Clinical Science (1991)81,59-63
59
Effect of a single test dose of lithium carbonate on sodium and
potassium excretion in man
D. G. SHIRLEY, D. R. J. SINGER*, G. A. SAGNELM*, M. G. BUCKLEF, M. A. MILLER*,
N. D. MARKANDU" AND G. A. MAcGREGOR*
Department of Physiology, Charing Cross and Westminster Medical School, London, and *Blood Pressure Unit, Department of
Medicine, St George's Hospital Medical School, London
(Received 19 September 1990/8 January 1991; accepted 18January 1991)
SUMMARY
Key words: lithium, plasma renin activity, potassium
excretion, sodium excretion.
1. The possible natriuretic and kaliuretic effects of a
single dose of lithium, as used in lithium clearance studies,
were investigated in 15 healthy subjects on fixed sodium
(100 mmo1/24 h) and potassium (70 mmo1/24 h) intakes.
Lithium carbonate (300 mg or 600 mg) or placebo tablets
were administered. double-blind and in random order,
midway through a 48 h urine collection (divided into six
8 h periods), at 23.00 hours.
2. During the three 24 h periods which preceded the
administration of lithium or placebo (control days), rates
of sodium and potassium excretion followed normal
circadian patterns, but no differences in excretion rates
between the 3 control days were observed. Placebo
tablets did not affect excretion rates.
3. After the 300 mg dose of lithium carbonate, 24 h
sodium excretion increased by approximately 17 mmol
( P < 0.05); almost all of the natriuretic effect occurred
during the first two 8 h periods. No effect on potassium
excretion was observed.
4. After the 600 mg dose of lithium carbonate, 24 h
sodium excretion increased by approximately 48 mmol
( P < O . O O l ) and 24 h potassium excretion increased by
approximately 19 mmol (P<O.Ol). These effects were
confined to the first two 8 h periods and thus occurred
before and during the usual lithium clearance period.
5. Plasma renin activity, measured in 1 0 subjects,
increased after the 600 mg dose of lithium carbonate
( P < O.OOS), but plasma concentrations of aldosterone and
atrial natriuretic peptide were not significantly affected.
Neither the 300 mg dose of lithium carbonate nor the
placebo tablets affected hormone levels.
6. It is recommended that the test dose of lithium
carbonate for use in lithium clearance studies should not
exceed 300 mg.
Correspondence: Dr D. G. Shirley, Department of Physiology,
Charing Cross and Westminster Medical School, London W6
8RF.
Abbreviations: ANP, atrial natriuretic peptide; pRA,
plasma refin activity.
INTRODUCTION
'
In recent years the renal clearance of lithium has achieved
prominence as a non-invasive method for assessment of
proximal tubular function in physiological and clinical
investigations [ l , 21. Lithium clearance studies in man
generally employ a single test dose of lithium carbonate,
usually in the dose range 300-600 mg, taken orally the
evening before the clearance measurements. There is little
doubt, that when administered chronically in large,
therapeutic doses for the treatment of affective disorders,
lithium can cause polyuria and a transient natriuresis [3].
It is important to establish, therefore, whether or not the
much smaller single test doses of lithium used in lithium
clearance studies can also influence sodium balance.
Although some groups have reported small increases in
the excretion of sodium (and in one case, potassium) after
administration of a test dose of lithium [4-61, the
protocols followed were not typical of that generally used
when measuring lithium clearance: the lithium was taken
either at a different time of day from usual [4] or at a dose
exceeding that normally employed [5, 61. In a study using
a lower dose of lithium carbonate (300 mg), Strazzullo et
al. [7] could find no effect on sodium or potassium
excretion, although it might be significant that urine was
collected for only a short period of time after the administration of lithium and the investigation was confined to
hypertensive subjects.
The present investigation was undertaken to define
more fully the possible natriuretic and kaliuretic effects of
a single dose of lithium carbonate (300 mg or 600 mg)
under the conditions normally employed for
measurements of lithium clearance in man. Endocrine
60
D. G. Shirley et al.
status was also assessed. The objective was to identify, if
possible, a suitable test dose of lithium for use in lithium
clearance studies.
METHODS
A double-blind study was performed on 15 healthy,
normotensive individuals (nine male, six female), aged
19-42 years. Written informed consent was obtained
from each subject and the study was approved by the
Hospital Ethical Committee. The experiments were
conducted over three 5-day periods separated by
intervals of at least 2 days. Under the supervision of the
Metabolic Unit dietitian, the subjects adhered to a diet
containing 100 mmol of sodium and 70 mmol of
potassium/day throughout each 5-day period. The
subjects were instructed to avoid alcohol. They followed
their normal daily activities, but strenuous exercise was
avoided. Urine was collected for the final 48 h of each
5-day period; collections were divided into six 8 h
periods, from 23.00 hours on day 3 to 23.00 hours on day
5. The intake of tea and coffee during the final 48 h was
standardized for each subject. On day 4, at 23.00 hours,
subjects took either lithium carbonate (Priadel; Delandale
Laboratories, Canterbury, Kent, U.K.) or matched
placebo tablets. The dose of lithium carbonate was 300
mg (8.1 mmol of Li) or 600 mg (16.2 mmol of Li). The
three treatments were assigned in random order.
In ten subjects, venous blood samples (50 ml) were
taken, after 10 min sitting, on days 4 and 5 beween 09.00
hours and 11.00 hours.
Urinary sodium and potassium concentrations were
measured by flame photometry (model 543; Instrumentation Laboratory, Warrington, Cheshire, U.K.), and
plasma lithium concentrations by atomic absorption
spectroscopy (model 151; Instrumentation Laboratory).
R i a . was used to measure plasma renin activity (PRA) [8]
and the plasma concentrations of aldosterone [9] and
(after Sep-Pak extraction) atrial natriuretic peptide (ANP)
[lo].
Results are presented as means k SEM. Statistical
analysis was by analysis of variance with repeated
measures. Where significant differences were found by
using analysis of variance, paired measurements (control
day versus treatment day) were then compared by using
the least significant difference derived from the pooled
error variance. Because of skew of the data for plasma
hormone measurements, these results were assessed by
using Wilcoxon's signed ranks test for matched pairs
(control day versus treatment day). Results are reported
as significant when the level of significance was P<O.O5
in a two-tailed test.
periods 2 (07.00-15.00 hours; P < O . O O l for each cation)
and 3 (15.00-23.00 hours; sodium, P<0.05; potassium,
P<O.Ol) than during period 1 (23.00-07.00 hours).
However, there was no significant time effect for urine
volume.
Excretion of electrolytes and water did not differ
between the 3 control days, either for the total 24 h
collection or for individual 8 h periods.
Effect of lithium
Twenty-four hour excretion rates (Fig. 1). Administration of placebo tablets had no discernible effect on the
Placebo
300 mg of
Li,CO,
150
4
600 rng of
Li,CO,
**
150
c
0
.+
;T
100
Urn
+:
%
'2
'C
50
3
0
2.5
G
**
2.0
e
CI
:
v
-5
~
2al
1.5
1.0
c
3
0.5
RESULTS
Comparison of control days
On each of the 3 control days that preceded administration of lithium carbonate or placebo, excretion of
sodium and potassium followed normal circadian
rhythms (see Fig. 2), with higher excretion rates during
0.0
Fig. 1. Twenty-four hour excretion rates for sodium,
potassium and water. Values are means with bars indicating SEM. 0, Control day; m, treatment day. Statistical
significance: *P< 0.05, **P< 0.001 compared with control day.
61
Effect of lithium on electrolyte excretion
and water during periods 1 and 2; excretion rates had
returned to normal by period 3.
Plasma hormone and lithium concentrations. There
were no significant differences in values for PRA,
aldosterone or ANP between the 3 control days (Table 1).
Neither placebo tablets nor the 300 mg dose of lithium
carbonate significantly affected the plasma concentration
of any of the hormones. After 600 mg of lithium
carbonate, PRA was increased, but there was no significant change in the plasma concentrations of aldosterone
or ANP.
Plasma lithium concentrations, 10-12 h after ingestion
of the 300 mg and 600 mg dose of lithium carbonate,
were 120 k 10 and 241 f 23 pmolll, respectively.
excretion of sodium, potassium or water. However, after
the 300 mg dose of lithium carbonate, sodium excretion
increased in 11 out of 15 subjects (mean 24 h increase
compared with control day= 17 mmol, P<0.05).
Although the natriuresis appeared to be accompanied by
a moderate increase in urine volume, this was not statistically significant over the 24 h period. Potassium excretion
was unaffected.
After the 600 mg dose of lithium carbonate, sodium
excretion increased in 14 out of 15 subjects (mean 24 h
increase compared with control day = 48 mmol,
P< 0.001). Urine volume also increased significantly, as
did potassium excretion.
Eight hour excretion rates (Fig. 2). Placebo tablets had
no effect on excretion rates during any of the 8 h periods.
After the 300 mg dose of lithium carbonate, sodium
excretion appeared to increase during the first two 8 h
periods (periods 1 and 2) as compared with the corresponding periods of the control day, but the changes were
not statistically significant. Urine volume increased significantly during period 1.
The 600 mg dose of lithium carbonate resulted in significant increases in the excretion of sodium, potassium
DISCUSSION
The use of the lithium clearance method to provide an
estimate of end-proximal fluid delivery has gained
increasing popularity in recent years. Although concerns
over its specificity have been expressed (see [l]),there is
little doubt that it remains the best method available for
assessment of proximal tubular function in man. A further
.-0
g-
+d
60-
:s
x c
2$
2-
2
'C
3
4020-
0-
60-
/'
40-
0-
1.0.
m
0.8-
0.8-
-59
0.6-
8
0.2-
h
3
0.4-
0.0-
g=====+=-
0-
'
l o - ***
*
o'61
0.4
0.2 -
'
:::**3
20-
20 -
1.0 -
c
&
***
0.0-
0.2.
,
0.0 -
'
D. G. Shirley et al.
62
Table 1. Plasma hormone levels in 10 subjects on control and treatment days
Values are means fSEM. Statistical comparisons between treatment days and corresponding control days were made by
using Wilcoxon’s signed ranks test for matched pairs: *P<0.005 compared with the corresponding control day.
PRA(pmolofAng1h-’ ml-I)
Aldosterone (pmol/l)
ANP (pg/mI)l
Control
Placebo
Control
300 mg of
Control
600 mg of
2.15 k 0.34
6 8 2 f 198
12.0 f 3.0
2.21 k 0.28
651 146
13.5 k 3.1
2.29 f 0.29
594k102
15.3 k 5.7
2.69 k 0.33
684+115
13.8 f 3.2
2.02 f 0.29
6 9 4 f 165
12.7k3.0
3.40 f 0.52*
8 2 9 f 192
13.0f2.9
+
Li,CO,
Li,CO,
?Seven subjects only; values in three subjects were below the limit of sensitivity of the assay (3 pg/ml).
concern, however, is whether the normal test dose of
lithium used in such studies might affect sodium and
potassium excretion (see the Introduction). It is self-evident that any method used to assess renal function should
not itself affect renal function.
The present results demonstrate a clear effect of the
usual test dose of lithium carbonate (600 mg) on sodium
excretion, amounting to some 50 mmol over a 24 h
period. Effects on potassium excretion and urine volume
were also observed (although since water intake was not
standardized, the data for urine flow can give only
qualitative information). Most of the effects occurred
during the first and second 8 h periods, when plasma
lithium levels will have been at their highest [ l l ] ;
excretion rates returned to control values during the final
8 h period. Thus at the time when lithium clearance
measurements would normally be performed, i.e. during
the second 8 h period, a significant sodium loss had
occurred and sodium excretion was still elevated. During
this period, PRA was also raised, possibly as a result of
the sodium deficit, although no significant changes in the
plasma concentrations of aldosterone or ANP were
observed. The restoration of normal sodium excretion
rates during the final 8 h period does not necessarily
imply that the lithium was no longer active. The normal
response to acute sodium loss is activation of compensatory mechanisms and consequent sodium retention.
Thus the absence of sodium retention after the initial
natriuresis in the present study suggests that the effect of
lithium may have persisted for up to 24 h.
Little is known about the mechanism(s) underlying the
lithium-induced natriuresis. It is possible that chronic,
therapeutic doses of lithium inhibit the action of aldosterone on the distal nephron [12, 131, but such an effect
seems inconsistent with the kaliuresis observed in the
present study after the 600 mg dose. Rather, the simultaneous increases in sodium and potassium excretion
are consistent with a proximal tubular effect. In this
context, a micropuncture study in rats has shown that
large intravenous infusions of lithium lead to reduced
fractional sodium reabsorption in proximal tubules [ 141,
an effect which might result from inhibition of the
sodium/proton antiport mechanism in the luminal
membrane [15, 161.
After administration of the lower dose of lithium
carbonate (300 mg), 24 h sodium excretion was only
moderately elevated (by < 2 0 mmol on average); again,
the natriuresis appeared to be confined to the first two
collection periods. Urine volume was also moderately
increased during the first collection period. However,
potassium excretion was not significantly affected, and the
subjects’ endocrine profiles remained unaltered.
The findings from this double-blind, placebocontrolled study are broadly consistent with previous
observations. Although Strazzullo et al. [7] found no
effect of 300 mg of lithium carbonate on sodium or potassium excretion in a group of hypertensive patients, their
observations were confined to a 4 h period that started
9 h after lithium administration and therefore missed
most of the critical period. Another study, published in
abstract form only, found that a dose of 500 mg of lithium
carbonate caused small but significant increases in the
excretion of water, sodium and potassium, measured over
a 12 h period immediately after administering the lithium
[4], and a recently published letter has reported an
increase in fractional sodium excretion after a 600 mg
dose of lithium carbonate [17] (although the latter study
could not provide quantitative information, since sodium
intake was uncontrolled). Finally, Lee and co-workers [5,
61 have shown that, when given in the evening, 750 mg of
lithium carbonate (a higher dose than that normally
employed for clearance measurements) results in a significant natriuresis the next morning [5, 61. No data were
given for potassium excretion.
In the present study we have attempted to provide a
more thorough assessment of the effect of lithium by
monitoring excretion rates for 24 h after its administration, the lithium being taken at the time generally used in
studies of lithium clearance. Our objective was to assess
any changes in sodium and potassium balance which
might occur during lithium clearance measurements, and
to identify, if possible, a suitable dose of lithium for such
studies. A further index of physiological status during the
‘normal’ clearance period was provided by assaying the
principal hormones whose concentrations would be
expected to be influenced by changes in sodium balance.
Our results suggest that measurements of lithium clearance after administration of the ‘standard’ 600 mg dose of
lithium carbonate should be interpreted with caution in
view of the clear effect of this dose on sodium balance. On
the other hand, the 300 mg dose, which still produces
plasma lithium concentrations which can be measured
accurately, caused only minor increases in sodium and
water excretion and had no detectable effect on potassium excretion or endocrine status. Whilst any effect on
sodium excretion is undesirable, it is likely that if the
Effect of lithium on electrolyte excretion
lithium dose were reduced much further the plasma
lithium concentration would approach the limits of precision of standard methodology (flame photometry or
atomic absorption spectroscopy), thereby introducing an
additional error into the clearance measurements.
Ideally, endogenous lithium clearance should be
measured, by using flameless atomic absorption [ 181.
However, until such measurements can be performed
routinely with the same precision as those of exogenously
administered lithium, we would recommend the use of
lithium carbonate at as low a dose as possible, certainly
not exceeding 300 mg, for future studies of lithium
clearance.
REFERENCES
1. Koomans, H.A., Boer, W.H. & Dorhout Mees, E.J. Evaluation of lithium clearance as a marker of proximal tubule
sodium handling. Kidney Int. 1989; 36,2-12.
2. Thomsen, K. Lithium clearance as a measure of sodium and
water delivery from the proximal tubules. Kidney Int. 1990;
37 (Suppl. 28), S10-16.
3. Singer, I. Lithium and the kidney. Kidney Int. 1981; 19,
374-87.
4. Peart, W.S., Reid, P. & Sutters, M. Urinary sodium,
potassium and water excretion in man after a single dose of
13.8 mmol of lithium carbonate [Abstract]. J. Physiol.
(London) 1987; 390,204P.
5. Jeffrey, R.F., MacDonald, T.M., Brown, J., Rae, P.W.H. &
Lee, M.R. The effect of lithium on the renal response to the
dopamine prodrug gludopa in normal man. Br. J. Clin.
Pharmacol. 1988; 25,725-32.
6. Freestone, S., Jeffrey, R.F., Bonner, C.V. & Lee, M.R. Effect
of lithium on the renal actions of a-human atrial natriuretic
peptide in normal man. Clin. Sci. 1990; 78,371-5.
7. Strazzullo, P., Iacoviello, L., Iacone, R. & Giorgione, N. Use
of fractional lithium clearance in clinical and epidemiologi-
63
cal investigation: a methodological assessment. Clin. Sci.
1988; 74,651-7.
8. Roulston, J.E. & MacGregor, G.A. Measurement of plasma
renin activity by radioimmunoassay after prolonged cold
storage. Clin. Chim. Acta 1978; 88,45-8.
9. James, V.H.T. &Wilson, G.A. Determination of aldosterone
in biological fluids. In: Reid, E., ed. Methodological
developments in biochemistry, Amsterdam: Elsevier, 1976:
VOI. 5,149-57.
10. Sagnella, G.A., Markandu, N.D., Shore, A.C. & MacGregor,
G.A. Effects of changes in dietary sodium intake and saline
infusion on immunoreactive atrial natriuretic peptide in
human plasma. Lancet 1985; ii, 1208-11.
11. Atherton, J.C., Bielinska, A., Davison, J.M., Haddon, I.,
Kay, C. & Samuels, R. Sodium and water reabsorption in the
proximal and distal nephron in conscious pregnant rats and
third trimester women. J. Physiol. (London) 1988; 396,
457-70.
12. Stewart, P.M., Grieve, J., Nairn, LM., Padfield, P.L. &
Edwards, C.R.W. Lithium inhibits the action of fludrocortisone on the kidney. Clin. Endocrinol. 1987; 27,63-8.
13. Stewart, P.M., Atherden, S.M., Stewart, S.E., Whalley, L.,
Edwards, C.R.W. & Padfield, P.L. Lithium carbonate - a
competitive aldosterone antagonist? Br. J. Psychiatry 1988;
153,205-7.
14. Hecht, B., Kashgarian, M., Forrest, J.N., Jr. & Hayslett, J.P.
Micropuncture study on the effects of lithium on proximal
and distal tubule function in the rat kidney. Pfliigers Arch.
1978; 377,69-74.
15. Bank, N., Lief, P.D., Aynedjian, H.S. & Mutz, B.F. Studies of
the urinary acidification defect induced by lithium. Am. J.
Physiol. 1982; 242, F23-9.
16. MacLaughin, M. & Mello Aires, M. Renal acidification
defect induced by lithium in control and acidotic rats. Clin.
Sci. 1990; 79,23-7.
17. Luik, A.J., Straub, J.P., Martens, H.J.M. & Donker, A.J.M.
Lithium carbonate orally increases renal sodium excretion
[Letter]. Nephron 1990; 54,355.
18. Durr, J.A., Miller, N.L. & Alfrey, A.C. Lithium clearance
derived from the natural trace blood and urine lithium
levels. Kidney Int. 1990; 37 (Suppl. 28), S58-62.