I !
Prior Treatment with Diuretic Augments the Hypokalemic and
Electrocardiographic Effects of Inhaled Albuterol
BRIANJ. LIPWORTH, M.R.c.P., DENISG. MCDEVITT, D.SC., M.D.,F.R.c.P., ALLAN D. STRUTHERS,M.D.,
M.R.C.P. Dundee,Scotland
PURPOSE: High doses of inhaled albuterol produce
substantial improvements in bronchodilatation
are associated with dose-related systemic side ef-
but
fects including hypohalemia and hypomagnesemia.
Concomitant diuretic therapy also produces these
metabolic sequelae and may therefore precipitate
cardiac arrhythmias in patients taking this combi-
nation of drugs. The purpose of this study was to
investigate the electrocardiographic (ECG) effects
of high-dose inhaled albuterol, and to evaluate
whether potentiation occurs with bendrofluazide.
PATIENTSANDMETHODS: Tennormalsubjects
(mean age [f SEMI: 29 f 2 years, four women, six
men) received seven days of treatment with either
bendrofluazide 5 mg or identical placebo in a sin-
gle-blind, randomized, cross-over design, with a lo-
day washout period. After each treatment period,
responses (potassium, magnesium, ECG) to cumu-
lative doubling doses of inhaled albuterol (100 pg to
2,000 pg) were measured.
RESULTS: Baseline potassium levels (mean and
95% confidence intervals) were lower after pre-
treatment with bendrofluazide compared with pla-
cebo (3.07 mmol/L [2.89 to 3.25 mmol/L] versus 3.78
mmol/L [3.62 to 3.93 mmol/L]; p <O.OOl). The com-
bination of bendrofluazide and albuterol produced
a lower absolute level of potassium than did place-
bo and albuteroh (2.72 mmol/L [2.50 to 2.95 mmol/
L] versus 3.18 mmol/L [3.09 to 3.27 mmol/L]; p
<O.OOl). Mean (& SEM) potassium fell to a lower
level with bendrofluazide and albuterol in women
than in men (2.45 f 0.04 mmol/L versus 2.90 f 0.13
mmol/L; p <0.005). Albuterol alone produced a
small but significant fall in magnesium (0.842
mmol/L [0.815 to 0.869 mmol/L] to 0.789 mmol/L
[0.757 to 0.820 mmol/L]; p <O.OOl), but no further
change after bendrofluazide. Pretreatment with
bendrofluazide increased the frequency (p <O.OOl)
and amplitude (p <0.05) of U waves due to albu-
terol. Albuterol also attenuated T-wave amplitude
(p <O.OOl) and prolonged the Q-Tc interval (p
<O.OOl), with no additive effect from bendroflua-
zide. ST-segment depression (p <O.OOl) occurred in
five subjects who inhaled albuterol.
CONCLUSION: These findings show that treatment
with bendrofluazide augments the hypohalemic
and ECG effects of high-dose inhaled albuterol.
The arrhythmogenic potential of this interaction
may be important in patients with acute exacerba-
tions of chronic airflow obstruction, who have con-
comitant hypoxemia and ischemic heart disease.
H igher than conventional doses (200 pg) of inhaled
albuterol dispensed by metered dose inhaler may
produce marked improvements in airways response in
patients with asthma [l]. However, these changes are
associated with dose-dependent side effects, including
hypokalemia and hypomagnesemia [2]. These electro-
lyte changes are thought to be important precipitating
factors in the production of ventricular arrhythmias.
There has been a recent increase in the use of home
nebulizers to deliver high doses of inhaled beta-ago-
nists. Delays in seeking medical help and overreliance
on high-dose inhaled beta-agonists occurred in up to
30% of cases of sudden death in asthmatic patients
using home nebulizers [3]. The development of cardiac
arrhythmias was postulated as a possible cause for
these deaths. Higgins and co-workers [4] found that
supraventricular and ventricular arrhythmias oc-
curred in five of 19 elderly patients with chronic air-
flow obstruction following nebulized beta-agonist
therapy and that these arrhythmias were independent
of blood oxygen saturation. Many patients with air-
flow obstruction are receiving long-term therapy with
diuretics for car pulmonale or coexisting hypertension.
Diuretics also produce hypokalemia and hypomagne-
semia, and an interaction may possibly result if high-
dose inhaled beta-agonists are administered concur-
rently.
The purposes of the current study were twofold.
First, we have investigated in detail the ECG effects of
inhaled albuterol. Second, we have assessed whether
prior treatment with bendrofluazide augments the hy-
pokalemic and ECG effects of inhaled albuterol in
normal subjects.
PATIENTS AND METHODS
Patients
After approval of the local ethics committee, written
informed consent was obtained from all subjects. Ten
healthy subjects (four women, six men) were studied
with a mean (ZIZSEM) age of 29 f 2 years (range: 21 to
40 years). A 12-lead electrocardiogram showed no ab-
normalities, and levels of blood electrolytes and creati-
nine were normal.
Study Design
Subjects were given seven days of treatment with
bendrofluazide 5 mg or identical placebo in a single-
blind, randomized, cross-over design. A washout peri-
od of 10 days was used between treatments with ben-
drofluazide and placebo. After each treatment period,
subjects visited the laboratory between 9 A.M. and 10
A.M. for a dose-response study. Subjects were supine
throughout the 130-minute study period. A cannula
was inserted into the antecubital vein and kept patent
with bolus injections of heparinized saline. A 30-min-
ute run-in period was used to establish a true baseline
with which to compare successive responses. Blood
June 1989 The AmericanJournalof Medicine Volume 86 653
 INHALED ALBUTEROL, DIURETICS, AND ECG CHANGES
I( (mmoi/l)
+ &I
0 100 200 500
Dose Salbutamol @g)
( (mmolll)
3.0
2.5
Dose Salbutamol (pg)
Figure 1. Potassium responses (mean and SEM) to cumulative
of inhaled albuterol (Salbutamol) in 10 normal subjects (top)following
pretreatment with either bendrofluazide (A) or placebo (a), (bot-
tom) in four women (- - - -) and six men (-).
pressure was measured at one-minute intervals, and
the mean of the eight lowest consistent readings was
taken as a baseline. Two blood samples were taken at
&minute intervals during the run-in period to assay
plasma potassium and magnesium level, and the mean
was used as a baseline.
Cumulative ‘doubling doses of inhaled albuterol
were then given from a 750-mL pear-shaped spacer
device (Nebuhaler, Astra Pharmaceuticals, King’s
Langley, Herts, United Kingdom), using the method
of Gleeson and Price [5] so as to eliminate differences
in inhaler technique. Dose increments were made ev-
ery 20 minutes as follows: 100 pg, 200 pg, 500 pg, 1,000
/*g, and 2,000 pg. Specially prepared metered-dose
aerosol canisters were used delivering either 100 pg or
500 pg per actuation (Glaxo Group Research Ltd,
Greenford, Middlesex, United Kingdom). Fifteen
minutes after each dose increment, blood pressure
measurements were made at one-minute intervals,
taking the mean of four consistent readings. Blood
samples were taken from the intravenous cannula af-
ter withdrawal of 2-mL dead space. The blood was
654 June 1989 The American Journal
/ LIPWORTH ET AL
immediately centrifuged at 3,500 rpm, separated, and
stored at -2OOC. Plasma potassium was measured at
each dose increment, and plasma magnesium only af-
ter the last dose of albuterol. An ECG trace was re-
corded on standard lead I for a period of 10 beats when
the heart rate (on the monitor) had settled to its lowest
reading. The following parameters were measured
from the mean of five beats: R-R interval (seconds), P-
R interval (seconds), Q-T interval (seconds), T-wave
amplitude (mV), U-wave amplitude (mV), and ST seg-
ment depression (mV). The Q-T interval was mea-
sured using the method described by Shamroth [6] to
account for the presence of U waves. The Bazett for-
mula [7] was used to correct the Q-T interval for heart
rate (Q-Tc). Heart rate was calculated from the R-R
interval.
1000 2000
Measurements
Systolic and diastolic blood pressures were recorded
with a semi-automated sphygmomanometer (Dina-
map Vital Signs Monitor, Critikon, Tampa, Florida)
with the subject in a supine position. The electrocar-
diogram was recorded with a Hewlett Packard (Palo
Alto, California) monitor and printer set at 50 mm/
second paper speed and 0.5 mV/cm gain. All biochemi-
cal analyses were performed in batches at the end of
the study and assayed in duplicate. Plasma potassium
was measured by Flame Photometry (IL943 analyzer;
Instrumentation Laboratory Ltd, Warrington, United
Kingdom). The within and between values for analyti-
cal imprecision (CVw and CVB) were 0.7% and 1.8%,
respectively. Plasma magnesium was analyzed by
spectrophotometry using a Cobas Bio (Roche, Belle-
ville, New Jersey) (CVw 1.3%, CVB 2.5%). Normal ref-
erence ranges for our laboratory are as follows: potassi-
um 3.5 to 5.0 mmol/L and magnesium 0.70 to 1.15
mmol/L.
Analyses
1 All analyses were performed on the raw data using
doses an “SPSS plus” software package (SPSS Inc., Chica-
go, Illinois). Repeated measures analysis of variance
(MANOVA) [8] was used to compare dose-response
curves in order to assess the effects of: (1) pretreat-
ment with bendrofluazide or placebo, (2) dose of albu-
terol, (3) sex of subject, and (4) interactions between
(l), (2), and (3). Where MANOVA showed significant
effects of pretreatment or sex, paired or unpaired Stu-
dent’s t-tests were applied, respectively, to compare
baseline values. A probability value of less than 5%
(two-tailed) was considered significant for all tests.
RESULTS
The results are presented either as means and stan-
dard errors (SEM) or as means and 95% confidence
intervals (95% CI). Baseline potassium was lower after
pretreatment with bendrofluazide than after pretreat-
ment with placebo (3.07 mmol/L [2.89 to 3.25 mmol/L]
versus 3.78 mmol/L [3.62 to 3.93 mmol/L]; p <O.OOl).
There were no significant sex differences in baseline
potassium levels after treatment with placebo or ben-
drofluazide. Dose-response curves for plasma potassi-
um levels are shown in Figure 1. The combination of
bendrofluazide and albuterol produced a lower abso-
lute level of potassium than placebo and albuterol
(2.72 mmol/L [2.50 to 2.95 mmol/L] versus 3.18 mmol/
L [3.09 to 3.27 mmol/L]; p <O.OOl). Plasma potassium
of Medicine Volume 86
 INHALED ALBUTEROL, DIURETICS, AND EGG CHANGES / LIPWORTH ET AL

0.16 -

0.14 -

0.12 -
0.10 -
u wave
amplitude 0.08 -
(mv) 0.06 -
0.04 -

0.02 -
0 -
2000
8 7

7 -
6 -
u wave 5 -
(Ircquency)
4

dzhaKl!
100
3 -
2

Figure 2. Cumulative frequency of U I -

waves and amplitude (mV) at each dose 0 -
0 200 500 1000 2000
of albuterol (Salbutamol) in 10 normal
subjects following pretreatment with Dose Salbufamol (pg)
bendrofluazide (B) or placebo (P). I

fell to a lower level after treatment with bendroflua-

zide in women than in men (2.45 f 0.04 mmol/L versus 0.7

2.90 f 0.13 mmol/L; p <0.005). The baseline plasma 0.6 I

magnesium level was lower after bendrofluazide than
after placebo (0.808 mmol/L [0.758 to 0.858 mmol/L] 0.5-

versus 0.842 mmol/L [0.815 to 0.869 mmol/L]), al- 0.4-

though this did not achieve statistical significance. Al- r wave (mV)
buterol alone produced a significant decline in magne- 0.3.

sium (0.842 mmol/L [0.815 to 0.869 mmol/L] to 0.789 0.2-

mmol/L [0.757 to 0.820 mmol/L], p <O.OOl), with a
further non-significant drop after pretreatment with 0. I
bendrofluazide (0.808 mmol/L [0.758 to 0.858 mmol/ 0 6 ,, ,
L] to 0.765 mmol/L [0.722 to 0.808 mmol/L]). 100 200 500 IO00 2000
There was a significant increase in the frequency Dose Salbutamol (pg)
and amplitude of U waves with albuterol (p <O.OOl)
(Figure 2). There was an additional increase in both
frequency (p <O.OOl) and amplitude (p <0.05) with
bendrofluazide and albuterol (Figure 2). Albuterol
caused attenuation of T-wave amplitude (Figure 3,
top) (0.47 f 0.08 mV to 0.22 f 0.05 mV; p <O.OOl),
with no further change after bendrofluazide. T-wave
amplitude began at and fell to a lower absolute level in
women than in men (0.325 f 0.06 mV to 0.15 f 0.09
mV versus 0.57 f 0.12 mV to 0.27 f 0.05 mV; p <0.05)
(Figure 3, bottom). There was no significant interac-
tion between the effect of pretreatment with bendro- 0’41 a
fluazide and sex on albuterol-induced T-wave flatten- 0 100 200 500 1000 2000
ing. Some examples of the effects of albuterol on T Dose Salbutamoi (pg)
L
waves and U waves and potentiation by bendroflua-
Figure 3. T-wave amplitude (mV) (mean and SEM) following cumula-
zide are shown in Figure 4. The Q-Tc interval was tive doses of inhaled albuterol (Salbutamol) (top) after treatment
prolonged with albuterol (Figure 5), from 0.38 f 0.005 with bendrofluazide (A) or placebo (*), (bottom) in women (- - - -)
second to 0.41 f 0.009 second (p <O.OOl). There was, and men (-).
however, no effect of bendrofluazide or sex on Q-Tc
prolongation due to albuterol. ST segment depression
occurred with albuterol (p <O.OOl) in five subjects, <O.OOl). The ECG changes were not, however, associ-
with a mean decline of 0.06 mV, and was not potentiat- ated with ventricular extrasystoles or arrhLythmias.
ed by bendrofluazide. There were no ST segment The hemodynamic responses are shown in Table I.
changes in the other five subjects. The P-R interval Prior treatment with bendrofluazide had no effect on
remained unchanged. There were significant correla- baseline heart rate, systolic blood pressure, or diastolic
tions between the level of plasma potassium and T- blood pressure. There were significant chronotropic
wave amplitude (r = 0.32, p <0.05) and between plas- responses to albuterol (p <O.OOl), with attenuation in
ma potassium and Q-Tc prolongation (r = -0.62, p men compared with women(65 f 4 beats/minute to 79

June 1989 The American Journal of Medicine Volume 86 655

 INHALED ALBUTEROL, DIURETICS, AND ECG CHANGES / LIPWORTH ET AL

Baseline
effects therefore occurred at doses that are higher than
After Salbutamol
those used with a conventional pressurized aerosol
(200 pg), but lower than those currently recommended
for nebulizer therapy (2.5 to 5.0 mg).
Changes in plasma magnesium levels with albuterol
were small, although plasma magnesium is a poor re-
flection of intracellular levels [14,15]. The hypomag-
b nesemic response to beta-agonists is largely due to
increased urinary excretion of magnesium [ll]. Evi-
dence suggests that beta-agonists may inhibit reab-
sorption of magnesium from the thick ascending limb
of the loop of Henle by a cyclic AMP-mediated mecha-
2a nism [16]. The unchanged baseline level of plasma
magnesium after treatment with a diuretic may be
explained by the relatively short period of treatment
with bendrofluazide in the current study.
Diuretics can produce ventricular extrasystoles
(VES) [17,18]. In the presence of ischemic heart dis-
b ease or heart failure, ventricular extrasystoles are an
important predictor of mortality [19,20]. Hypokale-
mia is a risk factor for ventricular arrhythmias in pa-
tients with acute myocardial infarction [21]. Infusion
of magnesium in diuretic-treated patients increases
3a muscle potassium content and decreases the frequen-
cy of ventricular extrasystoles, whereas potassium in-
fusion has no effect [14]. Oral supplementation with
potassium or magnesium alone in diuretic-treated pa-
tients has no effect on hypokalemia or ventricular ex-
trasystoles, although there is a marked reduction in
ventricular extrasystoles and hypokalemia when both
are given in combination [18]. Patients with chronic
obstructive pulmonary disease are usually cigarette
smokers and therefore commonly have coincidental
ischemic heart disease or indeed car pulmonale due to
Figure 4. Examples of single lead (I) electrocardiograms (50 mm/ the airway disease itself. They may therefore common-
second, 0.5 mV/cm) before and after 2,000 rg inhaled albuterol ly be taking high doses of inhaled beta-agonists along
(Salbutamol), following pretreatment with (top) placebo or (bottom) with diuretics for coexisting heart failure or hyperten-
bendrofluazide in three normal patients. Salbutamol alone produced sion. During acute exacerbations of airway obstruc-
decreased T-wave amplitude (Patients 1,2, and 3), U waves (Patients tion, the hypokalemia may be further increased by
1 and 2), and ST segment depression (Patients 1 and 2). Prior treat-
ment with bendrofluazide potentiated the effects of Salbutamol in all
steroid or theophylline therapy [22], along with the
three subjects. effects of high adrenergic drive [12]. Sensitization of
the myocardium by hypoxemia may also potentiate
the arrhythmogenic effects of hypokalemia.
f 5 beats/minute versus 60 f 3 beats minute to 85 f 5 It was of interest that albuterol-induced hypokale-
beats/minute; p <0.05). Albuterol produced a decrease mic and chronotropic responses were greater in wom-
in diastolic blood pressure (p <O.OOl) and a rise in en. Females have previously been shown to have a
systolic blood pressure (after 4,000 pg, p <O.OOl).
COMMENTS
The results of the current study show that higher
than conventional doses of inhaled albuterol produce
I 0.42-
I
marked ECG changes, including U waves, T-wave flat-
tening, Q-Tc interval prolongation, and, in some sub- 0.41 -
jects, ST segment depression. The ECG effects were
consistent with changes due to hypokalemia and hypo- 0.40 -

magnesemia [9,10]. Hypocalcemia prolongs the Q-T QTc(s’ 0.39: &

interval [6], although changes in serum calcium that
occur with beta-agonists are small [ll] and were there- 0.36
fore not measured in the current study. Alternatively,
0.37 -
the ECG changes with albuterol may in part be a result
of a direct sympathomimetic action on cardiac beta- 0.36 J rr I , I
adrenoceptors [12,13]. Prior treatment with bendro- 0 100 200 500 1000 2000
fluazide for seven days potentiated the hypokalemic Dose Salbuiamol (I-q)
and ECG effects of inhaled albuterol. Systemic re-
sponses to albuterol occurred in a dose-dependent Figure 5. Heart rate-corrected Q-Tc interval (Q-Tc) (mean and SEM)
fashion after 500 fig, although a plateau was not at- in response to albuterol (Salbutamol) following bendrofluazide (A) or
tained within our dose range (up to 2,000 pg). These placebo (0).

656 June 1989 The American Journal of Medicine Volume 86

 INHALED ALBUTEROL, DIURETICS, AND ECG CHANGES / LIPWORTH ET AL

TABLE I
Hemodynamic Responses to Inhaled Albuterol*

Baseline lfJ0 !-a 200Pfi 500 /.a 1,000 rg 2,000 pg

Heart rate (beats/minute) 63i2 63f 2 64&Z 7oi2 74f2 81zt3

Placebo 67 f 3 67f3 68 rt 3 70f3 71f4 79f 4
Bendrofluazide 5 mg
Sy;;$bb:od pressure (mm Hg) 112f2 112f3 lllf2 113f3 113f3 116f2
109f2 107 f 3 108f 2 107f2 109f3 117f4
Bendrofluazide 5 mg
Diasslic~lood pressure (mm Hg) 59 f 3 581t3 57 f 2 55f 2 56f 2 53 f 2
59f2 5812 58 f 2 56f 2
Bendrofluazide 5 mg
, 55*2 56f2
* Heart rate, systolic blood pressure, and diastolic blood pressure responses to cumulative doses of inhaled albuterol(100 pg to 2,000 rg), following pretreatment
with placebo or bendrofluazide 5 mg. Responses are shown as means (?&EM).

higher sensitivity to isoproteronol-induced tachycar-
dia than males [23]. This difference may simply reflect
lower body weight in females (60 f 3 kg versus 76 f 4
kg), and hence a proportionately greater dose effect
for weight. A difference in heart size may also explain
the lower T-wave voltages in females. Albuterol-in-
duced hypokalemia is mediated by betas-adrenocep-
tor-linked membrane-bound sodium/potassium
ATPase, resulting in intracellular influx of potassium
into muscle cells [24,25]. The increased hypokalemic
response in females may therefore also be explained by
a smaller muscle bulk, and consequently a reduced
homeostatic capacity to conserve potassium, when
stressed by beta-adrenergic stimulation.
In conclusion, these results show that high doses of
inhaled albuterol (over 500 pg) may produce marked
hypokalemic and ECG changes, which are augmented
by pretreatment with bendrofluazide. Further studies
are now indicated to examine the arrhythmogenic po-
tential of this interaction in patients during the acute
phase of obstructive pulmonary disease.
REFERENCES
1. Lipworth BJ, Clark RA. Dhillon DP, Brown RA. McDevitt DG: Beta-adrenoceptor
responses to high doses of inhaled salbutamol in patients with bronchial asthma.
Br J Clin Pharmacol 1988; 26: 527-533.
2. Lipworth BJ. Clark RA, Dhillon DP. Fraser CG, McDevitt DG: Metabolic effects of
cumulative doses of salbutamol from a metered-dose inhaler in bronchial asthma.
Br J Clin Pharmacol 1988; 25: 667P-668P.
3. Sears MR, Rea HH, Fenwick J, et a/: 75 deaths in asthmatics prescribed home
nebulisers. Br Med J 1987; 294: 477-480.
4. Higgins RM, Cookson WOCM, Lane DG, John SM, McCarthy GL, McCarthy ST:
Cardiac arrhythmias caused by nebulised beta-agonist therapy. Lancet 1987; II:
863-864.
5. Gleeson JGA, Price JF: Nebuhaler technique. Br J Dis Chest 1988; 82: 172-174.
6. Shamroth L: The Q-T interval. In: An introduction to electrocardiography. Ox-
ford: Blackwell Scientific Publications, 1982; 141-144.
7. Bazett HC: An analysis of the time relations of electrocardiograms. Heart 1920;
7: 353-370.
8. Nie NH, Hull CH, Jenkins JG, Steinbrumer K, Bent DH: SPSS: statistical package
for the Social Sciences, 2nd ed. New York: McGraw-Hill. 1975.
9. Weaver WF. Burchel H: Serum potassium and the electrocardiograph and hypo-
kalemia. Circulation 1960: 21: 505-521.
10. Eisenberg MJ: Magnesium deficrency and cardiac arrhythmias. NY State J Med
1986: 861X-135.
11. Bos WJW, Postma DS, Van Doormeal JJ: Magnesiuric and calc:iuric effects of
tenbutaline in man. Clin Sci 1988: 74: 595-597.
12. Struthers AD, Reid JL, Whitesmith R. Rodger JC: Effect of intravenous adrena-
line on electrocardiogram, blood pressure and serum potassium. Br Heart J 1983;
49: 90-93.
13. Taggart P, Donaldson R, Green J. eta/: Interrelation of heart rate and autonom-
ic activity in asymptomatic men with unobstructed coronary arteries. Studies with
atrial pacing, adrenaline infusion, and automatic blockade. Br Heart J 1982; 47:
19-25.
14. Dyckner T, Wester PO: Ventricular extrasystoles and intracellular electrolytes
before and after potassium and magnesium infusions in patients orr diuretic treat-
ment. Am Heart J 1979; 97: 12-17.
15. Levine BS. Coburn JW: Magnesium, the mimic/antagonist of calcium. N Engl J
Med 1984; 310: 1253-1255.
16. Quamme GA, Dirks JH: The physiology of renal magnesium handlrng. Renal
Physiol 1986; 9: 270-278.
17. Holland OB, Nixon JV, Kuhnert L: Diuretic-induced ventricular ectopic activity.
Am J Med 1981; 70: 762-768.
18. Hollifieid JW: Potassium and magnesium abnormalrties, diuretrc:s and arrhyth-
mias in hypertension. Am J Med 1984; 77: 28-32.
19. Coronary Drug Project Research Group: Prognostic importance of premature
beats following myocardial infarction: experience in the Coronary Drug Project.
JAMA 1973; 223: 1116-1124.
20. Dargie HJ, Cleland JGF. Leckie BJ, lnglis CG, East BW, Ford I: Relation of
arrhythmias and electrolyte abnormalities to survival in patients with severe chron-
ic heart failure. Circulation 1987; 75 (suppl IV): IV-98-IV-107.
21. Nordrehaug JE, Johannessen KA, Von der Lippe G: Serum potassium concen-
tration as a risk factor of ventricular arrhythmias early in acute myocardial infarc-
tion. Circulation 1985; 71: 645-649.
22. Whyte KF, Reid C, Addis GJ, Whitesmith R. Reid JL: Salbutamol induced hypo-
kolaemia: the effect of theophylline alone and in combination with adrenaline. Br J
Clin Pharmacol 1988; 25: 571-578.
23. Johansson SR, Hjalmarson A: Age and sex differences in cardiovascular reac-
tivity to adrenergic agonists, mental stress and isometric exercise in normal sub-
jects. Stand J Clin Lab Invest 1988; 48: 183-191.
24. Struthers AD, Reid JL, Whitesmith R, Rodger JC: The effects of c:ardioselective
and non-selective beta-adrenoceptor blockade on the hypokalaermic and cardio-
vascular responses to adrenomedullary hormones in man. Clin Sci 1983; 65: 143-
147.
25. Whyte KF. Addis GJ, Whitesmith R, Reid JL: The mechanism of dalbutamol
induced hypokalaemia. Br J Clin Pharmacol 1987; 23: 65-71.

June 1989 The American Journal of Medicine Volume 86 657