The Effect of Acidosis in Hypokalemic Periodic Maureen A.
Jarrell; Melvin Greer, MD; Thomas H. Maren,
\s=b\ Metabolic
acidosis was produced in patients with hypokalemic periodic paralysis by the administration of ammonium chloride over a period of three days. The challenging test of glucose and insulin produced a substantially smaller reduction of both serum potassium concentration and muscle strength than when the patients were tested in normal acid\x=req-\ two
base balance. The findings agree with earlier work on acetazolamide, suggesting that metabolic acidosis provides protection against episodes of muscle weakness in periodic
paralysis. (Arch Neurol 33:791-793, 1976) A cetazolamide has been effective in the treatment of the familial periodic paralyses. For the hypoka¬ lemic form of the disease, the mecha¬ nism appears to be protection against the fall in serum potassium level. Acetazolamide blunts the fall in serum potassium level precipitated by ad¬ ministration of glucose and insulin and reduces the induced paralysis.1 As -1 .
Accepted
From the
for
publication June 15, 1976. Departments of Pharmacology
and
Therapeutics (Miss Jarrell and Dr Maren) and Neurology (Dr Greer), University of Florida College of Medicine, Gainesville. Reprint requests to Department of Pharmacology, University of Florida College of Medicine, Gainesville, FL 32610 (Dr Maren).
Paralysis
MD
result, we postulated that the meta¬ bolic acidosis produced by acetazo¬ lamide was the major factor in achiev¬ ing the beneficial responses. This is in agreement with Viskoper et al,- who administered ammonium chloride in¬ travenously and reported an improved electromyogram in a patient with a
hypokalemic periodic paralysis. This report involves two previously studied patients' in whom ammonium
chloride was used to show the effect of acidosis in protecting against weak¬ ness produced by the administration of glucose and insulin. Our findings confirm the original theory that the production of metabolic acidosis is the mechanism of action of acetazolamide in the treatment of hypokalemic
periodic paralysis.
REPORT OF CASES The two brothers involved in this study were described in our earlier article' as cases 1 and 2. The same numbers are used here. General physical and laboratory findings were normal in both unless noted. Case l.-A 28-year-old man had taken acetazolamide intermittently since 1972. His weakness is generally mild. The drug decreased the duration, severity, and frequency of his episodes. He had not taken acetazolamide for several months prior to the present study. He was obese
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
and had a mildly elevated blood pressure. Case 2.—A 32-year-old man, the brother of patient 1, was placed on a regimen of acetazolamide in 1972 and experienced a decrease in number, duration, and severity of attacks. With treatment, he still had episodes of mild weakness approximately once a month. After a complete description of the experimental procedure, the patient volunteered to terminate acetazolamide therapy two weeks prior to admission. Shortly thereafter, he developed bother¬ some but not incapacitating periods of muscle weakness.
METHODS Blood samples and urine specimens were collected from the fasting patients in the mornings of days 1 and 2. On day 2, a fourhour glucose tolerance test was adminis¬ tered; serum potassium levels were moni¬ tored during this period. Following this procedure, the patients received 12 gm of ammonium chloride per day. Gelatin cap¬ sules containing 1 gm of ammonium chlo¬ ride were given in 2-gm doses at noon, 4 pm, and 8 pm, and in 3-gm doses at midnight and 8 am. The 4-am dose was omitted to avoid gastric upset. Patient 1 drank 100 ml of milk following each ammonium chloride administration. Venous blood pH, Pco,, Po,, and total CO., levels were determined
daily.
Both patients were challenged with glucose (1.75 gm/kg orally as Glucola) and insulin (20 units subcutaneously) on the third day of ammonium chloride adminis-
Acid-Base Values Following Oral Ammonium Chloride Administration to Two Patients Total NH CI
Hours After Initial Dose Case 1 Control
Pco,,
Given, gm
pH
26.5
15
52.5 68
29 37
7.44 7.30 7.30 7.22 7.23
15 29 37
7.45 7.32 7.34 7.28 7.26
Case 2 Control 26.5 52.5 68
Fig 1 .—Modification by ammonium chloride in patients with hypokalemic paralysis.
mm
Total CO., mM
Hg
26 46 38
23 19
41 28
12 28 20
37 32
of the
37
18 18
27
13
hypokalemic
effect of
glucose-insulin
tration. Venous blood samples were drawn every 30 minutes from a 19-gauge scalp vein needle, which was kept open by a slow, normal saline drip. Whole blood was analyzed for glucose, and serum for potas¬ sium, as in the previous study.1 Electrocar¬ diogram and muscle strength were moni¬ tored throughout the study.
RESULTS
Ammonium chloride treatment over
three-day period produced metabolic acidosis in both patients (Table). The patients suffered no gastric discom¬ fort but mentioned an occasional, mild, esophageal burning at the time of ammonium chloride ingestion. The glucose-insulin test produced no weakness in patient 1. Patient 2, who had the more severe disease, developed mild proximal weakness a
one
hour after the administration of
glucose-insulin. Right elbow flexor weakness was most prominent and
50-
CASE I 4.0
may have been related to the indwell¬
ing venous line. Three hours after the test, only mild right elbow flexor weakness remained; this returned to normal
about five hours. In after the glucoseinsulin test without other treatment, patient 1 had no weakness, but patient 2 showed severe muscle weakness, requiring oral administration of po¬ tassium.' The weakness appeared to be much milder following ammonium chloride pretreatment. The serum potassium response to glucose and insulin was blunted by ammonium chloride acidosis (Fig 1). In that figure, the control data are from our earlier study on the same patients (Table 2 of reference 1). Figure 2 compares present results with acidosis caused by acetazola¬ mide.1 In case 1, ammonium chloride had a greater effect than acetazola¬ mide; in case 2, less. Following ammo¬ nium chloride administration, the plasma bicarbonate concentration was much lower (13 mM) than in the aceta¬ zolamide experiment (22 mM). How¬ ever, a tissue acidosis of respiratory origin in the acetazolamide experi¬ ment is not reflected in the measured acid-base chemistry.' This is discussed below. Both patients were found to be diabetic during glucose tolerance test¬ ing. Following the administration of in
previous studies
3.0-
CONTROL 2.0
Ld
I
5.0
co
CASE 2 4.0-
N,
3.0
NH.CI
2.0-
CONTROL
30
Glucose +
60
90
120
150
180
210
240 270
MINUTES
Insulin
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
glucose alone,
serum
potassium
effect in
promoting bicarbonate loss. patients with hypokalemic periodic paralysis treated successfully with acetazolamide, blood pH declines mod¬ estly to about 7.35.14 But since the acidosis achieved by acetazolamide also has a respiratory component and a disequilibrium between HoCO,/CO, is created, the tissue Pco2 is actually greater than indicated by plasma
con¬
centrations decreased about 1 to 1.5 mM; this drop was less than that produced by glucose and insulin, usually 2 mM.
In
COMMENT The present data confirm our infer¬ that the mechanism of action of acetazolamide is through the induc¬ tion of a metabolic acidosis rather than a direct effect on the muscle.1 The metabolic acidosis engendered by acetazolamide is secondary to its renal ence
acid-base balance.3 The effects ob¬ served are an extension of known physiological effects in the normal: in the acidotic dog, the usual hypoka-
Fig 2.—Comparison of ammonium chloride and acetazolamide in protecting against hypokalemia following glucose-insulin in patients with hypokalemic paralysis. ACZ refers to earlier study with acetazolamide in same patients.' CASE I
,
-2.0 CONTROL
UJ
lemic effect of insulin is abolished.1 The effects of ammonium chloride have been shown to be similar to that of acetazolamide in blunting the effect of glucose-insulin on serum potassium concentrations by means of inducing an acidosis. The two agents reduce both the rate of potassium decrement and the total fall. Further studies are needed to show quantita¬ tive relations between plasma bicar¬ bonate levels and the effect of either acetazolamide or ammonium chloride in blocking the glucose-insulin effect. The data do suggest, however, that in the control of these patients, attention might be given to the level of plasma bicarbonate achieved by acetazola¬ mide. From these and other studies, it appears that, for maximum lowering of HCO, at least 250 mg every four to six hours should be given and that it may be wise to begin treatment with initial doses of 500 mg. We have provided a model of the quantitative factors involved in the flux of potassium from serum to muscle and have shown how this flux may be increased in this disease. This calculation shows that an attack of hypokalemic periodic paralysis that reduces serum potassium from 4 to 2 mM involves an increase in potassium flux from serum to muscle of only about 1%.' Acidosis reduces or abol¬ ishes this pathological change and protects the patient from attack.
-3.0
E
01 IO
or
o
References
-io-
< I
o
-2 0
CONTROL
3 0
I
Glucose +
30
60
90
120
150
180
210
240 270
MINUTES
Insulin
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
1. Vroom FQ, Jarrell MA, Maren TH: Acetazolamide treatment of hypokalemic periodic paralysis: Probable mechanism of action. Arch Neurol 32:385-392, 1975. 2. Viskoper RJ, Licht A, Fidel J, et al: Acetazolamide treatment in hypokalemic periodic paralysis: A metabolic and electromyographic study. Am J Med Sci 266:119-123, 1973. 3. Maren TH: Carbonic anhydrase: Chemistry, physiology and inhibition. Physiol Rev 47:595-781, 1967. 4. Resnick JS, Engel WK, Griggs RC, et al: Acetazolamide prophylaxis in hypokalemic paralysis. N Engl J Med 278:582-586, 1968. 5. Mackler B, Lichtenstein H, Guest GM: Effect of ammonium chloride acidosis on action of insulin in dogs. Am J Physiol 166:191-198, 1951.
Jarrell; Melvin Greer, MD; Thomas H. Maren,
\s=b\ Metabolic
acidosis was produced in patients with hypokalemic periodic paralysis by the administration of ammonium chloride over a period of three days. The challenging test of glucose and insulin produced a substantially smaller reduction of both serum potassium concentration and muscle strength than when the patients were tested in normal acid\x=req-\ two
base balance. The findings agree with earlier work on acetazolamide, suggesting that metabolic acidosis provides protection against episodes of muscle weakness in periodic
paralysis. (Arch Neurol 33:791-793, 1976) A cetazolamide has been effective in the treatment of the familial periodic paralyses. For the hypoka¬ lemic form of the disease, the mecha¬ nism appears to be protection against the fall in serum potassium level. Acetazolamide blunts the fall in serum potassium level precipitated by ad¬ ministration of glucose and insulin and reduces the induced paralysis.1 As -1 .
Accepted
From the
for
publication June 15, 1976. Departments of Pharmacology
and
Therapeutics (Miss Jarrell and Dr Maren) and Neurology (Dr Greer), University of Florida College of Medicine, Gainesville. Reprint requests to Department of Pharmacology, University of Florida College of Medicine, Gainesville, FL 32610 (Dr Maren).
Paralysis
MD
result, we postulated that the meta¬ bolic acidosis produced by acetazo¬ lamide was the major factor in achiev¬ ing the beneficial responses. This is in agreement with Viskoper et al,- who administered ammonium chloride in¬ travenously and reported an improved electromyogram in a patient with a
hypokalemic periodic paralysis. This report involves two previously studied patients' in whom ammonium
chloride was used to show the effect of acidosis in protecting against weak¬ ness produced by the administration of glucose and insulin. Our findings confirm the original theory that the production of metabolic acidosis is the mechanism of action of acetazolamide in the treatment of hypokalemic
periodic paralysis.
REPORT OF CASES The two brothers involved in this study were described in our earlier article' as cases 1 and 2. The same numbers are used here. General physical and laboratory findings were normal in both unless noted. Case l.-A 28-year-old man had taken acetazolamide intermittently since 1972. His weakness is generally mild. The drug decreased the duration, severity, and frequency of his episodes. He had not taken acetazolamide for several months prior to the present study. He was obese
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
and had a mildly elevated blood pressure. Case 2.—A 32-year-old man, the brother of patient 1, was placed on a regimen of acetazolamide in 1972 and experienced a decrease in number, duration, and severity of attacks. With treatment, he still had episodes of mild weakness approximately once a month. After a complete description of the experimental procedure, the patient volunteered to terminate acetazolamide therapy two weeks prior to admission. Shortly thereafter, he developed bother¬ some but not incapacitating periods of muscle weakness.
METHODS Blood samples and urine specimens were collected from the fasting patients in the mornings of days 1 and 2. On day 2, a fourhour glucose tolerance test was adminis¬ tered; serum potassium levels were moni¬ tored during this period. Following this procedure, the patients received 12 gm of ammonium chloride per day. Gelatin cap¬ sules containing 1 gm of ammonium chlo¬ ride were given in 2-gm doses at noon, 4 pm, and 8 pm, and in 3-gm doses at midnight and 8 am. The 4-am dose was omitted to avoid gastric upset. Patient 1 drank 100 ml of milk following each ammonium chloride administration. Venous blood pH, Pco,, Po,, and total CO., levels were determined
daily.
Both patients were challenged with glucose (1.75 gm/kg orally as Glucola) and insulin (20 units subcutaneously) on the third day of ammonium chloride adminis-
Acid-Base Values Following Oral Ammonium Chloride Administration to Two Patients Total NH CI
Hours After Initial Dose Case 1 Control
Pco,,
Given, gm
pH
26.5
15
52.5 68
29 37
7.44 7.30 7.30 7.22 7.23
15 29 37
7.45 7.32 7.34 7.28 7.26
Case 2 Control 26.5 52.5 68
Fig 1 .—Modification by ammonium chloride in patients with hypokalemic paralysis.
mm
Total CO., mM
Hg
26 46 38
23 19
41 28
12 28 20
37 32
of the
37
18 18
27
13
hypokalemic
effect of
glucose-insulin
tration. Venous blood samples were drawn every 30 minutes from a 19-gauge scalp vein needle, which was kept open by a slow, normal saline drip. Whole blood was analyzed for glucose, and serum for potas¬ sium, as in the previous study.1 Electrocar¬ diogram and muscle strength were moni¬ tored throughout the study.
RESULTS
Ammonium chloride treatment over
three-day period produced metabolic acidosis in both patients (Table). The patients suffered no gastric discom¬ fort but mentioned an occasional, mild, esophageal burning at the time of ammonium chloride ingestion. The glucose-insulin test produced no weakness in patient 1. Patient 2, who had the more severe disease, developed mild proximal weakness a
one
hour after the administration of
glucose-insulin. Right elbow flexor weakness was most prominent and
50-
CASE I 4.0
may have been related to the indwell¬
ing venous line. Three hours after the test, only mild right elbow flexor weakness remained; this returned to normal
about five hours. In after the glucoseinsulin test without other treatment, patient 1 had no weakness, but patient 2 showed severe muscle weakness, requiring oral administration of po¬ tassium.' The weakness appeared to be much milder following ammonium chloride pretreatment. The serum potassium response to glucose and insulin was blunted by ammonium chloride acidosis (Fig 1). In that figure, the control data are from our earlier study on the same patients (Table 2 of reference 1). Figure 2 compares present results with acidosis caused by acetazola¬ mide.1 In case 1, ammonium chloride had a greater effect than acetazola¬ mide; in case 2, less. Following ammo¬ nium chloride administration, the plasma bicarbonate concentration was much lower (13 mM) than in the aceta¬ zolamide experiment (22 mM). How¬ ever, a tissue acidosis of respiratory origin in the acetazolamide experi¬ ment is not reflected in the measured acid-base chemistry.' This is discussed below. Both patients were found to be diabetic during glucose tolerance test¬ ing. Following the administration of in
previous studies
3.0-
CONTROL 2.0
Ld
I
5.0
co
CASE 2 4.0-
N,
3.0
NH.CI
2.0-
CONTROL
30
Glucose +
60
90
120
150
180
210
240 270
MINUTES
Insulin
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
glucose alone,
serum
potassium
effect in
promoting bicarbonate loss. patients with hypokalemic periodic paralysis treated successfully with acetazolamide, blood pH declines mod¬ estly to about 7.35.14 But since the acidosis achieved by acetazolamide also has a respiratory component and a disequilibrium between HoCO,/CO, is created, the tissue Pco2 is actually greater than indicated by plasma
con¬
centrations decreased about 1 to 1.5 mM; this drop was less than that produced by glucose and insulin, usually 2 mM.
In
COMMENT The present data confirm our infer¬ that the mechanism of action of acetazolamide is through the induc¬ tion of a metabolic acidosis rather than a direct effect on the muscle.1 The metabolic acidosis engendered by acetazolamide is secondary to its renal ence
acid-base balance.3 The effects ob¬ served are an extension of known physiological effects in the normal: in the acidotic dog, the usual hypoka-
Fig 2.—Comparison of ammonium chloride and acetazolamide in protecting against hypokalemia following glucose-insulin in patients with hypokalemic paralysis. ACZ refers to earlier study with acetazolamide in same patients.' CASE I
,
-2.0 CONTROL
UJ
lemic effect of insulin is abolished.1 The effects of ammonium chloride have been shown to be similar to that of acetazolamide in blunting the effect of glucose-insulin on serum potassium concentrations by means of inducing an acidosis. The two agents reduce both the rate of potassium decrement and the total fall. Further studies are needed to show quantita¬ tive relations between plasma bicar¬ bonate levels and the effect of either acetazolamide or ammonium chloride in blocking the glucose-insulin effect. The data do suggest, however, that in the control of these patients, attention might be given to the level of plasma bicarbonate achieved by acetazola¬ mide. From these and other studies, it appears that, for maximum lowering of HCO, at least 250 mg every four to six hours should be given and that it may be wise to begin treatment with initial doses of 500 mg. We have provided a model of the quantitative factors involved in the flux of potassium from serum to muscle and have shown how this flux may be increased in this disease. This calculation shows that an attack of hypokalemic periodic paralysis that reduces serum potassium from 4 to 2 mM involves an increase in potassium flux from serum to muscle of only about 1%.' Acidosis reduces or abol¬ ishes this pathological change and protects the patient from attack.
-3.0
E
01 IO
or
o
References
-io-
< I
o
-2 0
CONTROL
3 0
I
Glucose +
30
60
90
120
150
180
210
240 270
MINUTES
Insulin
Downloaded From: http://archneur.jamanetwork.com/ by a Oakland University User on 06/09/2015
1. Vroom FQ, Jarrell MA, Maren TH: Acetazolamide treatment of hypokalemic periodic paralysis: Probable mechanism of action. Arch Neurol 32:385-392, 1975. 2. Viskoper RJ, Licht A, Fidel J, et al: Acetazolamide treatment in hypokalemic periodic paralysis: A metabolic and electromyographic study. Am J Med Sci 266:119-123, 1973. 3. Maren TH: Carbonic anhydrase: Chemistry, physiology and inhibition. Physiol Rev 47:595-781, 1967. 4. Resnick JS, Engel WK, Griggs RC, et al: Acetazolamide prophylaxis in hypokalemic paralysis. N Engl J Med 278:582-586, 1968. 5. Mackler B, Lichtenstein H, Guest GM: Effect of ammonium chloride acidosis on action of insulin in dogs. Am J Physiol 166:191-198, 1951.
