435 J. Physiol. (I946) I04, 435-442 6I2.464.I THE EFFECT OF ETHYL ALCOHOL AND SOME OTHER DIURETICS ON CHLORIDE EXCRETION IN MAN BY M. GRACE EGGLETON AND ISABEL G. SMITH, From the Physiology Department, University College, London (Received 8 November 1945) Evidence as to the nature of the diuresis following ingestion of ethyl alcohol (ethanol) suggests that it is similar to the diuresis following ingestion of water. In view of the lack of absolute proof of their identity, however, a further comparison of the two has been made in respect of certain properties by which a water diuresis can be differentiated from other types of diuresis. One such property is the response of the kidney in respect of excretion of chloride; an increased rate of urine flow is in general associated with an increased output of chloride, but, in a water diuresis, the great increase in rate of flow is accompanied by a diminished chloride output (Eggleton, 1943). The results of a series of class experiments on the effect of exercise on renal function, when different diuretics were used to promote urine flow, suggested that alcohol and water had the same effect in diminishing excretion of chloride (Eggleton, 1945), whereas tea increased it. The matter has now been investigated more fully and under more standardized conditions, without the additional effects of exercise. METHODS The general procedure was the same as that used in previous investigations. Each experiment was performed at least 5 hr. after the last meal, usually in the early afternoon; if in the morning, breakfast was omitted. Usually a glass of water was taken 2-3 hr. before the experiment began, to ensure a reasonable state of body hydration. In the experiments now reported, all substances under investigation were taken by mouth, and at least one urine sample was collected before their ingestion. In a few of the earlier experiments, urinary chloride was determined by the electrometric titration method (Eggleton, Eggleton & Hamilton, 1937); in the remainder, a direct titration method, using tartrazine as indicator, was used with complete satisfaction. This method is simple and was foun& to be unexpectedly accurate over a wide range of chloride concentration in urine. The urine was delivered slowly from a 10 c.c. micro-burette into a boiling tube containing 2 c.c. 0-1 N-AgNO3 solution, one drop of 2% tartrazine and 0-15 c.c. conc. HNO3. At the end-point, the buff-yellow colour adsorbed on the precipitate changes to a greenish yellow colour diffused through the solution. Under these conditions, the range of chloride concentration encountered in the urine samples during the course of this research (10-250 m.equiv./l.) could be determined with an accuracy of ±1%, although the titration values ranged from 0-8 to 20 c.c. 29-2
436 M. G. EGGLETON AND I. G. SMITH RESULTS, In a group of subjects from which data on the changes in urine ph -following ingestion of water and of ethyl alcohol have been already published (Eggleton, 1945), chloride output was also determined in three individuals. The average results given in Fig. 1 indicate that, in these three, chloride output was diminished as greatly during the alcohol diuresis as during the water diuresis. Fig. 1. 560 c.c. HpO or Minutes 40 g. alcohol in 250 c.c. The average changes in chloride output during water diuresis and during alcohol diuresis in three subjects. + + Water diuresis. + + Alcohol diuresis. Opportunity arose later, during an investigation into some further effects of alcohol, to extend the observations to a larger and different group of subjects. The average results obtained on the twelve subjects studied are shown in Fig. 2. There was a considerable individual variation in the degree to which the chloride output was decreased during alcohol diuresis, as had also been noted in connexion with water diuresis, but, in all but one subject, some diminution was observed. On the average, the chloride output, which rose with the first increase in rate of urine flow, fell again at the peak of
ALCOHOL DI URESIS AND CHLORIDE EXCRETION IN MAN 437 diuresis and later reached a value of 60% of the original resting value at a comparable rate of flow. This average final value ranged from 20 to 100% of the original; in no subject was any actual increase observed, apart from the characteristic early rise shown in Figs. 1 and 2. This evidence is decisive as to the identical effects of water and of alcohol in promoting a diuresis with a diminished excretion of chloride; but its value in differentiating these two diuretics from all others depends on the accuracy of the commonly accepted fact that these others increase the output of chloride. Evidence in support of this concept is scattered and much of it is + ~~~~~~~~~~~~~~~12 Chloride output (m.equiv./hr.) 10 14 8 6 Fig. 2. 6-4- Urine flow 2 = (c.c./min.) 0 I I I f I - 30 60 90 120 32 g. alcohol Minutes in 200 c.c. The average change in chloride output during alcohol diuresis in twelve subjects. derived from experiments on anaesthetized animals and on the isolated kidney. It seemed of interest, therefore, to attempt to demonstrate in an unanaesthetized subject the opposing actions on chloride excretion of water and of alcohol on the one hand, and of a series of typical diuretics on the other. The experiments were carried out under the usual standard conditions over a period of 3-4 months. No attempt was made to standardize the diet during this period, but the general regime did not suffer any major changes. Variations in atmospheric temperature and humidity were great during the period in question, and were probably largely responsible for the variable rate of urine flow encountered in the resting sample on different experimental days. The results given in Table 1 suggest that this variable rate of flow was the major factor responsible for variations in the value of the resting chloride output.
438 M. G. EGGLETON AND I. G. SMITH TABLE 1. The relationship between rate of urine flow and chloride output under resting conditions in one subject Rate of urine flow Chloride output Date c.c./min. m.equiv./hr. 3. v. 45 1.87 14-9 12. iii. 45 1*82 13-2 4. v. 45 1-45 13-4 15. iii. 45 1.4 10-8 14. vii, 45 0-93 8-3 7.v.45 0.9 7.4 21. vi. 45 0-66 5.9 22. iii. 45 0-6 4-7 25 g. alcohol 5g. pot. cit. Minutes in 200 c.c. in 250 c.c. Fig. 3. A comparison between the chloride outputs during diureses of the same magnitude induced by alcohol and by potassium citrate. +- + Potassium citrate. +- + Alcohol. In addition to experiments with water and with ethyl alcohol, the effects of the following substances were examined: caffeine, urea, potassium citrate and glucose. It may be said at once that any diuresis resulting from ingestion of these substances was invariably accompanied by an increase, usually a large
ALCOHOL DIURESIS AND CHLORIDE EXCRETION IN MAN 439 increase, in excretion of chloride. In Fig. 3 the results obtained after,ingestion of potassium citrate have been compared with those after ingestion of alcohol, since the starting-point, as regards rate of flow and chloride output, was practically identical on the two occasions, and the diuretic actions comparable in magnitude. This subject showed less depression of chloride output after alcohol than the average: at the height of diuresis chloride excretion fell from 8*4 to 5*4 m.equiv./hr. In comparison, the rather larger diuresis after ingestion of 10 g. potassium citrate in 500 c.c. water was accompanied by an increase in chloride output from 7-4 to 30'8 m.equiv./hr. I II Chloride output + 24 (m.equiv./hr.) Caffeine 20 50 cwwater 1 +~~~~~~~~~~~~+1 9 - + 8 16 6 3 -Urinle flow. - ~~(c.c.imin.) L Fig. 4. 0 30 tf 60 90 120 150 500 c.c. water or Minutes 0-6 g. caffeine in 20 c.c. A comparison between the chloride outputs during caffeine diuresis and during water diuresis. + + Caffeine. +-----+ Water. Similarly, in Fig. 4, there can be seen the slight depression of chloride output accompanying a large water diuresis, and a rise from 11 to 25 m.equiv./hr. chloride with the smaller diuresis occasioned by ingestion of 0-6 g. caffeine in 20 c.c. water. Urea (32 g. in 200 c.c.) gave a response similar to that of caffeine, the chloride output rising from 8-7 to 22 m.equiv./hr., and in this experiment, as in those depicted in Figs. 3 and 4, the peak of chloride excretion lagged behind the peak of the diuresis. Glucose ingestion (200 g. in 220 c.c.) gave rise to no appreciable diuresis nor to any great rise in chloride output:
440 M. G. EGGLETON AND I. G. SMITH the flow rose only from 066 to 1.0 c.c./min. during the height of glycosuria and the chloride output increased from 5-9 to 7-3 m.equiv./hr. just after the peak of diuresis. Further experiments were performed on the group of subjects on whom the results shown in Fig. 2 were obtained; these provided some evidence concerning the closeness of association between the chloride-depressing and the diuretic actions of alcohol, since, in the case of water diuresis, it has been suggested that two distinct mechanisms may be involved. In one subject (not included in the average shown in Fig. 2), ingestion of alcohol produced practically no diuresis, nevertheless there was a typical fall in chloride excretion (Fig. 5). The alcohol was undoubtedly absorbed: it was noted at the time as having a prolonged moderate action on the higher nervous centres; but the slow rate l + 1 _# _10 Chloride output 8 (m.equiv./hr.) +_ 6 Fig. 5. 2 Urine flow 4 (c.c/min.) 1 _ 0 30 60 90 120 I 150 180 32 g. Minutes alcohol (200(c.c.) Showing the suppression of chloride output following ingestion of alcohol in one subject when the diuresis was inhibited for an unknown reason. of increase in blood alcohol concentration, probably resulting from slow and erratic emptying of the stomach, may have been insufficient to affect a rather insensitive pituitary mechanism. Such a state of affairs was encountered on a previous occasion (Eggleton, 1942) and in that subject also the pituitary mechanism was insensitive, i.e. there was an unusually small diuretic response to ingestion of water. In the subject used in the experiment illustrated in Fig. 5, 500 c.c. water was followed by a diuresis of only 355 c.c.; again, however, the usual fall in chloride output was observed, from 14-4 to 541 m.equiv./hr. (at identical rates of urine flow). In several subjects, subcutaneous administration of nicotine (1 mg. in the form of tartrate) before ingestion of alcohol almost suppressed the diuresis. It can be seen from their average results, shown in Fig. 6, that the typical decrease in chloride output nevertheless occurred. In other subjects, nicotine enhanced the diuresis, but the- typical diminution of chloride excretion resulting from ingestion of alcohol
ALCOHOL DIURESIS AND CHLORIDE EXCRETION IN MAN 441 was again observed. The conclusion that nicotine itself cannot be held responsible for the decrease in chloride output is strengthened by the further observation that the magnitude of this decrease in the twelve subjects studied was of the same order, whether the alcohol was preceded by nicotine or not. A detailed report of these nicotine experiments will be given elsewhere. + I0 Chloride output (m.equiv./hr.) 8 + 4 + 4 Urine flow 2 (c.c./min.) 12 6 0 0 t30_ 60 90 120 150 180 Fig. 6. 1mg. 32 g. alcohol Minutes nicotine in 200 c.c. subcutaneously Showing the average suppression of chloride output in four subjects following ingestion of alcohol when the diuresis was inhibited by nicotine. DISCUSSION It has been demonstrated clearly that ethyl alcohol shares with water the property of promoting a diuresis with diminution of chloride excretion, in contrast to such other and varied diuretics as caffeine, potassium citrate and urea, which promote a large increase in chloride output. Of the mechanism responsible for this differentiation between the two groups, little fresh evidence has come to light. It is now generally agreed (Goodman & Gilman, 1940) that the diuretics in both groups act by diminishing water reabsorption in the tubules, though caffeine and alcohol may also increase the glomerular filtration rate in the early stages. If this view of the action of diuretics in general be accepted, they might all be expected to cause an increased output of chloride, as part of the general 'flushing-out' process accompanying a faster rate of flow. There is now considerable doubt, however, as to the validity of this conception in the whole animal as distinct from the isolated kidney preparation. Various other explanations of the increase in chloride output after ingestion of caffeine and of potassium citrate are equally tenable, and it has been shown (McCance & Young, 1944) that increase in chloride excretion after
442 M. G. EGGLETON AND I. G. SMITH ingestion of urea may, in some circumstances (dehydration), be lacking in spite of an increase in rate of urine flow. Thus, many factors can apparently override the simple 'flushing-out.' process. Until conclusive evidence concerning the mechanism responsible for increased chloride reabsorption during water diuresis is forthcoming, the view that some pituitary factor, associated with the anti-diuretic hormone, is concerned remains the most probable. The fact that ethyl alcohol initiates the same chain of events, i.e. a large diuresis after a 20-30 min. lag accompanied by a decreased output of chloride, suggests that the two factors, antidiuretic and chloride-reabsorbing, are closely linked. That they are not identical has been already suggested in the case of water diuresis (Eggleton, 1943), corroborating an earlier conclusion of Fee (1928) on dogs. He found that inhibition of a water diuresis by various anaesthetics did not prevent the typical fall in chloride output normally accompanying the diuresis. Later, Bayliss & Brown (1940), working on hypophysectomized decerebrate dogs, were led to the tentative conclusion that some structure in the tuber cinereum was concerned with the excretion of chloride. The fact that a similar dissociation of the two responses of the kidney, in regard to the excretion of water and of chloride, can be observed when the diuresis is stimulated by alcohol in place of water is again strongly suggestive that the two agents are working through the same intermediary mechanism. SUMMARY 1. The depression of chloride output in the urine accompanying water diuresis is observed also during alcohol diuresis (Figs. 1, 2). 2. Comparison, in the same individual, of these two diuretics with others shows that they are unique in this respect. Potassium citrate (Fig. 3), caffeine (Fig. 4) and urea all produce a large increase in chloride output during the diuresis. 3. The decrease in chloride output after ingestion of alcohol, as in that after ingestion of water, is not closely correlated with the degree of diuresis and may be observed when this is almost completely suppressed (Figs. 5, 6). REFERENCES Bayliss, L. E. & Brown, A. (1940). J. Ph'y8iol. 98, 190. Eggleton, M. G. (1942). J. Physiol. 101, 172. Eggleton, M. G. (1943). J. Phy8iol. 102, 140. Eggleton, M. G. (1945). J. Physiol. 104, 312. Eggleton, M. G., Eggleton, P. & Hamilton, A. M. (1937). J. Physiol. 90, 16. Fee, A. R. (1928). J. Pharmacol. 34, 305. Goodman, L. & Gilman, A. (1940). The Pharmacological Basis of Therapeutics. New York: The Macmillan Company. McCance, R. A. & Young, W. F. (1944). J. Physiol. 102, 415.