POSTGRAD. MED. J. (I963), 39, 19

ELECTROLYTES AND DEPRESSIVE ILLNESSJ. L. GIBBONS, M.R.C.P., D.P.M.

The Department of Psychiatry, Institute of Psychiatry, Maudsley Hospital, London, S.E.5

Interest in water and electrolyte metabolism inmental illness is not new, nor is it confined todepressive illness. In fact schizophrenic patientshave been intensively studied and there have been,for example, reports of phasic variations in sodiumexcretion in the acute episodes of periodic catatonia(Gjessing, I953) and of delayed diuretic responseto a water load taken by mouth in chronic schizo-phrenic subjects (Pfister, I937). The first studiesin depressives derived from the old observationthat the urinary output tended to be low in de-pressive and high in manic illness (Allers, I914).The obvious explanation, that manic patients areoften thirsty and drink more, was not thought tobe the complete explanation in every case and, aswe shall see, there is justification for this view.The first thorough attempt to correlate varia-

tions in metabolism, including the excretion ofwater and sodium, with variations in mental illnesswas made by Gjessing (1932, I953) in his famousstudies of periodic catatonia. This is a rarecondition in w'hich attacks of stupor or excitementoccur at regular and predictable intervals.Gjessing's patients were given a fixed intake offood and fluids for many months, so that a detailedbalance study was made during several cycles ofillness. Most depressive patients, however, eitherhave only one attack of illness or have their attacksat long and quite irregular intervals. A very fewpatients suffer from rapid alternations of mood, ashort period of elation and overactivity beingfollowed by a similar period of depression andretardation, with perhaps a day or more of normalmood at some stage of the cycle. Such patientscan obviously be studied in Gjessing's way-always provided that they will cooperate with theinvestigator. Only a handful of studies have beenreported, but they are worth considering in somedetail for the light they may throw on the possiblemetabolic accompaniments of the commoner andmore prolonged depressive illnesses.The most intensive study is that reported by

Klein and Nunn (I945). Their patient, a managed 67, had suffered from chronic depression for14 years from the age of 5 I. Later his illness tookthe form of regular weekly cycles consisting of fivedays of depression and retardation and two days ofelation and overactivity. Various autonomicchanges accompanied the changes in mood: thushis pulse was slower and his blood pressure lowerduring the depressed phases. For a period of

several months he was given a constant mixeddiet, with a constant fluid intake, and 12-hourlyurine collections were made. When biochemicalinvestigations were in progress he was kept in bed.The findings of interest in the present context maybe summarized as follows. During the depressivephases the patient retained sodium and water,which was excreted again during the manic phase.The daily volume of urine averaged 780 ml. in thedepressive and i,670 ml. in the manic phases, andthe urinary excretion of sodium and chloride wasincreased by 6o to 70% during mania. Potassiumexcretion was unrelated to mood change and thesame was true of other urinary constituents(calcium, ammonia, phosphate and nitrogen).Daily blood samples were taken throughout threeconsecutive cycles but no variations in hlimatocritor hemoglobin concentration were seen. Thechange in weight was in the direction expected-again of o.9 to I.3 kg. during depression which waspromptly lost again during mania. More detailedstudy of the rate of urine flow during the change-over from depression to mania showed that anabrupt increase in urine flow actually preceded themood change.The second of the two cases reported by

Crammer (1959) was very similar. The patientwas a 49-year-old man with a regular six to sevenday cycle for at least five years. Two days werespent in stupor and four in manic overactivity.On a constant intake of food and fluids this patientalso showed retention of water and sodium andgain in weight during his depressed phase-about225 mEq. of sodium and I.75 kg. in weight.Polyuria and excretion of sodium followed duringthe overactive phase. Hamodilution during thestuporous period was indicated by appropriatefluctuations in packed cell volume. In thispatient also there were no regular variations in theserum concentration or urinary excretion ofpotassium. Crammer observed that the magnitudeof the weight gain varied directly with the amountof sodium in the diet. Thus the gain was onlyabout 700 g. on a diet containing I5 mEq. ofsodium a day, while fluid restriction (to 500 ml. aday) practically abolished the weight changes.It should be mentioned that none of thesemanouvres had any effect on either the timing orthe intensity of the mood change. Moreover thetime relationship between mood and fluid retentionwas reversed after several months of treatment with

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

chlorpromazine. Then the peak of the sodiumand weight loss occurred at the start of the inactivephase while most of the weight was regained afterthe beginning of the overactive phase.

Cases of such short and regular alternation,while highly convenient for the biochemical in-vestigator, are both uncommon and uncharacter-istic. The patient described by Klein and Nunn,for example, slept a great deal during his depressedphases, whereas the typical depressive complainsof insomnia. There are a few published studiesof patients with rather longer episodes of de-pression and mania. The first patient of Crammer(1959), a man of 48 with a six-week cycle, wasstudied through one complete cycle (including afortnight of mania and a fortnight of depression)while on a constant intake of fluids but a quitevariable food intake. This patient lost weightduring the first few days of his period of de-pression, thus behaving in a manner opposite tothat of the two cases already discussed. Study ofCrammer's graph shows that the weight loss wasdue to diuresis of water and sodium on two daysonly and that during the rest of the depressivephase sodium output was less than in the manicphase. The lack of control of the patient's dietmakes it difficult to evaluate the findings. As willbe mentioned below, two diets of equal caloricvalue but different composition may have quitedifferent effects on water and sodium excretion.On the other hand it must be conceded thatCrammer's other patient still showed regularfluctuations in weight and urinary volume when hisdiet was uncontrolled. Klein (1950) also describeda second case, a man of 40 with recurrent attacks ofdepression and mania of varying duration. Thedepressive attack usually lasted about I3 days andthe mania i8 days; the manic phase was followedby about I4 days of normal mood. This patientalso was kept in bed, with a constant intake offood and fluids, and various urinary and bloodconstituents were measured. There was noindication of fluid or electrolyte retention at anystage of the cycle, although the patient's ability toexcrete a water load taken by mouth varied, as willbe discussed below. A study by Str6m-Olsenand Weil-Malherbe (1958) may also be mentioned,since they give brief details of a 30-year-oldwoman with two to three week phases of mania ordepression and one to three days interveningnormal mood. The periods of depression andelation were reported to coincide with minima andmaxima of the total urinary volume and of theexcretion of sodium and potassium. Since foodand fluid intake were uncontrolled, however, toomuch should not be made of this case.We may conclude that in some cases of short-

lived alternating mania and depression changes in

water and sodium excretion accompany the varia-tions in mood. The exact time relationshipsprobably vary in different patients or may be variedin the same patient by drugs. Such changes arenot invariably seen and, particularly if they areslight, cannot be interpreted unless the patient isreceiving a constant food and fluid intake. Theconstancy of food intake is important not onlybecause different foodstuffs contain differentamounts of water and electrolytes. It is wellknown that obese subjects lose weight less rapidlyon a high carbohydrate diet than on a high proteinand fat diet of equal calorie value. Russell (I962)has shown that this variation in weight loss occurseven though the amount of water and sodium in thediet is kept constant and that it is a result of a morepositive sodium and water balance. Similar effectsof high and low carbohydrate di&s on waterbalance have been observed in normal subjects.It is quite conceivable that a patient subject tomood swings might, for example, prefer foodstuffsrich in carbohydrate at one particular stage of hisillness.The conclusions drawn above are based on

findings in only four patients and further studiesof such patients will be of great interest. Changesof a similar type have also been reported in a ratherlarger number of patients suffering from periodiccatatonia, an illness with a different clinical picture.In five patients receiving a constant intake sodiumretention was noted during the phase of illness inall of them and water retention in three (Gjessing,I953). Essentially similar changes were observedin the two cases reported by Rowntree and Kay(I952). Recurrent episodes of fluid retentionhave also been reported in association with psycho-logical symptoms of organic type-clouding ofconsciousness, disorientation, etc. (Biemond, 1949;Broser, 195I). In these cases, however, there wasevidence of structural disease of the brain and thepsychological symptoms, quite different in typefrom mania or depression, may well have beendue to water retention, itself in turn the result ofdisordered central nervous control of osmoticequilibrium. Disorders of electrolyte and watermetabolism have in recent years been recognizedas occasional complications of brain lesions and thesubject has recently been reviewed by Gilbert andGlaser (I96I).When we turn to the more usual form of de-

pressive illness we find a disorder of severalweeks' or months' duration, often of slow onset,which may occur only once in the patient's lifetime.It is at once apparent that it is almost impossible tostudy any metabolic changes which may occur atthe beginning of the illness. There are patientswith fairly frequent although irregular episodes ofdepression and it may very occasionally be possible

Yanuary I96320

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

GIBBONS: Electrolytes and Depressive Illness

to observe such a patient at the beginning of anattack. Generally, however, the depressed patientis not seen until the illness has developed. At thisstage one can make various biochemical measure-ments and see how they compare with normalvalues. It may be said at once that plasma electro-lytes have long been known to be in the normalrange in depressed patients (Altschule, 1953;McFarland and Goldstein, I939). The same istrue of electrolytes in the cerebrospinal fluid(Eichhorn, I954) and also of total body sodium andpotassium measured by isotope dilution (Gibbons,I960; Coppen, Shaw and Mangoni, I962).Comparisons of groups of depressed patients and

of subjects without mood disorder are unlikely toreveal any differences unless they are gross.Since the depressed patient usually recovers,however, it is possible to use him as his own controland compare findings in the phase of illness witht:hose obtained when he is well again. In addition,repeated determinations of certain variables maybe made during the process of recovery in thereasonable expectation that any persistent changeswhich developed at the beginning of the illness willbe reversed at its end. If recovery is hastened byelectrical convulsive treatment or by anti-depressive drugs, it has also to be demonstratedthat any changes found are not merely artefactsof the treatment.

For example, if sodium retention is a feature ofthe early stages of depression, and if such retentionpersists during the illness, diuresis of sodiumwould be expected to accompany recovery. Thesituation is complicated of course by the fact thatmost depressives experience decrease in appetite,and some of them eat so little that they lose aconsiderable amount of weight. Malnutrition, ifsevere, is itself accompanied by considerableperturbations of electrolyte metabolism which willmask any that may be part of the syndrome ofdepression. Similarly recovery from depressionis usually accompanied by restoration of theappetite to normal and consequent gain in weight.Studies of recovering depressed patients shouldtherefore be restricted to those who are not severelymalnourished and, in order to avoid confusion withthe effects of increased food intake, should pre-ferably be made with patients receiving a constantintake of food and fluids.The present author (Gibbons, I960) investi-

gated a group of 24 adult patients who showed theclinical picture of so-called endogenous depres-sion. Total exchangeable sodium was measuredin all the patients with 24Na and total exchangeablepotassium in I3 patients with 42K. The resultswere all in the normal range and the mean values,expressed as mEq. per kg. of body weight, were42.8 and 45.I for sodium and potassium re-

spectively in the male patients, 37.o and 32.8 forthe female. The estimations were repeated afterseveral weeks, by which time i6 patients hadrecovered or improved considerably, while eightwere scarcely changed in mood because of failure torespond to treatment (in most cases electro-convulsive treatment) or because of relapse afterinitial improvement. In the recovered group themean change in exchangeable sodium was a highlysignificant decrease of about 200 mEq.-9% of thetotal-whereas in the control group the meanchange, an increase of about 30 mEq., was notsignificant. No consistent change in total ex-changeable potassium occurred. Dietary varia-tions were not thought to be responsible for thedifference between the two groups since both werenursed on the same metabolic ward, partook of thesame diet and showed similar fluctuations inweight. Moreover ten of the patients received aconstant diet throughout the period of the in-vestigation and the same changes were seen inthem-a decrease of some 200 mEq. in six whorecovered, an increase of 40 mEq. in four whofailed to do so-as in the series as a whole. Theconclusion drawn at the time was that the resultssupported the hypothesis that depression is ac-companied by retention of sodium, which isexcreted during recovery.The patients who received the constant diet

were subjects of another investigation carried outby Russell (I960). Fifteen depressed patients inall were investigated by the metabolic balancetechnique for periods of two to five weeks, duringwhich time they received electrical convulsivetreatment (ECT); i i patients recovered orimproved considerably during the period of study.It was found that ECT usually, but not always,caused slight and quite transient retention of waterand sodium. The mean amounts retained wereapproximately I50 ml. and 12 mEq.; the largestamounts in individual subjects were 450 ml. and50 mEq. Essentially similar results were obtainedwhen patients were subjected to a 'mock' ECTprocedure in which atropine and intravenousbarbiturate were administered but a convulsionwas not induced. These slight and transitoryeffects of ECT or, more probably, of the emotionalreaction aroused by the procedure may becom-pared with the large increases in extracellularfluid, measured by the thiocyanate method,reported by Altschule, Ascoli and Tillotson (I949)in their patients after ECT. This clearly illu-strates the importance of avoiding malnourishedsubjects in this type of work and also of main-taining control of diet. Another interestingfinding reported by Russell (I960) is that improve-ment in the mental state of his patients was notaccompanied by any gain in weight so long as the

Yanuary I963 21

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

calorie intake was kept constant; when the calorieintake was increased, a prompt gain in weightensued.The overall balance in water, sodium and potas-

sium during 66 days of ' recovery ' (i.e. periodswhen improvement in mental state was occurring)was compared with 66 days of ' illness' (periodsbefore improvement had occurred). The groupas a whole showed a slight loss of sodium duringthe period of recovery of 5.5 mEq. per day. Thisfinding was not statistically significant, but it wasin fact of the direction and magnitude to beexpected from the study of the present author(Gibbons, I960) already referred to. Whenindividual patients were considered, four of themactually showed a slight retention of sodium.Changes in water and potassium balance for thegroup as a whole were negligible, although onepatient showed considerable water retention(300 ml. per day) and another even greater loss(430 ml.). Russell concluded that no substantialalteration in the balance of water and electrolytesoccurs as a depressed patient recovers and thatthe main immediate cause of weight gain duringrecovery is an increased calorie intakeresultingfrom an increase in appetite.

At this stage, then, an isotope dilution studysuggested that recovery from depression was ac-companied by loss of sodium while a metabolicstudy, carried out on several of the same patients,showed that if any change did occur it must be ofvery small magnitude from day to day. Theproblem was investigated further by Coppen andShaw (I962), who studied I2 depressed patients,all of them seriously ill, with the isotope dilutionmethod. They were able to use the long-livedradioisotope 22Na, which is administered to thepatient on one occasion only. Several days areallowed for full equilibration to occur and then thetotal exchangeable sodium can be measured dailyfor several weeks by determining the specificactivity of sodium in plasma or urine. Thenecessary correction for radiosodium lost from thebody presents no difficulty since the amountremaining is easily measured by a whole-bodycounter. The 22Na technique will detect smallerchanges in total body sodium and will also showexactly when any changes occur. In fact Coppenand Shaw found no overall change in sodiumduring recovery and the mean change for the groupwas infinitesimal. Their finding is at first sightat direct odds with the present writer's work, butthe two results can become compatible if it besupposed that there is no real change in bodysodium but that changes occur in the amount ofbody sodium which equilibrates rapidly withingested radiosodium. Normally most of thesodium in bone equilibrates only very slowly with

radiosodium. If mixing occurred more rapidlyduring depressive illness, then an estimation oftotal exchangeable sodium at 24 hours would givea higher value during the illness than afterrecovery. It must be admitted that this sounds arather far-fetched explanation, but the point isworthy of further investigation. Further workis in fact being carried out by Dr. Coppen and hiscolleagues, who are also measuring total body waterand extracellular water in their patients. Theirfindings will be awaited with great interest. Thesame investigators will also have data on the bodycontent of 42K, the naturally occurring radio-isotope of potassium. Its estimation gives areliable measurement of the actual total bodypotassium, not merely that fraction which isreadily exchangeable with ingested potassium.It will be interesting to see if results with thistechnique support the conclusions of the presentauthor and of Russell that no significant change inpotassium balance occurs in depressed patients.Some previous measurements of extracellular

fluid (ECF) by the thiocyanate method may bementioned here. Altschule and Tillotson (I949)found ECF values in depressed patients in thelower part of the normal range with an increase,still within the normal range, after treatment.Dawson, Hullin and Crocket (1956) found ratherhigh values during healthy phases in four patientswith mood swings, with a reduction in ECF inassociation with both manic and depressive phases.How far the changes were a result of variation infood or fluid intake is not clear.There is some evidence that the sodium ion is

handled rather differently during depressiveillness. Coppen (I960) measured the rate ofentry of radiosodium (24Na) into the lumbarcerebrospinal fluid (CSF) after rapid intravenousinjection and found that this rate was lower indepressed patients than in a control group (patientsbefore surgical operations) or in a group of schizo-phrenic patients or in a group of patients who hadrecovered from depression. The determinationof the rate of entry requires repeated bloodsampling during the first hour. When the moreusual one-hour ratio was calculated (i.e. theconcentration of radiosodium in the CSF ex-pressed as a percentage of its concentration in theplasma at one hour after injection) exactly similardifferences were found between the depressivesand the other groups. The mean values were2.7 for the depressives, 3.9 for the controls, 4.0 forthe schizophrenics, 3.8 for the recovered depres-sives. There were no 'differences, however, in therate of entry of radioactive water from plasma tolumbar CSF. The author, while not claiming thatthe changes in sodium entry were related to thecauses of depression, pointed out that ECT, an

3tanuary I96322

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

GIBBONS: Electrolytes and Depressive Illness

effective method of treatment of depression, hasbeen shown to increase the permeability of theblood brain barrier.An attempt to confirm this interesting work has

recently been made by a group in Edinburgh(Fotherby, Ashcroft, Affleck and Forest, I962)who measured the one hour CSF-plasma ratio ofradiosodium. They failed to find any significantdifference between a group of 12 depressivepatients and groups of 14 schizophrenic patientsand seven patients with organic brain disease;the mean values were 3.5, 3.4 and 4.4, respectively.It may be significant that the four lowest valueswere found in depressive patients. Moreover themean ratio was only 2.4 in the four patients whowere suffering from severe ' endogenous ' de-pression in fairly pure form. There was greatervariation from patient to patient than in Coppen'sseries. It should be noted that almost all theEdinburgh patients, unlike Coppen's, were re-ceiving phenothiazine drugs. It appears that theonly way to settle the discrepancy between the twoinvestigations will be to measure sodium entry inthe same patients during illness and after recovery.If a change in the blood-CSF barrier is an im-portant feature of recovery this method will showit without need for controls, whether of schizo-phrenics (who are not immune from depression ofmood) or of normal subjects (who are still liable toundergo emotional turmoil).

Certain aspects of water metabolism have fromtime to time been studied in depressives. Lange(1928) mentioned how he had carried out oralwater loading tests in depressed patients someyears previously and had several times observeda delayed diuretic response. He presented agraph from one patient with adelayed responseduring illness and a normal response after recovery.Hoff and Potzl (1930) also found delayed diuresisin seven out of eight depressives (while two manicpatients showed a normal response). They alsoreported that posterior pituitary extract had lessthan the expected degree of antidiuretic activity indepressives. After recovery from depression theresponse to a water load was normal. Similarobservations were reported by Altschule andTillotson (I949). Thorvardsson (I942) also founda delayed diuretic response in nine severelydepressed patients; when the test was repeated,however, after the patients had had an adequatefluid intake (I,500 ml.) for several days, a normalresponse was found. He concluded that the delayin diuresis was merely a manifestation of dehydra-tion, consequent upon the failure of the severedepressive to take enough to drink. Thisexplanation may not be the whole one, however,because the manic-depressive patient reported byKlein (1950) received a constant diet with a con-

stant and adequate amount of fluid and yetshowed a delayed diuretic response to a water loadduring his depressive phases and a brisk diuresisduring his elated phases. It may be that in somedepressives it is not so much a question of de-hydration as of delayed absorption of largequantities of water.

Trolle (I945) reported an investigation intowater excretion by 14 depressed patients in whichhe concluded that recovery may be accompaniedby an increase or a decrease in the volume of urine,the change persisting for about a month. Graphswere presented of four patients who had received aconstant diet and fluid intake: in two cases urinaryvolume increased and in two it decreased byIoo to 300 ml. per day. (These findings may becompared with those of Russell (I960) referred toabove). No data of electrolyte excretion aregiven. Trolle did not think that the variationscould be explained by changes in motor activity,although an explanation on these lines might beconstructed. On the other hand Trolle's workmight be an indication that different patients reactdifferently and a warning that an attempt to find aconstant pattern of electrolyte metabolism may bea pursuit of a will-o'-the-wisp. It is also con-ceivable that variations occur according to the typeof the depression-if it be allowed that there aretypes of depression-or according to the presenceor absence of certain clinical features. ThusAnderson and Dawson (I962) reported that abouthalf of a series of 98 depressives had a high bloodlevel of acetyl-methyl carbinol (AMC). Therewas a strong correlation between AMC level andcertain clinical features, especially slow and hesi-tant speech and preoccupation with depressiveideas. The interest that this observation has forus at the moment lies in the fact that one of thedeterminants of a high blood level of AMC is anincrease in intracellular sodium concentration.

Finally we may refer to the observations ofBussow (I950), who administered water and vaso-pressin to manic-depressive patients in orderto produce water retention. Most of the de-pressive patients became temporarily worse andseveral manic patients became almost stuporous orlapsed into a transient depressive state. Acolleague of Biissow (Karstens, 195 i) repeated themanceuvre in six healthy subjects and reported thatthe psychological effects resembled the symptomsof depressive illness.The investigations reviewed above provide

suggestive but equivocal evidence of some de-rangement of water and sodium metabolism indepressive illness. If further work establishesthat derangements frequently occur, there remainsthe question of the relationship between thechange in mood and the metabolic changes.

3'anuary i 963 23

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAI, January I963

Some authors (for example, Crammer, I959)have suggested that abnormalities in electrolytemetabolism may cause mood changes, at least inoccasional cases. It is customary to refer to thework of Woodbury, who reported that the change inbrain excitability produced in experimental animalsby the administration of steroid hormones was theresult of changes in the concentration of intra- andextracellular sodium in the brain (Woodbury,Timiras and Vernadakis, 1957). Moreoverfluctuations in brain excitability, measured by thephotoconvulsive technique, have been demon-strated to accompany variations in mental state inpatients with recurrent catatonic stupor (Leiber-man and Hoenig, 1953) while fluctuations in waterand electrolyte excretion are a feature of periodiccatatonia. Unfortunately for this line of argu-ment, however, no abnormality in brain excita-bility could be detected when the photoconvulsivethreshold was measured in a series of depressivepatients (Driver and Eilenberg, I960). It is truethat reports of experimental depletion of electro-lytes in otherwise healthy subjects speak oflethargy or apathy as notable symptoms (McCance,1936; Fourman, 1954). These subjects, however,had lost a considerable proportion of their bodystores of sodium or potassium, while it is clear thatdepressed patients show no significant depletion ofthese electrolytes. Sirmilarly the observations ofBussow and of Karstens, that forced water re-tention makes depressives worse and healthysubjects depressed, cast no light on the patho-genesis of depression since the depressed patient isobviously not overhydrated. The notion thatwater and sodium retention may somehow causedepression of mood is also to be found in thedoctrine that such retention causes premenstrualtension-a protean syndrome in which depressionis one of the most characteristic symptoms-andthat the disorder may be successfully treated withdiuretics. (See, for example, Greene and Dalton,1953). It may be mentioned here that it has yetto be demonstrated that the occurrence of pre-menstrual psychological symptoms is regularlyaccompanied by significant water and sodiumretention. A recent balance study in the MetabolicUnit of this hospital failed to demonstrate it in agroup of psychoneurotic women who complainedof severe premenstrual exacerbation of theirpsychological symptoms (Bruce and Russell,I962). Similarly an investigation of normalyoung women with 22Na failed to show the pre-menstrual retention of sodium which is often

spoken of as if it were a quite usual phenomenon(Klein and Carey, 1957).

It seems much more likely that any electrolytechanges in depression are secondary effects of theillness. Some may be due to changes in the*amount or composition of the diet or to variationsin motor activity. Others may be the result of theemotional disturbance itself. It is well known,that emotional upset can cause an increased or*decreased rate of urine flow in human subjects(Chalmers and Lewis, I95I; Miles and deWardener, 1953). Moreover increased adreno-cortical activity, which might lead to variations inrelectrolyte excretion, is known to occur in mostcases of severe depression (Board, Wadeson and&Persky, I9S7, Gibbons and McHugh, I962.),There is also one study, which deserves to berepeated because of its great theoretical interest,which claims to show that different varieties ofemotional experience, whether induced naturallyiby life experience or artifically contrived in thelaboratory, are associated with different patterns4of renal excretion of water and electrolytes(Schottstaedt, -Grace and Wolff, I956). In partiJtcular depression of mood is accompanied by re-tention of salt and water. This might then be thephysiological counterpart of the changes noted by~Klein and Nunn (I945) and by Crammer (I959)in their patients with cyclic manic-depressivepsychosis.SummaryA review of investigations on electrolyte-,

metabolism in depressive illness suggests thefollowing conclusions.

I. In some cases of rapid alternation of maniaand depression variations in water and sodium-metabolism accompany the fluctuations in mood.

2. In the more usual single attacks of depressive&illness the metabolic balance technique failed toreveal such variation, but various studies withWradiosodium indicate that there may be somealteration in sodium metabolism.

3. Changes in water metabolism have been,found in some cases. They may be partly ex-plained in terms of changes in fluid intake, and in4all investigations of this type dietary control isimportant.

4. Changes in water and electrolyte metabolismwhich cannot be explained as the result of alterationin diet or activity are more likely to be the effectrather than the cause of the mood change.

I should like to thank Dr. Coppen, Dr. Forrest andDr. Russell for allowing me to cite papers awaitingpublication.

REFERENCESALLERS, R. (I9I4): Ergebnisse stoffwechselpathologischer Untersuchungen bei Psychosen. III Das manisch-depressive

Irresein, Z. ges. Neurol. Psychiat., 9, 585.ALTSCHULE, M. D. (I953): 'Body Physiology in Mental and Emotional Disorders'. New York: Grune and Stratton.

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from

January I963 GIBBONS: Electrolytes and Depressive Illness 25

, AsCOLI, I., and TILLOTSON, K. J. (1949): Extracellular Fluid and Plasma Volumes in Depressed Patients GivenElectric Shock Therapy, Arch. Neurol. Psychiat. (Chicago), 62, 6i8.and TILLOTSON, K. J. (1949): Effect of Electroconvulsive Therapy on Water Metabolism in Psychotic Patients,Amer. J. Psychiat., IO5, 829.

ANDERSON, W. M., and DAWSON, J. (I962): The Clinical Manifestations of Depressive Illness with Abnormal AcetylMethyl Carbinol Metabolism, Y. ment. Sci., Io8, 8o.

BIEMOND, A. (1949): Un nouveau syndrome hypothalamique: oligurie intermittente avec oedeme du visage et troublessexuels, Folia psychiat. (Amst.), 52, 343.

BoARD, F., WADESON, R., and PERSKY, H. (1957): Depressive Affect and Endocrine Functions, Arch. Neurol. Psychiat.(Chicago), 78, 612.

Bussow, H. (1950): tlber die Wirkung des Wasser-Tonephin-Versuches auf manisch-depressive Zustandsbilder, Arch.Psychiat. Nervenkr., 184, 357.

BROSER, F. (195 I): Periodische Bewusstseinsst6rungen und paroxysmale Comata bei einem Falle von primiirer Oligurienach Hirntrauma, Ibid., I87, 3 1.

BRUCE, J., and RUSSELL, G. F. M. (I962): Weight Changes and Balance of Water, Sodium and Potassium in Womertwith Premenstrual Tension Symptoms, Lancet, ii, 267.

CHALMERS, T. M., and LEWIS, A. A. G. (I951): Stimulation of the Supraopticohypophysial System in Man, Clin. Sci.,10, 127.

COPPEN, A. J. (I960): Abnormality of the Blood-cerebrospinal Fluid Barrier of Patients Suffering from a DepressiveIllness, J. Neurol. Neurosurg. Psychiat., 23, I56.SHAW, D. M., and MANGONI, A. (I962): Total Exchangeable Sodium in Depressive Illness, Brit. med.J., ii, 295-

CRAMMER, J. L. (1959): Water and Sodium in Two Psychotics, Lancet, i, 1122.DAWSON, J., HULLIN, R. P., and CROCKET, B. M. (I956): Metabolic Variations in Manic-depressive Psychosis, J. ment.

Sci., 102, i68.DRIVER, M. V., and EILENBERG, M. D. (I960): Photoconvulsive Threshold in Depressive Illness and the Effect of

E.C.T., Ibid., xo6, 6I i.EICHHORN, 0. (1954): Zur Frage der Elektrolytfunktionen bei Geisteskrankheiten, Nervenarzt, 25, 207.FOURMAN, P. (I954): Depletion of Potassium Induced in Man with an Exchange Resin, Clin. Sci., 13, 93.FOTHERBY, K., ASHCROFT, G. W., AFFLECK, J. W., and FORREST, A. D. (I962): 'Studies on Sodium Transfer and

s-Hydroxyindoles in Depressive Illness' (to be published).GIBBONS, J. L. (I960): Total Body Sodium and Potassium in Depressive Illness, Clin. Sci., I9, I33.

and McHUGH, P. R. (I962): Plasma Cortisol Levels in Depressive Illness, 7. psychiat. Res., I, I62.GILBERT, G. J., and GLASER, G. H. (I96I): On the Nervous System Integration of Water and Salt Metabolism, Arch.

Neurol. Psychiat. (Chicago), 5, I79.GJESSING, R. (1932): Beitriige zur Kenntnis der Pathophysiologie des katatonen stupors, Arch. Psychiat. Nervenkr., 96,

319.(1953): Beitrage zur Somatologie der periodischen Katatonie, Mitteilung VII, Ibid., 191, 247.

GREENE, R., and DALTON, K. (1953): The Premenstrual Syndrome, Brit. med. J., i, 1007.HOFF, H., and POTZL, 0. (1930): Untersuchungen Ober den Wasserhaushalt bei periodischen Psychosen, Z. ges. Neurol.

Psychiat., 124, 200.KARSTENS, P. (1951): tber die Beeinflussung des psychischen Zustandes Normaler durch Aufnahme und Retention

unphysiologisch grosser Wassermengen, Arch. Psychiat. Nervenkr., I86, 231.KLEIN, L., and CAREY, J. (1957): Total Exchangeable Sodium in the Menstrual Cycle, Amer. J. Obstet. Gynec., 74, 956.KLEIN, R. (1950): Clinical and Biochemical Investigations in a Manic-depressive with Short Cycles,J. ment. Sci.,96, 293.

and NUNN, R. F. (1945): Clinical and Biochemical Analysis of a Case of Manic-depressive Psychosis ShowingRegular Weekly Cycles, Ibid., 91, 79.

LANGE, J. (1928): ' Die endogenen und reaktiven Gemiitserkrankungen und die manisch-depressive Konstitution, VKorperliche Erscheinungen, in Bumke, 0. (Ed.) Handbuch der Geisteskrankheiten, Band 6, p. 52. Berlin: JuliusSpringer.

LEIBERMAN, D. M., and HOENIG, J. (I953): Photometrazol Sensitivity in Catatonic Schizophrenia, Y. Neurol. Neurosurg.Psychiat., I6, 194.

MCCANCE, R. A. (1936): Medical Problems in Mineral Metabolism, Lancet, i, 823.McFARLAND, R. A., and GOLDSTEIN, H. (1939): The Biochemistry of Manic-depressive Psychosis. A Review, Amer. J.

Psychiat., 96, 2I.MILES, B. E., and DE WARDENER, H. E. (1953): Effect of Emotion on Renal Function, Lancet, ii, 539.PFISTER, H. 0. (1937): Die neurovegetativen Storungen der Schizophrenien und ihre Beziehungen zur Insulin-,

Cardiazol-und Schlafkurbehandlung, Schweiz. Arch. Neurol. Psychiat., 39 (supplement), 77.ROWNTREE, D. W., and KAY, W. W. (1952): Clinical, Biochemical and Physiological Studies in Cases of Recurrent

Schizophrenia, 5'. ment. Sci., 98, IOO.RUSSELL, G. F. M. (I960): Body Weight and Balance of Water, Sodium and Potassium in Depressed Patients Given

Electro-convulsive Therapy, Clin. Sci., 19, 327.(1962): The Effect of Diets of Different Composition on Weight Loss, Water and Sodium Balance in Obese

Patients, Ibid., 22, 269.SCHOTTSTAEDT, W. W., GRACE, W. J., and WOLFF, H. G. (1956): Life Situations, Behaviour, Attitudes, Emotions and:

Renal Excretion of Fluid and Electrolytes, I to V, Y. psychosom. Res., I, 75, 147, 203, 287 and 292.STROM-OLSEN, R., and WEIL-MALHERBE, H. (1958): Humoral Changes in Manic-depressive Psychosis with Particular

Reference to the Excretion of Catechol Amines in Urine, Y. ment. Sci., 1O4, 696.THORVARDSSON, K. (1942): Wasserbelastungsversuche an Melancholikern, Acta psychiat. (Kbh.), 17, 89.TROLLE, C. (1945): Studies of the Water Excretion in Recovery from Manic-depressive Psychosis (Depressive Phase),

Ibid., 20, 235.WOODBURY, D. M., TIMIRAS, P. S., and VERNADAKIS, A. (1957): Influence of Adrenocortical Steroids on Brain Function

and Metabolism, in 'Hormones, Brain Function and Behaviour' (edited by H. Hoagland), p. 27. New York:Academic Press.

copyright. on June 7, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.39.447.19 on 1 January 1963. Dow

nloaded from