.
Distal convoluted tubule: Disorders affecting
Na and Cl in this nephron segment have intact

Masterclasses
concentrating ability (unless there is a second
lesion) and a low rate of excretion of Ca.9 This
nephron segment is the last one that regulates the
excretion of magnesium (Mg) and hypokalaemia is
often accompanied by hypomagnesemia, especially
later in the disease.
Based on the above, patient 5 with a modestly
high Uosm (not tested after dDAVP), a low UCa/
Ucreatinine, and a low PMg fits a DCT lesion.9 Patient 4
would best fit a loop of Henle lesion, if the Uosm
failed to rise after dDAVP was administered. Given
the age of the patient, the most likely lesion in the
loop of Henle is an acquired problem such as the
presence of a cation that occupies the calciumsensing receptor (Ca-SR) on the basolateral membrane of the medullary thick ascending limb of
the loop of Henle.22 Ligands that may bind to this
receptor include Ca2 (hypercalcaemia), cationic
drugs such as aminoglycosides, and cationic proteins (globulins) as might be present in multiple
myeloma.

313

To block this Na ion channel in the luminal

membrane in the CCD, it is likely that K-sparing
diuretics will resemble Na in some way. Compounds with an amine group have the potential
of bearing a positive charge when they bind H
(equation 2). One must also examine the pK of these
compounds, suggested McCance. The nephrology
consultant informed the group that because the
pK of triamterene or amiloride is much higher than
the luminal fluid pH in the CCD in vivo, they
would always bear a positive charge and therefore
compete with Na for entry into ENaC. In contrast,
trimethoprim has a pK
7.2, so the luminal fluid pH
can affect its ability to be cationic (luminal pH in
the high 68 range). Thus, making the luminal fluid
pH more alkaline by giving NaHCO3 can amputate
the ability of trimethoprim to block ENaC, but not
so for the other K-sparing diuretics.23 Professor
McCance therefore had to find another mechanism
to amputate the renal actions of triamterene. How
can one lower its concentration of in the lumen of
the CCD? he asked.
H Trimethoprim ! H  Trimethoprim 2

The team now turned their attention to the current

consultation. Following the steps outlined in
Flow Chart 1, there were no medical emergencies
as judged from her clinical picture, EKG and her
PaCO2. There were no urine data available in the
time when the PK fell, so I cannot comment on the
UK/Ucreatinine, said Professor McCance. Notwithstanding, given the patients age and therapy, I will
presume that renal K loss was an important
component of the very low PK. Therefore I will
turn to the question posed by the medical team that
cared for this patient.

no effect on ENaC

Why might the renal excretion of K

so high while on a K-sparing diuretic?
For a K-sparing diuretic to be taken but not act,
its effect on the kidney must be amputated,
pronounced Professor McCance. I was told that
this class of diuretics should prevent the generation
of a lumen-negative voltage in the CCD and thereby
diminish the secretion of K. Although there are
several possibilities to do this (Figure 2), I would
guess that blocking Na reabsorption rather than
accelerating Cl reabsorption in the CCD would be
the most likely mechanism, because this class of
drugs promotes the excretion of Na (they are
diuretics). The nephrology consultant confirmed that
Na reabsorption in the CCD occurs via a specific
ion channel, ENaC and that this class of K-sparing
diuretics acts by blocking the ENaC in the luminal
membrane of the principal cells of the CCD.

blocks ENaC

Question 6. How can one lower the

concentration of triamterene in the
lumen of the CCD?
Physiology principle 6. Concentration terms have
numerators and denominators.
Return to the bedside. If the drug entered the
tubular lumen as expected, one can lower its concentration by having a very large flow rate in the
CCD. Hence the next question is, How can the flow
rate in the CCD be raised?

Question 7. How can the flow rate

in the CCD be raised?
Physiology principle 7. The CCD is permeable to
water when vasopressin acts. Therefore the osmolality in its luminal fluid will equal that of the
interstitial fluid in the cortex (or the plasma osmolality, Posm). Hence the number of osmoles delivered
to the CCD will determine the volume in the lumen
of the CCD (equation 3) (Figure 5).24
Flow rateCCD Osmoles deliveredCCD =Posm

Return to the bedside. The major osmoles delivered

to the CCD are urea and electrolytes. The other
factor that permits a high rate of secretion of K

Downloaded from by guest on September 27, 2015

Current consultation

314

J.H.M. Groeneveld et al.

Figure 5. Non-invasive estimate of the flow rate in the terminal CCD. The barrel-shaped structure represents the CCD.
When ADH acts, the Posm and the osmolality in the luminal fluid of the CCD are equal (represented as 300 mOsm/kg
H2O for easy calculation). For example, if 600 mOsm are excreted in a given time, the minimum flow rate in the terminal
CCD is 2 l.

if the b-adrenergic effect was dominant, this could

activate the Na-K-ATPase, make the ICF voltage
become more negative, and thereby shift K into
cells (Figure 1).
Metabolic alkalosis: The patient had a severe
degree of metabolic alkalosis so this could be an
important factor that caused K to shift into cells.
Before Professor McCance could continue, the
razor-sharp student raised his hand and asked:
Without vomiting and HCl loss, the presence of
little dietary intake of alkali, and poor renal function
(adjusting her PCreatinine for muscle mass), how could
this patient have such a high PHCO3?

Question 8. Why might this patient

have such a high PHCO3?
Physiology principle 8. The PHCO3 can rise because
of an addition of HCO
3 or a fall in the ECF volume.
Return to the bedside. While there is little
evidence for a gain of HCO
3 , there was a marked
reduction in her ECF volume (loss of oedema fluid).
It would not be surprising for this to almost double
her PHCO3 by halving her ECF volume.14 This is
a contraction form of metabolic alkalosis.26 In fact,
this should be the most common setting for
contraction alkalosis, the administration of diuretics
to a person with a reasonable GFR and a large
volume of retained oedema fluid.
Professor McCance nodded in agreement, but
he had not finished his analysis of the basis for
the patients low PK. There are two factors I can
identify, he said. First, for the same amount of K

Downloaded from by guest on September 27, 2015

in the CCD is aldosterone, because it causes the

ENaC to be open. The low effective circulating
volume will provide this signal to release aldosterone, because it activates the renin-angiotensin
system. In this patient, the higher dose of thiazides
led to a larger delivery of Na and Cl to the CCD.
Because there was such a large excess of Na and
Cl in her body (the oedema fluid), she was able to
have a large enough volume in her CCD to dilute
the concentration of triamterene and minimize its
renal effect.
If you want to ensure a K -sparing effect, perhaps
one should use drugs that block aldosterone action
at a site other than ENaC, such as the mineralocorticoid receptor in principal cells, volunteered a
final year student. An excellent suggestion and one
I will keep in mind when next I have a similar case,
agreed the nephrology consultant.
Not wishing to omit anything, Professor McCance
also considered factors that could cause K to shift
into cells in this patient. Following the rationale of
physiology principle 1, the major factors that cause
K to shift into cells are hormones, metabolic
alkalosis and cell growth. While Professor McCance
could rule out cell growth with confidence in this
patient, he could see a possible role for hormones
and metabolic alkalosis.
Hormones: Because of the actions of the diuretics,
her effective circulating volume was low. The latter
could invoke an adrenergic response. If the
a-adrenergic effect dominated, there would be an
inhibited release of insulin from b-cells (25) and a
tendency for the PK to rise. On the other hand,

Masterclasses
loss, the presence of a low K intake would lead to
larger K deficit and thereby, a lower PK. While this
was one factor the medical team had considered,
they were unable to describe a second one that
could have caused such a severe degree of
hypokalaemia. Therefore Professor McCance asked
his question in a more direct fashion. What could
cause the PK to be lower in one of two patients, if
both patients had the same total body K deficit and
identical factors influencing the distribution of K
across cell membranes?

Question 9. What could cause the

PK to be lower in one of two patients
given the same K deficit and identical
factors influencing the distribution
of K across cell membranes?

Therefore this could contribute to the severity of

the degree of hypokalaemia.

Concluding remarks
Professor McCance said that he enjoyed participating in the K-quiz. What he found even more
rewarding was the opportunity to introduce his
younger colleagues to the application of principles
of integrative physiology (Table 4), to permit
them to deduce answers that may be correct.
He also emphasized the importance of a quantitative analysis. After considerable thought, he
developed Flow Charts 13 to illustrate an
approach to accurate diagnosis and therefore
logical therapy for patients with a severe degree of
hypokalaemia.
Regarding the new consultation, his message
was that more than one mechanism is likely to be
involved to produce such a severe abnormality.
Adding to this diuretic-induced K loss, were a low
intake of K, a shift of K into cells (hormone actions
and metabolic alkalosis), and an ability to amputate the renal effects of a K-sparing diuretic. The
latter could be explained when he considered
both the numerator and denominator of the concentration of this drug in the lumen of the CCD.

Table 4 Principles of physiology in this case

Physiology principle

Comments

1.

The negative voltage in cells is created by

Na-K-ATPase using Na that is already
inside cells or Na that entered (via NHE)

Assess the EKG for cardiac arrhythmias. K moves into cells

due to hormones, anabolism, metabolic alkalosis

2.

The vast majority of K in the body is in

cells (
4000 mmol in a 70 kg adult)

It takes many weeks of low K intake to decrease the PK to

52 mmol/l

3.

Examine intake, shift, and excretion of K to

establish the cause of the low PK

There is no acid-base disturbance with an acute K shift into

cells (e.g. HPP)

4.

K excretion rate in a spot urine is reflected

by the UK/Ucreatinine ratio

Expected UK/Ucreatinine is 51.5 in mmol terms when renal

hyperkaluria is absent

5.

Patients with metabolic acidosis and a normal

anion gap can be divided into two groups
depending on UNH4

With a low UNH4  V, look next at the urine pH. Those with high
UNH4  V have direct or indirect non-renal NaHCO3 loss

6.

Concentration terms have numerators

and denominators

Drug concentration in the CCD will fall with high flow rates to
the CCD

7.

The flow rate in the CCD is determined by the

osmole delivery rate when vasopressin acts

The major osmoles delivered to the CCD are urea and Na Cl

8.

Metabolic alkalosis is due to a rise in the

content of HCO
3 in the ECF and/or a low
ECF volume. The kidney does not excrete
this extra HCO
3

Measuring the UCl helps define the cause of metabolic alkalosis.

Look for diuretics or LOH lesion if the urine contains
abundant Cl

9.

The [K]CCDUK/(U/P)osm

TTKG should be 51.5 in K depletion

Downloaded from by guest on September 27, 2015

Physiology principle 9. For the same ratio of K

concentrations in the ICF to ECF compartments,
the smaller the ICF compartment, the lower the PK
for a given total K deficit.
Return to the bedside. Since the vast majority of
the ICF volume is in skeletal muscle and the patient
has muscle wasting due to Parkinsons disease,
her total ICF K content should be very low.

315

316

J.H.M. Groeneveld et al.

Finally, he added that one must consider the

magnitude of the K deficit and the size of the total
body K store (mainly muscle mass) to determine
how a modest deficit of K might produce such a
severe degree of hypokalaemia. The basis for this
low K store capacity was muscle wasting.

References
1. Rosa RM, Williams ME, Epstein FH. Extrarenal potassium
metabolism. In: Seldin DW, Giebisch G, eds. The Kidney:
Physiology and Pathophysiology, 2nd edn. New York, Raven
Press, 1992:216590.
2. Russell JM. Sodium-potassium-chloride cotransport. Physiol
Rev 2000; 80:21176.
3. Clausen T. Regulation of active Na-K transport in skeletal
muscle. Physiol Rev 1986; 66:54280.
4. Clausen T, Flatman JA. The effect of catecholamines on Na-K
transport and membrane potential in rat soleus muscle.
J Physiol 1977; 270:383424.

13. Ethier JH, Kamel KS, Magner PO, Lemann JJ, Halperin ML.
The transtubular potassium concentration in patients with
hypokalemia and hyperkalemia. Am J Kidney Dis 1990;
15:30915.
14. Zalunardo N, Lemaire M, Davids MR, Halperin ML. Acidosis
in a patient with cholera: A need to redefine concepts.
Q J Med 2004; 97:68196.
15. Carlisle E, Donnelly S, Vasuvattakul S, Kamel K, Tobe S,
Halperin M. Glue-sniffing and distal renal tubular acidosis:
sticking to the facts. J Amer Soc Nephrol 1991; 1:101927.
16. Kamel KS, Briceno LF, Sanchez MI, Brenes L, Yorgin P,
Kooh SW, et al. A new classification for renal defects in
net acid excretion. Am J Kidney Dis 1997; 29:13646.
17. Goldstein M, Bear R, Richardson R, Marsden P, Halperin M.
The urine anion gap: a clinically useful index of ammonium
excretion. Am J Med Sci 1986; 292:198202.
18. Halperin ML, Margolis BL, Robinson LA, Halperin RM,
West ML, Bear RA. The urine osmolal gap: a clue to estimate
urine ammonium in hybrid types of metabolic acidosis.
Clin Invest Med 1988; 11:198202.
19. Carlisle EJF, Donnelly SM, Halperin ML. Renal tubular
acidosis (RTA): Recognize The Ammonium defect and pH
or get the urine pH. Pediatr Nephrol 1991; 5:2428.

6. Lin SH, Lin YF, Halperin ML. Hypokalemia and Paralysis:

Clues on admission to help in the differential diagnosis.
Q J Med 2001; 94:1339.

20. Kaitwatcharachai C, Vasuvattakul S, Yenchitsomanus

P, Thuwajit P, Malasit P, Chuawatana D, et al. Distal Renal
Tubular Acidosis in a Patient with Southeast Asian Ovalocytosis: Possible interpretations of a high urine PCO2. Am J
Kidney Dis 1999; 33:114752.

7. Cockcroft DW, Gault MH. Prediction of creatinine clearance

from serum creatinine. Nephron 1976; 16:3141.

21. Lin SH, Davids MR, Halperin ML. Hypokalemia and

paralysis. Q J Med 2003; 96:1619.

8. Huth EJ, Squires RD, Elkinton JR. Experimental potassium

depletion in normal human subjects. II. Renal and hormonal
factors in the development of extracellular alkalosis during
depletion. J Clin Invest 1959; 38:114965.

22. Hebert SC. Extracellular calcium-sensing receptor: Implications for calcium and magnesium handling in the kidney.
Kidney Int 1996; 50:212939.

9. Ring T, Knoers N, Oh MS, Halperin ML. Reevaluation of

the criteria for the clinical diagnosis of Gitelmans syndrome.
Ped Nephrol 2002; 17:61216.
10. Lin SH, Lin YF. Propranolol rapidly reverses paralysis,
hypokalemia and hypophosphatemia in thyrotoxic periodic
paralysis. Am J Kidney Dis 2001; 37:6204.
11. Carlisle EJF, Donnelly SM, Ethier J, Quaggin SE, Kaiser U,
Vasuvattakul S, et al. Modulation of the secretion of
potassium by accompanying anions in humans. Kidney Int
1991; 39:120612.

23. Schreiber MS, Chen C-B, Lessan-Pezeshki M, Halperin ML,

Schlanger LE, Patnaik A, et al. Antikaliuretic action of
trimethoprim is minimized by raising urine pH. Kidney Int
1996; 49:827.
24. Halperin ML, Kamel KS. Potassium. Lancet 1998; 352:
13542.
25. Porte DJ. Sympathetic regulation of insulin secretion. Arch
Intern Med 1969; 123:25260.
26. Cannon PJ, Heinemann HO, Albert MS, Laragh JH, Winters
RW. Contraction alkalosis after diuresis of edematous
patients with ethacrynic acid. Ann Int Med 1965; 62:97990.

Downloaded from by guest on September 27, 2015

5. Counillon LL, Pouyssegur RJ. The members of the Na/H

exchanger gene family: Their structure, function, expression,
and regulation. In: Seldin DW, Giebisch G, eds. The Kidney:
Physiology & Pathophysiology. Philadelphia PA, Lippincott
Williams & Wilkins, 2000:22334.

12. Steele A, deVeber H, Quaggin SE, Scheich A, Ethier J,

Halperin ML. What is responsible for the diurnal variation in
potassium excretion? Am J Physiol 1994; 36:R55460.