- Email: [email protected]
The effects of antidepressant medication on the diurnal plasma cortisol levels in depressed patients
J. Psychosomattc
Res,
1966. Vol
10. PP 197 to 202.
Pergamon Press Ltd
Printed m Northern Ireland
THE EFFECTS OF ANTIDEPRESSANT MEDICATION ON THE DIURNAL PLASMA CORTISOL LEVELS IN DEPRESSED PATIENTS* D. J. MCCLURE (Recezved 25 February 1966)
IN A PREVIOUS paper [I], a study of the dmrnal variation of plasma cortisol levels in psychotically depressed patients revealed that when these values were plotted, they formed a characteristrc diurnal plasma cortisol curve for each individual patient The present study is concerned wrth the effects on the diurnal plasma cortisol levels m the same group of depressed patients, before, during and after they had been treated with anti-depressant medication. Of the orrgmal seven cases two could not complete the study. One (No. 6) fell and fractured her left leg and was transferred to an orthopedic umt. The other (No. 7) was found to have diabetes melhtus and was removed to a general hosprtal for assessment of this conditron. The remaining five patrents completed 28 days of treatment wrth a selected antrdepressant drug. An attempt was made to carry out a control study wrth normal SubJects of comparable age and sex However, the first two controls reacted so badly to the medrcatton that this aspect of the investlgatron had to be abandoned. METHODOLOGY Clrnical Each patient m the mvesttgatton was placed on placebo three times dally and once at night for the first 48 hr During this time, bloods were drawn over two consecuttve 24-hour periods at intervals which would give most mformation about the dmrnal adrenocorttcal activity, namely, 7 00 a m , 12 noon and 10 00 p m. When these were completed, the pattents were placed on either amttnptylme (Elavtl), or lmtpramme (Toframl), 150 mg daily m divided doses and the placebo was continued at bedtime This procedure was repeated after 14 days and agam after 28 days of treatment A chmcal assessment of the depression was made by the author on each patient before treatment started and then after two weeks, and again after four weeks of the therapeutic regime. The rating was done on a 5 pomt scale as follows. 0 = absence of depression; + = mildly depressed; + + = moderately depressed, + + + = markedly depressed, + + + + = severely depressed. Scormg was of a reasonably htgh order because each pattent was interviewed every day of the investigation and the ward nursmg reports were taken into constderatron before the final assessment was made. The chntcal ratmg was done Independent of the later chemtcal estimations Bzochemml Laboratory estimations of the plasma corttsol were determined by a very shght modtficatron of the dialysis method described by Murphy et al [2] Fifteen mtlhhtres of venous blood were drawn at 6 45 a m , 11 45 a m. and 10 00 p.m on two consecutive days mto 20 ml test tubes contammg heparm, then centrifuged, and the plasma separated and frozen untd assayed This method measures both corttsol and corttcosterone m the plasma, but smce the amount of the cortrcosterone ISonly about 1/13th that of cortisol, the results obtamed here will be referred to as plasma corttsol [3] Statzstrcal The material of the five patients taken three times per day before treatment, and 14 and 28 days after treatment began, was analysed by an analysis of vanance for triple classrficatton [4] The data were then analysed by means of the 7040 IBM computer at McGill Umverstty * From the Allan Memortal Institute of Psychiatry, McG111University, Montreal. 197
198
D. J MCCLURE TABLE I-PLASMA CORTISOL LEVELS (pg/lOO ml) IWAM, 12 NOON AND 10mPM BEFORE AND
Patients
700am.
OF EACH DURING
PATIENT AT TREATMENT
12 noon
10 00 p.m
Pre-treatment
1 2 3 4 5
29 00 34 00 3100 34 00 53 00
20 50 18 75 19 25 1600 32 75
1875 1550 12 50 11 75 31 25
After 14 days treatment
1 2 3 4 5
22 25 19 75 1650 1300 43 50
14 75 1400 1200 1150 25 00
1175 9 15 6 75 8 00 16 25
After 28 days treatment
1 2 3 4 5
16 75 16 00 1575 11 25 32 00
11 25 11 25 9 50 8 50 1975
5 25 8 50 6 50 600 9 75
2 -THE
TABLE
ANALYSISOF VARIANCEFOR TRIPLE CLASSIFICATION
Sum of squares
Source
Degrees of freedom
Variance estimate
Patients (P) Treatment (r)
1489 3139 1268 3583
Time (t)
1528 0333 82 9611
8
10 3701
T X t mteraction
201 2444 118 9083
8 4
25 1556f 29 72711
P x T x t Interaction
114 9805
16
7 1863
4803 7999
44
P X Tmteractlon P X t mteraction
Total
4 2 2
372 3285* 634 17921_ 764 0167-F
* Slgmficant at 0 1 ‘A level. t Slgmficant at 1 ‘A level. $ Slgmficant at 2 5 % level
TABLE
3 -PLASMA
CORTWL
LEVELS IN THE VARIOUS PATIENTS
Age
Sex
(yrs)
pg/lOO ml*
1
F
2 3 4 5
F M M F
48 62 70 71 56
167 164 144 13 3 29 3
Patients
* The value shown IS the mean of 9 determmatlons during the entlre penod of mvestlgation.
taken
Antidepressant
medtcatton on the diurnal plasma cortisol levels m depressed patients
199
RESULTS The raw data of each of the five patients taken at 7 00 a m , 12 noon and 10 00 p m , before treatment and 14 and 28 days after treatment started, are shown m Table 1 The results of the analysis of variance are given m Table 2 and are as follows Differences between patients wrth regard to their plasma corttsol levels were highly significant (0 001 > P < 0 0005) These differences are due to the fact that patient number 5 had very s~pficantly higher values than all the other subjects (P < 0 001, Table 3). Dtfferences in plasma cortisol levels between the other four subjects were not stattstlcally stgmficant (P > 0 10, Table 3) TABLE
J.-PLASMA
BEFORE
AND
CORTISOL DURING
LEVELS
TREATMENT
Length of treatment (days)
pg/lOO ml*
0 14 28
25 2 163 12 5
* The value shown is the mean of 15 determmattons taken durmg the entire period of investigation Dtfferences between different penods m the course of treatment were also statistically stgmficant. This can be seen from Table 4 where the values before treatment started were higher than those taken 14 days after treatment began, and these m turn were htgher than those taken after 28 days of treatment t-tests, using a standard error based on the variance estimate obtained from the analysts of variance showed that the difference between the before treatment and 14 day values were htghly stgmficant (P < 0 001) whtle differences between the 14 day and 28 day periods were only of borderline significance (0 10 > P > 0 05) TABLE
5.--DIURNAL
PLASMA
VARIATION
CORTISOL
OF
LEVELS
Time of day
pg/lOO ml*
7.00 a m 12 noon 10.00 p m.
25 9 16 3 119
* The value shown 1s the mean of 1.5 determmattons taken during the entire period of mvestlgatton. Differences between values obtained at different times of the day were statisttcally stgnrfrcant (0 01 > P > 0 005). Thus, it can be seen from Table 5 that the values taken early m the mormng were higher than those taken around noon, which m turn were higher than those taken m the evening. t-tests showed that the differences between the early mormng and noon samples were highly stgruticant (P < 0 001) while those between noon and 10 00 p m were also stgmficant but less so (0 05 > P > 0 02). As for the mteractlons there was no sigmficant patient x treatment mteractton (0 30 > P > 0 20, Fig 1) The meaning of thts 1s that the plasma cortlsol levels changed m essentially the same manner m all patients during the course of treatment There was, however, stgrnficant patient x time mteractlon (0 025 > P > 0 01, Fig 2) and treatment x time mteractlon (0 025 > P > 0 01, Fig 3) The meaning of the stgmficant patient x time mteractton 1s that whereas all the patients showed a decline m the values between 7 00 a m. and noon, the values fell significantly more m the case of patient number 5 than they did m the remammg four pattents The meanmg of the treatment x ttme mteractton IS that the shape of the dmrnal curve changes m the course of treatment, the decline between 7 00 a m and noon being more sharp before treatment than it was 14 and 28 days after treatment started. It should be pointed out that because the patient x treatment interaction was not stgmficant, Its sum of squares and degrees of freedom were incorporated with those of the error term whtch m
D. J. MCCLURE
200
, __--I_ 2
E
40
3 -------25’
l--._
-.._
--._
--._
?
--__
--o._
“i”
.-.
_____.--
---_
t <-;------_____*_ -1
___
---__*
---_
B i-l1
-_-
-’
--i:_.
------- -----
l
---8
IO
I
I
I
28
14
0
Days
of
treatment
FIG 1. Changes m the plasma cortlsol levels of each patlent durmg treatment Pohent
I ----------
2
! g40
4 _----5 -----
‘\,
I
FIG. 2. Dmrnal vatxatlon of the plasma cortlsol levels of each patlent durmg the entire period of mvestlgatlon.
Duration
of
treatment
days 0 --------14
4-
I 7am
m
12 “co”
I lOpm
FIG. 3. Changes m the diurnal varlatlon of the plasma cortlsol levels durmg treatment
Antidepressant
medmatron on the dmrnal plasma cortrsol levels m depressed patients
201
this analysis was the patient X treatment X trme Interaction. This triple mteractron was used as the error term to test the patient x trme mteractron and the treatment x time mteractton, the results bemg stgmficant as mdrcated above. NOW, because the patient x trme and treatment X time mteractions were significant, they were used m appropriate places as error terms to test the stgmficance of patrents, treatment and tnne, with the results being as mdrcated above It should be noted that m testmg for the stgmficance of time or diurnal varratron, the error term used was the treatment x time mteractron whtch had the higher varumce estimate and the lower number of degrees of freedom than the patient x trme mteractton The results by this analysis were nevertheless statistically srgmficant. DISCUSSION
While the number of patients treated is small, the results show an effect which For the first two days of the mvestlgatron, no patrent appears to be consistent improved on placebo, m fact, they became worse wrth aggravation of the depressrve mood and msomma. All five patrents, however, responded favourably to the prescribed antidepressant medrcatron (Table 6). Of the five cases who completed the TABLE6 -CLINICAL RATING
Patient 1 2 3 4 5
Pre-treatment ++++ ++++ ++++ ++++ ++++
After 14 days treatment ++ ++ ++ ++ +++
After 28 days treatment 0 0 + 0 +
study, four were treated wrth amitriptyhne and one with rmipramme (No. 4). The two cases who had to drop out of the study had also been placed on rmrpramine. The effectrveness of the amrtriptyhne was clinically evident during the first week of treatment, but much more so at the end of the second week. The clinical improvement was noticeable to the medical staff and also to the patrents’ relatrves. The improvement was accompanied by a fall in the plasma cortrsol levels. This fall was greatest in the peak mornmg values after the first two weeks. After four weeks, there was a further dechne m the cortrsol values wrth the levels moving towards more normal ones. The case treated wrth lmlpramme followed a similar course. Chmcally, the depressron was reheved m three of the cases (Nos 1,2, and 4) and in the other two (Nos 3 and 5) only minor symptoms remamed, which cleared up after a further two weeks of treatment. Smce sedative medrcatlons are known to lower the hydrocortrsone values m the peripheral blood and smce the only addmonal medrcatton grven to these patients was placebo, we can conclude that the drug Influence on the plasma cortrsol levels was due entirely to the prescribed antrdepressant medlcatron. The amltrrptyhne showed a consistently good response in lowering the plasma cortrsol and the one case of lmlpramme did hkewlse. However, not all patients respond as consistently and favourably as this. For example, the five depressed patrents treated by Gibbons and McHugh [5] wrth imipramme showed different responses. One improved clmrcally and showed a dechne m plasma cortlsol values but the other four had no chmcal Improvement. Two of them showed no change in the levels of cortisol and the other two had transient falls followed by a subsequent rise to the previous levels. The effectiveness of a drug may vary with mdlvrdual patients, and the questIon frequently arrses as to whether to contmue treatment with a specific medication whrch
202
D. J. MCCLURE
clinically does not appear to be havmg the desired effect. Under these circumstances in depressed cases, it seems reasonable to suggest that if there is not a substantial fall in the plasma cortisol levels, especially the early morning values, after two weeks It 1s probably a good indication that the drug 1s not the medication of choice for that patient and should be changed. The shape of the diurnal plasma cortisol curve m the depressed patients changed durmg the course of treatment. The high pre-treatment levels fell considerably after two weeks treatment, with a further smaller decline after four weeks of treatment. The levels changed in exactly the same way m all patients so that with remission or significant improvement, the cortisol values resembled those of normal subjects which meant that the diurnal variation curves were also m the normal range. In all five patients the exacerbation of depressive mood was present at the same time as the peak plasma cortisol levels, namely 7 00 a m. This may have been a coincidental finding but since this morning exacerbation of mood disappeared when the cortlsol levels fell, it IS concluded that the high mornmg steroid values may be responsible for the diurnal variation of mood which occurs in psychotically depressed subjects or ISclosely associated with a central nervous system mechamsm which is more fundamentally responsible. SUMMARY
The diurnal plasma cortlsol levels were observed m five psychotically depressed patients before, during, and after treatment with antidepressant medication. Chnical improvement was accompanied by a fall m the cortisol values and clinical cure by the values returning to normal. The shape of all the diurnal plasma cortlsol curves changed in an identical way during the course of treatment, the high pre-investigation levels falling gradually over the four weeks of treatment. It is suggested that the plasma cortisol levels can be used as a guide to the progress of treatment and to the effectiveness of the prescribed medication m psychotic depressive illness; and further, that the cortisol itself may be responsible for the early morning exacerbation of mood which occurs m this condrtron. partrcularly wish to thank Professor R A Cleghorn, whose supervrslon and gmdance were a constant source of strmulatron throughout thrs work Dr. B Grad helped with the statrstrcal data The study was made possible by a Fellowshrp from the Medrcal Research Council of Canada
Acknowledgements-I
REFERENCES 1 MCCLURE D. J. The dmrnal varratron of plasma cortrsol levels m depressron J. Psychosom. Res. 10, 189 (1966). 2. MURPHY B. P., ENGELBERGW. and PATTEE C. J Simple method for the determmatron of plasma cortlcords. J Clm Endocr. Metabol. 23, 293 (1963) 3. EIK-NES K Methods for measurement of cortrcosterords m blood, m Hormones zn Human PIasma ed. H. N. ANTONIADES,p 358 Little, Brown, Boston, Mass (1960) 4. MCNEMAR 0 Psycholopzcal Statzstzcs, chapter 16, Analysis of variance for trmle I classlficatlon John Wiley, New’York”(l959) 5. GIBBONSJ. L and MCHUGH P R. Plasma cortlsol m depresswe illness J Psychrat. Res 1, 162 (1963).
Res,
1966. Vol
10. PP 197 to 202.
Pergamon Press Ltd
Printed m Northern Ireland
THE EFFECTS OF ANTIDEPRESSANT MEDICATION ON THE DIURNAL PLASMA CORTISOL LEVELS IN DEPRESSED PATIENTS* D. J. MCCLURE (Recezved 25 February 1966)
IN A PREVIOUS paper [I], a study of the dmrnal variation of plasma cortisol levels in psychotically depressed patients revealed that when these values were plotted, they formed a characteristrc diurnal plasma cortisol curve for each individual patient The present study is concerned wrth the effects on the diurnal plasma cortisol levels m the same group of depressed patients, before, during and after they had been treated with anti-depressant medication. Of the orrgmal seven cases two could not complete the study. One (No. 6) fell and fractured her left leg and was transferred to an orthopedic umt. The other (No. 7) was found to have diabetes melhtus and was removed to a general hosprtal for assessment of this conditron. The remaining five patrents completed 28 days of treatment wrth a selected antrdepressant drug. An attempt was made to carry out a control study wrth normal SubJects of comparable age and sex However, the first two controls reacted so badly to the medrcatton that this aspect of the investlgatron had to be abandoned. METHODOLOGY Clrnical Each patient m the mvesttgatton was placed on placebo three times dally and once at night for the first 48 hr During this time, bloods were drawn over two consecuttve 24-hour periods at intervals which would give most mformation about the dmrnal adrenocorttcal activity, namely, 7 00 a m , 12 noon and 10 00 p m. When these were completed, the pattents were placed on either amttnptylme (Elavtl), or lmtpramme (Toframl), 150 mg daily m divided doses and the placebo was continued at bedtime This procedure was repeated after 14 days and agam after 28 days of treatment A chmcal assessment of the depression was made by the author on each patient before treatment started and then after two weeks, and again after four weeks of the therapeutic regime. The rating was done on a 5 pomt scale as follows. 0 = absence of depression; + = mildly depressed; + + = moderately depressed, + + + = markedly depressed, + + + + = severely depressed. Scormg was of a reasonably htgh order because each pattent was interviewed every day of the investigation and the ward nursmg reports were taken into constderatron before the final assessment was made. The chntcal ratmg was done Independent of the later chemtcal estimations Bzochemml Laboratory estimations of the plasma corttsol were determined by a very shght modtficatron of the dialysis method described by Murphy et al [2] Fifteen mtlhhtres of venous blood were drawn at 6 45 a m , 11 45 a m. and 10 00 p.m on two consecutive days mto 20 ml test tubes contammg heparm, then centrifuged, and the plasma separated and frozen untd assayed This method measures both corttsol and corttcosterone m the plasma, but smce the amount of the cortrcosterone ISonly about 1/13th that of cortisol, the results obtamed here will be referred to as plasma corttsol [3] Statzstrcal The material of the five patients taken three times per day before treatment, and 14 and 28 days after treatment began, was analysed by an analysis of vanance for triple classrficatton [4] The data were then analysed by means of the 7040 IBM computer at McGill Umverstty * From the Allan Memortal Institute of Psychiatry, McG111University, Montreal. 197
198
D. J MCCLURE TABLE I-PLASMA CORTISOL LEVELS (pg/lOO ml) IWAM, 12 NOON AND 10mPM BEFORE AND
Patients
700am.
OF EACH DURING
PATIENT AT TREATMENT
12 noon
10 00 p.m
Pre-treatment
1 2 3 4 5
29 00 34 00 3100 34 00 53 00
20 50 18 75 19 25 1600 32 75
1875 1550 12 50 11 75 31 25
After 14 days treatment
1 2 3 4 5
22 25 19 75 1650 1300 43 50
14 75 1400 1200 1150 25 00
1175 9 15 6 75 8 00 16 25
After 28 days treatment
1 2 3 4 5
16 75 16 00 1575 11 25 32 00
11 25 11 25 9 50 8 50 1975
5 25 8 50 6 50 600 9 75
2 -THE
TABLE
ANALYSISOF VARIANCEFOR TRIPLE CLASSIFICATION
Sum of squares
Source
Degrees of freedom
Variance estimate
Patients (P) Treatment (r)
1489 3139 1268 3583
Time (t)
1528 0333 82 9611
8
10 3701
T X t mteraction
201 2444 118 9083
8 4
25 1556f 29 72711
P x T x t Interaction
114 9805
16
7 1863
4803 7999
44
P X Tmteractlon P X t mteraction
Total
4 2 2
372 3285* 634 17921_ 764 0167-F
* Slgmficant at 0 1 ‘A level. t Slgmficant at 1 ‘A level. $ Slgmficant at 2 5 % level
TABLE
3 -PLASMA
CORTWL
LEVELS IN THE VARIOUS PATIENTS
Age
Sex
(yrs)
pg/lOO ml*
1
F
2 3 4 5
F M M F
48 62 70 71 56
167 164 144 13 3 29 3
Patients
* The value shown IS the mean of 9 determmatlons during the entlre penod of mvestlgation.
taken
Antidepressant
medtcatton on the diurnal plasma cortisol levels m depressed patients
199
RESULTS The raw data of each of the five patients taken at 7 00 a m , 12 noon and 10 00 p m , before treatment and 14 and 28 days after treatment started, are shown m Table 1 The results of the analysis of variance are given m Table 2 and are as follows Differences between patients wrth regard to their plasma corttsol levels were highly significant (0 001 > P < 0 0005) These differences are due to the fact that patient number 5 had very s~pficantly higher values than all the other subjects (P < 0 001, Table 3). Dtfferences in plasma cortisol levels between the other four subjects were not stattstlcally stgmficant (P > 0 10, Table 3) TABLE
J.-PLASMA
BEFORE
AND
CORTISOL DURING
LEVELS
TREATMENT
Length of treatment (days)
pg/lOO ml*
0 14 28
25 2 163 12 5
* The value shown is the mean of 15 determmattons taken durmg the entire period of investigation Dtfferences between different penods m the course of treatment were also statistically stgmficant. This can be seen from Table 4 where the values before treatment started were higher than those taken 14 days after treatment began, and these m turn were htgher than those taken after 28 days of treatment t-tests, using a standard error based on the variance estimate obtained from the analysts of variance showed that the difference between the before treatment and 14 day values were htghly stgmficant (P < 0 001) whtle differences between the 14 day and 28 day periods were only of borderline significance (0 10 > P > 0 05) TABLE
5.--DIURNAL
PLASMA
VARIATION
CORTISOL
OF
LEVELS
Time of day
pg/lOO ml*
7.00 a m 12 noon 10.00 p m.
25 9 16 3 119
* The value shown 1s the mean of 1.5 determmattons taken during the entire period of mvestlgatton. Differences between values obtained at different times of the day were statisttcally stgnrfrcant (0 01 > P > 0 005). Thus, it can be seen from Table 5 that the values taken early m the mormng were higher than those taken around noon, which m turn were higher than those taken m the evening. t-tests showed that the differences between the early mormng and noon samples were highly stgruticant (P < 0 001) while those between noon and 10 00 p m were also stgmficant but less so (0 05 > P > 0 02). As for the mteractlons there was no sigmficant patient x treatment mteractton (0 30 > P > 0 20, Fig 1) The meaning of thts 1s that the plasma cortlsol levels changed m essentially the same manner m all patients during the course of treatment There was, however, stgrnficant patient x time mteractlon (0 025 > P > 0 01, Fig 2) and treatment x time mteractlon (0 025 > P > 0 01, Fig 3) The meaning of the stgmficant patient x time mteractton 1s that whereas all the patients showed a decline m the values between 7 00 a m. and noon, the values fell significantly more m the case of patient number 5 than they did m the remammg four pattents The meanmg of the treatment x ttme mteractton IS that the shape of the dmrnal curve changes m the course of treatment, the decline between 7 00 a m and noon being more sharp before treatment than it was 14 and 28 days after treatment started. It should be pointed out that because the patient x treatment interaction was not stgmficant, Its sum of squares and degrees of freedom were incorporated with those of the error term whtch m
D. J. MCCLURE
200
, __--I_ 2
E
40
3 -------25’
l--._
-.._
--._
--._
?
--__
--o._
“i”
.-.
_____.--
---_
t <-;------_____*_ -1
___
---__*
---_
B i-l1
-_-
-’
--i:_.
------- -----
l
---8
IO
I
I
I
28
14
0
Days
of
treatment
FIG 1. Changes m the plasma cortlsol levels of each patlent durmg treatment Pohent
I ----------
2
! g40
4 _----5 -----
‘\,
I
FIG. 2. Dmrnal vatxatlon of the plasma cortlsol levels of each patlent durmg the entire period of mvestlgatlon.
Duration
of
treatment
days 0 --------14
4-
I 7am
m
12 “co”
I lOpm
FIG. 3. Changes m the diurnal varlatlon of the plasma cortlsol levels durmg treatment
Antidepressant
medmatron on the dmrnal plasma cortrsol levels m depressed patients
201
this analysis was the patient X treatment X trme Interaction. This triple mteractron was used as the error term to test the patient x trme mteractron and the treatment x time mteractton, the results bemg stgmficant as mdrcated above. NOW, because the patient x trme and treatment X time mteractions were significant, they were used m appropriate places as error terms to test the stgmficance of patrents, treatment and tnne, with the results being as mdrcated above It should be noted that m testmg for the stgmficance of time or diurnal varratron, the error term used was the treatment x time mteractron whtch had the higher varumce estimate and the lower number of degrees of freedom than the patient x trme mteractton The results by this analysis were nevertheless statistically srgmficant. DISCUSSION
While the number of patients treated is small, the results show an effect which For the first two days of the mvestlgatron, no patrent appears to be consistent improved on placebo, m fact, they became worse wrth aggravation of the depressrve mood and msomma. All five patrents, however, responded favourably to the prescribed antidepressant medrcatron (Table 6). Of the five cases who completed the TABLE6 -CLINICAL RATING
Patient 1 2 3 4 5
Pre-treatment ++++ ++++ ++++ ++++ ++++
After 14 days treatment ++ ++ ++ ++ +++
After 28 days treatment 0 0 + 0 +
study, four were treated wrth amitriptyhne and one with rmipramme (No. 4). The two cases who had to drop out of the study had also been placed on rmrpramine. The effectrveness of the amrtriptyhne was clinically evident during the first week of treatment, but much more so at the end of the second week. The clinical improvement was noticeable to the medical staff and also to the patrents’ relatrves. The improvement was accompanied by a fall in the plasma cortrsol levels. This fall was greatest in the peak mornmg values after the first two weeks. After four weeks, there was a further dechne m the cortrsol values wrth the levels moving towards more normal ones. The case treated wrth lmlpramme followed a similar course. Chmcally, the depressron was reheved m three of the cases (Nos 1,2, and 4) and in the other two (Nos 3 and 5) only minor symptoms remamed, which cleared up after a further two weeks of treatment. Smce sedative medrcatlons are known to lower the hydrocortrsone values m the peripheral blood and smce the only addmonal medrcatton grven to these patients was placebo, we can conclude that the drug Influence on the plasma cortrsol levels was due entirely to the prescribed antrdepressant medlcatron. The amltrrptyhne showed a consistently good response in lowering the plasma cortrsol and the one case of lmlpramme did hkewlse. However, not all patients respond as consistently and favourably as this. For example, the five depressed patrents treated by Gibbons and McHugh [5] wrth imipramme showed different responses. One improved clmrcally and showed a dechne m plasma cortlsol values but the other four had no chmcal Improvement. Two of them showed no change in the levels of cortisol and the other two had transient falls followed by a subsequent rise to the previous levels. The effectiveness of a drug may vary with mdlvrdual patients, and the questIon frequently arrses as to whether to contmue treatment with a specific medication whrch
202
D. J. MCCLURE
clinically does not appear to be havmg the desired effect. Under these circumstances in depressed cases, it seems reasonable to suggest that if there is not a substantial fall in the plasma cortisol levels, especially the early morning values, after two weeks It 1s probably a good indication that the drug 1s not the medication of choice for that patient and should be changed. The shape of the diurnal plasma cortisol curve m the depressed patients changed durmg the course of treatment. The high pre-treatment levels fell considerably after two weeks treatment, with a further smaller decline after four weeks of treatment. The levels changed in exactly the same way m all patients so that with remission or significant improvement, the cortisol values resembled those of normal subjects which meant that the diurnal variation curves were also m the normal range. In all five patients the exacerbation of depressive mood was present at the same time as the peak plasma cortisol levels, namely 7 00 a m. This may have been a coincidental finding but since this morning exacerbation of mood disappeared when the cortlsol levels fell, it IS concluded that the high mornmg steroid values may be responsible for the diurnal variation of mood which occurs in psychotically depressed subjects or ISclosely associated with a central nervous system mechamsm which is more fundamentally responsible. SUMMARY
The diurnal plasma cortlsol levels were observed m five psychotically depressed patients before, during, and after treatment with antidepressant medication. Chnical improvement was accompanied by a fall m the cortisol values and clinical cure by the values returning to normal. The shape of all the diurnal plasma cortlsol curves changed in an identical way during the course of treatment, the high pre-investigation levels falling gradually over the four weeks of treatment. It is suggested that the plasma cortisol levels can be used as a guide to the progress of treatment and to the effectiveness of the prescribed medication m psychotic depressive illness; and further, that the cortisol itself may be responsible for the early morning exacerbation of mood which occurs m this condrtron. partrcularly wish to thank Professor R A Cleghorn, whose supervrslon and gmdance were a constant source of strmulatron throughout thrs work Dr. B Grad helped with the statrstrcal data The study was made possible by a Fellowshrp from the Medrcal Research Council of Canada
Acknowledgements-I
REFERENCES 1 MCCLURE D. J. The dmrnal varratron of plasma cortrsol levels m depressron J. Psychosom. Res. 10, 189 (1966). 2. MURPHY B. P., ENGELBERGW. and PATTEE C. J Simple method for the determmatron of plasma cortlcords. J Clm Endocr. Metabol. 23, 293 (1963) 3. EIK-NES K Methods for measurement of cortrcosterords m blood, m Hormones zn Human PIasma ed. H. N. ANTONIADES,p 358 Little, Brown, Boston, Mass (1960) 4. MCNEMAR 0 Psycholopzcal Statzstzcs, chapter 16, Analysis of variance for trmle I classlficatlon John Wiley, New’York”(l959) 5. GIBBONSJ. L and MCHUGH P R. Plasma cortlsol m depresswe illness J Psychrat. Res 1, 162 (1963).