Biological Psychiatry
Original Paper Neuropsychobiology 1992:25:130-133
Department of Psychiatry, University of Oklahoma Medical School, Oklahoma City, Okla., USA
Electroconvulsive Therapy-Induced Cortisol Release after Dexamethasone in Depression
Key Words
Abstract
Electroconvulsive therapy Cortisol Depression Dexamethasone
ECT-induced cortisol release was distinctly seen, and fell along a course of ECT in each of 12 inpatient male melancholics (p = 0.00024, binomial), with dexamethasone given to diminish the elevated baseline cortisol levels typically seen in depression. Cortisol release dropped on average by 55% (p = 0.015), from 16.6 ± 6.8 |ig/dl (p = 0.000002) with the first ECT to 8.0 ± 7 .7 pg/dl (p = 0.000003) after 6 or more ECTs. The fall along the course was larger with unilateral ECT than bilateral ECT (p = 0.042), although significant regardless of electrode placement, suggesting that unilateral ECT tends to lose thera peutic impact along a course in comparison to bilateral ECT.
Introduction
Because the brief elevations of hormone concentra tions in the blood after electroconvulsive therapy (ECT) are readily measurable consequences of a CNS event [Mitchell et al., 1990; Swartz. 1985b], they have been measured as potential reflections of psychiatric illness in over a hundred reported studies. Serum prolactin levels have been most frequently reported; they typically rise 400-800% over baseline after ECT [Swartz, 1985b], In comparison, serum cortisol levels of depressives rise only 0-50% after ECT [Deakin et al., 1983; Arato et al., 1980; Mitchell el al., 1990], The marginal elevation of serum cortisol has presumably resulted from the tendency of depression to elevate baseline serum cortisol levels and deplete pituitary ACTH [Gold and Chrousos, 1985].
The only study [Swartz and Chen, 1985] to observe ECT-induced cortisol release in the 400-800% range gave dexamethasone before each measured ECT with the in tention of decreasing baseline cortisol levels and allowing replenishment of pituitary ACTH stores. That study found a 52% decrease of ECT-induced cortisol release along the course of treatment with bilateral ECT; 10 of the 11 ECT responders showed such decrease, and the 1 non responder did not. There have been no other reports of ECT-induced cortisol release after dexamethasone, al though reports of marginal cortisol elevations without dexamethasone have continued [Turner et al., 1987; Whalley et al., 1987; Weizman et al., 1987; Widerlôv et al.. 1989; Mitchell etal.. 1990], This study of ECT-induced cortisol release after dexa methasone and its fall with treatment was conducted to
Dr. Conrad Swart/ Department o f Psychiatric Medicine East Carolina University School o f Medicine Greenville, NC 27858-4354 (USA)
© 1992 S. Kargcr AG. Basel 0302—282X/92/0253— 0130 $2.75/0
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
Conrad M. Swartz
Methods We prospectively studied 12 consecutive melancholic male inpa tients selected to receive ECT according to the clinical judgment of their physicians during a 9-month period in an academic veterans hospital. We required absence of exposure to drugs known to affect cortisol levels (e.g.. anticonvulsants, ethanol, sedative hypnotics) for at least 3 weeks, no recent trauma or physical illness, and no known brain lesions. Subject ages ranged from 47 to 75 years. All were phys ical! healthy, with no recent trauma or illness. Informed consent was obtained. The protocol was virtually identical to our initial study [Swartz and Chen, 1985]. Symptoms were verified by checklist examination prior to ECT for all DSM-11I-R [American Psychiatric Association, 1987] criteria for major depression. Each patient suffered severe dys phoria or apathy, distinct recent deterioration, inadequate sleep, prominent motor retardation or agitation, anhedonia. fatigue, and low self-esteem, without concurrent organic mental disorder. For diagnostic specificity, depression was required for at least 4 weeks, rather than the 2 weeks o f DSM-III-R. Each subject had failed to respond to antidepressants, 8 had suicidal thoughts, and 7 had psy chotic features. To identify large response from ECT each subject was prospectively rated on a 4-point global severity scale (0 = entirely w'ell, 1 = marginal or questionable, 2 = definite. 3 = severe) before and after the ECT course, with a large difference understood between rat ings of 1 and 2. Pretreatment ratings were required to be at least 2, and ECT response was regarded as decrease to a rating of 0 or I. An interviewer-conducted checklist rating equivalent to the Hamilton rating was conducted pre- and post-ECT course [Carroll et al.. 1981], to confirm this global rating, w'hich it did. Subjects remained medication-free throughout their ECT course, except for anesthesia in the usual manner [American Psychiatric Association. 1990]. with intravenous glycopyrrolate (0.0044 mg/kg. maximum 0.4 mg), methohexital 50-90 mg, and succinylcholine 3090 mg. and hyperoxygenation by mask throughout. Stimuli had a charge of 378 mC (140 pulse pairs/s. 0.9 amp. 1 ms pulse width for 3 s), within 10% of our prior study and about 2.6 times average male seizure threshold [Sackeim ct al.. 1987], Judging by the cuffed-arm method [Fink and Johnson. 1982]. the duration of each ECT seizure was at least 25 s. Patients received 3 ECT sessions per week for a total of 6-13 ECTs (mean 8.2, SD 2.6), concluding with either remission or failure to progress, excepting 1 who stopped against medical advice after 6 sessions, without remission. Electrode placement was assigned randomly. Bilateral frontotemporal electrode placement was used in 4 subjects, and unilateral frontotemporal-vertex place-
Table 1. Post-ECT serum cortisol levels and elevations over base
line (pg/dl) Subject No.
1 2 3 4 5 6 7 8 9 I0 1 11 12
PLCMT
B U U U B U B U B U U U
First ECT
Last ECT with same PLCMT
level
elevation
level
elevation
15.7 26.5 34.1 17.1 17.4 5.1 22.8 30.3 24.0 9.1 15.5 19.8
13.3 20.4 20.7 15.0 16.2 3.3 20.2 28.1 22.3 7.0 14.0 ¡8.1
13.7 8.2 20.9 1.1 17.2 2.7 12.4 2.1 23.5 2.3 3.9 17.9
11.4 5.6 18.4 0.3 15.2 0.6 10.3 0.4 21.5 0.6 1.2 10.7
U = Unilateral: B = bilateral electrode placement. 1 Nonresponder.
ment in 8 [d’Elia and Raotma, 1975]. but changed for clinical pur poses to bilateral placement when more than 6 treatments were needed. Nine hours before the first ECT 2 mg dexamethasone was given. Blood was drawn before ECT medications and again 30 min after the end of the ECT seizure, the time of peak serum cortisol [Mitchell et al., 1990]. This procedure was repeated with the last ECT given with the same electrode placement (except subject No. 3 started unilateral but final measurement was not taken until after change to bilateral: his data weakened statistical differences between unilateral and bilat eral and went against the hypothesis). Serum cortisol was analyzed by radioimmunoassay kit (Diagnostic Products. Los Angeles, interas say coefficient of variation 8.4%) by technicians blind to sample identity.
Results
Elevations of serum cortisol concentrations and postECT levels are shown in table 1. Eleven subjects achieved remission: 7 had global rating 0. and 4 were rated 1: prior to ECT there were 8 ratings of 2 and 4 ratings of 3. The single nonresponder showed only small cortisol eleva tions, which did not affect the pattern of results. Every patient showed an increase in serum cortisol level with the first ECT. on average 16.6 ± SD 6.8 pg/dl and 750% over baseline (range 150-1,350%, p = 0.000002 1-tailed, t = 8.5, df = 11). With the last measured treatment cortisol levels rose on average 8.0 ± 7.7 pg/dl. 380% over base
131
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
test two expectations. The first is that these phenomena occur reliably enough to appear in a second, independent subject group. The second is that the change in cortisol release along the course differs between unilateral ECT and bilateral ECT: testing this requires comparison of patients receiving unilateral ECT with others receiving bilateral ECT. In contrast, comparison of the amount of cortisol release is best accomplished by giving the same patients both treatments alternately, to minimize con founding effects from interindividual variability.
Discussion
These observations confirm initial findings that ECT elevates the serum cortisol concentration by multiples of the baseline after dexamethasone pretreatment, and that ECT-induced cortisol decreases along a course of ECT by about 50%, on average. This decrease resembles the fall of DST cortisol levels seen with response to drug treatment by nonsuppressor depressives [Greden et al„ 1983]. It supports the view that ECT resembles pharmacotherapy by abating hypercortisolism in depressives and that con flicting observations [Fink et al., 1987] were affected by ECT-induced acceleration of elimination of 1-mg doses of dexamethasone [Swartz and Saheba. 1990]. Conversely, acceleration of dexamethasone elimination cannot ex plain the present results because it would generate in creasing rather than decreasing cortisol levels. The inter individual variations in cortisol elevation seen, although wide, are of the same magnitude as variations in ECTinduced prolactin release [Swartz. 1985a] and DST corti sol levels in depressives [Greden et al., 1983], ECT-induced cortisol release presumably reflects both physiological impact of the ECT seizure and state of ill ness. The fall of ECT-induced cortisol release indepen dent of electrode placement presumably reflects clinical improvement. Because the physiological and therapeutic impact of the seizure depend on electrode placement [Sackeim et al.. 1987; Swartz and Abrams, 1984; Swartz and Larson, 1986] and stimulus dose relative to seizure threshold for unilateral ECT [Abrams et al., 1991; Abrams and Swartz, 1985], it is reasonable to expect stim
132
Swartz
ulus placement and dose to affect cortisol release. In the present study, cortisol release with unilateral ECT was as vigorous as with bilateral ECT with the first but not the last treatment - it dropped near zero in 5 of 8 subjects after unilateral ECT but remained over 10 pg/dl in all 4 bilateral ECT subjects. This difference between unilateral and bilateral ECT might be explained by treatment method or state of illness. The former possibility is that the physiological impact fell rapidly with unilateral but not bilateral ECT. The latter possibility is that clinical improvement was achieved faster with unilateral than bilateral ECT; this is unlikely because unilateral ECT clearly is not more effective than bilateral ECT with the stimulus used in this study [Abrams et al., 1991], Therefore, the results provide evidence that unilateral ECT tends to lose physiological impact along a course of treatment, in comparison to bilateral ECT. This can be understood through increases in seizure threshold along a course of ECT. Such increases progressively narrow the difference between an unchanging ECT stimulus (as used in this study) and seizure threshold, and such narrowing decreases the impact of unilateral ECT more than of bilat eral ECT [Sackeim et al.. 1987], The results imply that preservation of the physiological impact of unilateral ECT tends to require that the stimulus dose be raised sub stantially along the course, to exceed 378 mC by the 6th treatment in patients similar to our subjects, and that a switch to bilateral ECT should be considered if the rate of improvement becomes undesirably slow. These implica tions are consistent with typical clinical guidelines [Amer ican Psychiatric Association, 1990], If ECT did not affect cortisol release more than sham ECT, ECT method would similarly not affect cortisol release; therefore, the results indicate that ECT releases more cortisol than sham ECT or anesthesia alone. Although DST cortisol levels do not fall with response to drug treatment in DST suppressors [Greden et al., 1983], ECT-induced cortisol release does. DST results were indicated by pre-ECT baseline cortisol levels, which were 8 a.m. measurements on a 2-mg DST. Ten of the 12 subjects showed suppression, with DST cortisol levels below 3.0 pg/dl; the others were 6.1 and 13.4 pg/dl. For DST suppressors, ECT-induced cortisol release decreased by 8.5 ± 8.3 pg/dl (p = 0.010 2-tailed, t = 3.23, d.f. = 9), the same as the two DST nonsuppressors. Our initial study [Swartz and Chen. 1985] found the same pattern (t = 2.2, d.f. = 9. p = 0.028), although it was not com mented on. ECT-induced cortisol release appears to be a conse quence of serial liberation of corticotropin releasing hor-
ECT-induced Cortisol
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
line (range 20-1,075%, p = 0.000003 1-tailed, t = 8.0. d.f. = 11). ECT-induced cortisol elevation decreased 51.5% between first and last ECTs (8.5 ± 8.0 pg/dl. p = 0.015 repeated measures ANOVA, placement covaried). Each subject showed greater ECT-induced cortisol elevation with the first ECT (p = 0.00024, binomial). Likewise, ECT-induced cortisol elevation decreased (11.1 ± 8 .4 pg/dl, p = 0.0035 1-tailed, t = 3.76. d.f. = 7; p = 0.0039 binomial) in the subset of 8 unilateral ECT sub jects. This decrease exceeded (p = 0.042, Mann-Whitney for skew) that for the 4 bilateral ECT subjects (3.4 ± 4.4 pg/dl) by 225%. Correspondingly. last-ECT cortisol release was lower (p = 0.041, Mann-Whitney) for unilat eral treatment (4.7 ± 6.6 pg/dl) than bilateral (14.6 ± 5.1 pg/dl). Age of unilateral ECT patients (61 ± 7.9) was not substantially different (t = 1.28. d.f. = 10. p = 0.23) from bilateral ECT patients (66.7 ± 5.9).
mone (CRH) and ACTH [Gold and Chrousos, 1985: Smoak et al., 1991], Because ECT does not change CRHinduced cortisol release [Dored et al., 1990], the clear fall of ECT-induced cortisol release suggests that the sensitiv ity of CRH release to ECT falls with treatment. Because plasma CRH concentrations do not reliably correspond to hypothalmic CRH release [Sasaki et al., 1987], direct plasma CRH measurements are not conclusive [Widerlov et al., 1989]. Ultimately, measurements of ACTH and cortisol release sensitivities coordinated with measure
ments of ECT-induced ACTH and cortisol should pro vide direct evaluation of the sensitivity of CRH release to ECT and further elucidation ofHPA axis dysrégulation in depression.
Acknowledgment Neeta C. Saheba. MD assisted with the protocol, which was con ducted at the North Chicago Veterans Affairs Medical Center.
References Fink M. Johnson L: Monitoring the duration of electroconvulsive therapy seizure. Arch Gen Psychiatry 1982:39:1189-1191. Fink M, Gujavarty K. Greenberg L: Serial déxa méthasone suppression tests and clinical out come in ECT. Convulsive Thcr 1987:3:111-
120. Gold PW. Chrousos G P: Clinical studies with cor ticotropin releasing factor: Implications for the diagnosis and pathophysiology of depression. Cushing’s disease, and adrenal insufficiency. Psychoneurocndocrinology 1985; 10:401- 4 19. Greden JF, Gardner R, King D, Grunhaus L, Carroll BJ. Kronfol Z: Dexamethasone suppres sion tests in antidepressant treatment of melan cholia. Arch Gen Psychiatre 1983:40:493— 500. Mitchell P. Smythe G. Torda T: Effect of the anes thetic agent propofol on hormonal responses to ECT. Biol Psychiatry 1990;28:315-324. Sackeim HA, Decina P, Portnoy S, Kanzler M. Kerr B, Malitz S: Effects of electrode place ment on the efficacy of titrated low-dosage ECT. Am J Psychiatry 1987;144:1449-1455. Sasaki A, Sato S, Murakami O, Go M, Inoue M. Shimizu Y, Hanew' K, Andoh N, Sato I, Sasano N, Yoshinaga K: Immunoreactive corticotro pin-releasing hormone present in human plasma may be derived from both hypotha lamic and extrahypothalamic sources. J Clin Endocrinol Metab 1987;65:176-182. Smoak B, Deuster P, Rabin D, Chrousos G: Corti cotropin-releasing hormone is not the sole fac tor mediating exercise-induced adrenocorlicotropin release in humans. J Clin Endocrinol Metab 1991;73:302-306. Swartz CM: The time course of post-ECT prolactin levels. Convulsive Ther 1985a; 1:81 -89.
Swartz CM: Characterization of the total amount of prolactin released by electroconvulsive ther apy. Convulsive Ther 1985b: 1:252-257. Swartz CM, Abrams R: Prolactin levels after bilat eral and unilateral ECT. Br J Psychiatry 1984: 144:643-645. Swartz CM, Chen JJ: Electroconvulsive therapvinduccd cortisol release: Changes with depres sive state. Convulsive Ther 1985; 1:15-21. Swartz CM, Saheba NC: Dose effect on dexametha sone suppression testing with ECT. Ann Clin Psychiatry 1990;2:155-160. Sw artz CM, [.arson G: Generalization of the effects of unilateral and bilateral ECT. Am J Psychia try 1986:143:1040-1041. Turner TH, Ur E, Grossman A: Naloxone has no effect on hormonal responses to ECT in man. Psychiatry Res 1987;22:207-212. Weizman A, Gil-Ad 1. Grupper D. Tyano S, Laron Z: The effect of acute and repeated electrocon vulsive treatments on plasma beta-endorphin, growth hormone, prolactin and cortisol secre tion in depressed patients. Psychopharmacol ogy 1987;93:122-126. Whalley U . Eagles JM. Bowler GM, Bennie JG. Dick HR. McGuire RJ. Fink G: Selective ef fects of ECT on hypothalamic-pituitary activi ty. Psychol Med 1987;17:319-328. Widerlov E. Ekman R. Jensen L. Borglund L, Ny man K: Arginine vasopressin, but not cortico tropin releasing factor, is a potent stimulator of adrenocorticolropin hormone following elec troconvulsive therapy. J Neural Transm 1989; 75:131-109.
133
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
Abrams R. Swartz CM. Vcdak C: Antidepressant effects of high-dose right unilateral ECT. Arch Gen Psychiatry 1991:48:746-748. Abrams R. Swartz CN: Electroconvulsive therapy and prolactin release: Effect of stimulus param eters. Convulsive Thcr 1985; 1:115—119. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, cd 3 rev. Washington, American Psychiatric Asso ciation. 1987. American Psychiatric Association: The Practice of Electroconvulsive Therapy: Recommenda tions for Treatment. Training, and Privileging. Washington, American Psychiatric Associa tion. 1990. Arato M. Szontagh L, Tclegdy G: Effect of electro convulsive treatment on brain monoamine content, pituitary-adrenal function and plasma prolactin. Acts Physiol Acad Sci Hung 1980: 56:163-171. Carroll BJ. Feinberg M. Smouse P. Rawson SG. Grcden JF: The Carroll rating scale for depres sion. I. Development, reliability and validity. Br J Psychiatry 1981; 138:194-200. Deakin JFW. Ferrier IN, Crow TJ. Johnstone EC, Lawler P: Effects of ECT on pituitary hormone release: Relationship to seizure, clinical vari ables and outcome. Br J Psychiatry 1983:143: 618-624. d'Elia G, Raotma H: Is unilateral ECT less effective than bilateral ECT? Br J Psychiatrv 1975:126: 83-89. Dored G, Stcfansson S, d ’Elia G, Kigedal B. Karlberg E, Ekman R: Corticotropin, cortisol, and beta-endorphin responses to the human corti cotropin-releasing hormone during melan cholia and after unilateral electroconvulsive therapy. Acta Psychiatr Scand 1990:82:204209.
Original Paper Neuropsychobiology 1992:25:130-133
Department of Psychiatry, University of Oklahoma Medical School, Oklahoma City, Okla., USA
Electroconvulsive Therapy-Induced Cortisol Release after Dexamethasone in Depression
Key Words
Abstract
Electroconvulsive therapy Cortisol Depression Dexamethasone
ECT-induced cortisol release was distinctly seen, and fell along a course of ECT in each of 12 inpatient male melancholics (p = 0.00024, binomial), with dexamethasone given to diminish the elevated baseline cortisol levels typically seen in depression. Cortisol release dropped on average by 55% (p = 0.015), from 16.6 ± 6.8 |ig/dl (p = 0.000002) with the first ECT to 8.0 ± 7 .7 pg/dl (p = 0.000003) after 6 or more ECTs. The fall along the course was larger with unilateral ECT than bilateral ECT (p = 0.042), although significant regardless of electrode placement, suggesting that unilateral ECT tends to lose thera peutic impact along a course in comparison to bilateral ECT.
Introduction
Because the brief elevations of hormone concentra tions in the blood after electroconvulsive therapy (ECT) are readily measurable consequences of a CNS event [Mitchell et al., 1990; Swartz. 1985b], they have been measured as potential reflections of psychiatric illness in over a hundred reported studies. Serum prolactin levels have been most frequently reported; they typically rise 400-800% over baseline after ECT [Swartz, 1985b], In comparison, serum cortisol levels of depressives rise only 0-50% after ECT [Deakin et al., 1983; Arato et al., 1980; Mitchell el al., 1990], The marginal elevation of serum cortisol has presumably resulted from the tendency of depression to elevate baseline serum cortisol levels and deplete pituitary ACTH [Gold and Chrousos, 1985].
The only study [Swartz and Chen, 1985] to observe ECT-induced cortisol release in the 400-800% range gave dexamethasone before each measured ECT with the in tention of decreasing baseline cortisol levels and allowing replenishment of pituitary ACTH stores. That study found a 52% decrease of ECT-induced cortisol release along the course of treatment with bilateral ECT; 10 of the 11 ECT responders showed such decrease, and the 1 non responder did not. There have been no other reports of ECT-induced cortisol release after dexamethasone, al though reports of marginal cortisol elevations without dexamethasone have continued [Turner et al., 1987; Whalley et al., 1987; Weizman et al., 1987; Widerlôv et al.. 1989; Mitchell etal.. 1990], This study of ECT-induced cortisol release after dexa methasone and its fall with treatment was conducted to
Dr. Conrad Swart/ Department o f Psychiatric Medicine East Carolina University School o f Medicine Greenville, NC 27858-4354 (USA)
© 1992 S. Kargcr AG. Basel 0302—282X/92/0253— 0130 $2.75/0
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
Conrad M. Swartz
Methods We prospectively studied 12 consecutive melancholic male inpa tients selected to receive ECT according to the clinical judgment of their physicians during a 9-month period in an academic veterans hospital. We required absence of exposure to drugs known to affect cortisol levels (e.g.. anticonvulsants, ethanol, sedative hypnotics) for at least 3 weeks, no recent trauma or physical illness, and no known brain lesions. Subject ages ranged from 47 to 75 years. All were phys ical! healthy, with no recent trauma or illness. Informed consent was obtained. The protocol was virtually identical to our initial study [Swartz and Chen, 1985]. Symptoms were verified by checklist examination prior to ECT for all DSM-11I-R [American Psychiatric Association, 1987] criteria for major depression. Each patient suffered severe dys phoria or apathy, distinct recent deterioration, inadequate sleep, prominent motor retardation or agitation, anhedonia. fatigue, and low self-esteem, without concurrent organic mental disorder. For diagnostic specificity, depression was required for at least 4 weeks, rather than the 2 weeks o f DSM-III-R. Each subject had failed to respond to antidepressants, 8 had suicidal thoughts, and 7 had psy chotic features. To identify large response from ECT each subject was prospectively rated on a 4-point global severity scale (0 = entirely w'ell, 1 = marginal or questionable, 2 = definite. 3 = severe) before and after the ECT course, with a large difference understood between rat ings of 1 and 2. Pretreatment ratings were required to be at least 2, and ECT response was regarded as decrease to a rating of 0 or I. An interviewer-conducted checklist rating equivalent to the Hamilton rating was conducted pre- and post-ECT course [Carroll et al.. 1981], to confirm this global rating, w'hich it did. Subjects remained medication-free throughout their ECT course, except for anesthesia in the usual manner [American Psychiatric Association. 1990]. with intravenous glycopyrrolate (0.0044 mg/kg. maximum 0.4 mg), methohexital 50-90 mg, and succinylcholine 3090 mg. and hyperoxygenation by mask throughout. Stimuli had a charge of 378 mC (140 pulse pairs/s. 0.9 amp. 1 ms pulse width for 3 s), within 10% of our prior study and about 2.6 times average male seizure threshold [Sackeim ct al.. 1987], Judging by the cuffed-arm method [Fink and Johnson. 1982]. the duration of each ECT seizure was at least 25 s. Patients received 3 ECT sessions per week for a total of 6-13 ECTs (mean 8.2, SD 2.6), concluding with either remission or failure to progress, excepting 1 who stopped against medical advice after 6 sessions, without remission. Electrode placement was assigned randomly. Bilateral frontotemporal electrode placement was used in 4 subjects, and unilateral frontotemporal-vertex place-
Table 1. Post-ECT serum cortisol levels and elevations over base
line (pg/dl) Subject No.
1 2 3 4 5 6 7 8 9 I0 1 11 12
PLCMT
B U U U B U B U B U U U
First ECT
Last ECT with same PLCMT
level
elevation
level
elevation
15.7 26.5 34.1 17.1 17.4 5.1 22.8 30.3 24.0 9.1 15.5 19.8
13.3 20.4 20.7 15.0 16.2 3.3 20.2 28.1 22.3 7.0 14.0 ¡8.1
13.7 8.2 20.9 1.1 17.2 2.7 12.4 2.1 23.5 2.3 3.9 17.9
11.4 5.6 18.4 0.3 15.2 0.6 10.3 0.4 21.5 0.6 1.2 10.7
U = Unilateral: B = bilateral electrode placement. 1 Nonresponder.
ment in 8 [d’Elia and Raotma, 1975]. but changed for clinical pur poses to bilateral placement when more than 6 treatments were needed. Nine hours before the first ECT 2 mg dexamethasone was given. Blood was drawn before ECT medications and again 30 min after the end of the ECT seizure, the time of peak serum cortisol [Mitchell et al., 1990]. This procedure was repeated with the last ECT given with the same electrode placement (except subject No. 3 started unilateral but final measurement was not taken until after change to bilateral: his data weakened statistical differences between unilateral and bilat eral and went against the hypothesis). Serum cortisol was analyzed by radioimmunoassay kit (Diagnostic Products. Los Angeles, interas say coefficient of variation 8.4%) by technicians blind to sample identity.
Results
Elevations of serum cortisol concentrations and postECT levels are shown in table 1. Eleven subjects achieved remission: 7 had global rating 0. and 4 were rated 1: prior to ECT there were 8 ratings of 2 and 4 ratings of 3. The single nonresponder showed only small cortisol eleva tions, which did not affect the pattern of results. Every patient showed an increase in serum cortisol level with the first ECT. on average 16.6 ± SD 6.8 pg/dl and 750% over baseline (range 150-1,350%, p = 0.000002 1-tailed, t = 8.5, df = 11). With the last measured treatment cortisol levels rose on average 8.0 ± 7.7 pg/dl. 380% over base
131
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
test two expectations. The first is that these phenomena occur reliably enough to appear in a second, independent subject group. The second is that the change in cortisol release along the course differs between unilateral ECT and bilateral ECT: testing this requires comparison of patients receiving unilateral ECT with others receiving bilateral ECT. In contrast, comparison of the amount of cortisol release is best accomplished by giving the same patients both treatments alternately, to minimize con founding effects from interindividual variability.
Discussion
These observations confirm initial findings that ECT elevates the serum cortisol concentration by multiples of the baseline after dexamethasone pretreatment, and that ECT-induced cortisol decreases along a course of ECT by about 50%, on average. This decrease resembles the fall of DST cortisol levels seen with response to drug treatment by nonsuppressor depressives [Greden et al„ 1983]. It supports the view that ECT resembles pharmacotherapy by abating hypercortisolism in depressives and that con flicting observations [Fink et al., 1987] were affected by ECT-induced acceleration of elimination of 1-mg doses of dexamethasone [Swartz and Saheba. 1990]. Conversely, acceleration of dexamethasone elimination cannot ex plain the present results because it would generate in creasing rather than decreasing cortisol levels. The inter individual variations in cortisol elevation seen, although wide, are of the same magnitude as variations in ECTinduced prolactin release [Swartz. 1985a] and DST corti sol levels in depressives [Greden et al., 1983], ECT-induced cortisol release presumably reflects both physiological impact of the ECT seizure and state of ill ness. The fall of ECT-induced cortisol release indepen dent of electrode placement presumably reflects clinical improvement. Because the physiological and therapeutic impact of the seizure depend on electrode placement [Sackeim et al.. 1987; Swartz and Abrams, 1984; Swartz and Larson, 1986] and stimulus dose relative to seizure threshold for unilateral ECT [Abrams et al., 1991; Abrams and Swartz, 1985], it is reasonable to expect stim
132
Swartz
ulus placement and dose to affect cortisol release. In the present study, cortisol release with unilateral ECT was as vigorous as with bilateral ECT with the first but not the last treatment - it dropped near zero in 5 of 8 subjects after unilateral ECT but remained over 10 pg/dl in all 4 bilateral ECT subjects. This difference between unilateral and bilateral ECT might be explained by treatment method or state of illness. The former possibility is that the physiological impact fell rapidly with unilateral but not bilateral ECT. The latter possibility is that clinical improvement was achieved faster with unilateral than bilateral ECT; this is unlikely because unilateral ECT clearly is not more effective than bilateral ECT with the stimulus used in this study [Abrams et al., 1991], Therefore, the results provide evidence that unilateral ECT tends to lose physiological impact along a course of treatment, in comparison to bilateral ECT. This can be understood through increases in seizure threshold along a course of ECT. Such increases progressively narrow the difference between an unchanging ECT stimulus (as used in this study) and seizure threshold, and such narrowing decreases the impact of unilateral ECT more than of bilat eral ECT [Sackeim et al.. 1987], The results imply that preservation of the physiological impact of unilateral ECT tends to require that the stimulus dose be raised sub stantially along the course, to exceed 378 mC by the 6th treatment in patients similar to our subjects, and that a switch to bilateral ECT should be considered if the rate of improvement becomes undesirably slow. These implica tions are consistent with typical clinical guidelines [Amer ican Psychiatric Association, 1990], If ECT did not affect cortisol release more than sham ECT, ECT method would similarly not affect cortisol release; therefore, the results indicate that ECT releases more cortisol than sham ECT or anesthesia alone. Although DST cortisol levels do not fall with response to drug treatment in DST suppressors [Greden et al., 1983], ECT-induced cortisol release does. DST results were indicated by pre-ECT baseline cortisol levels, which were 8 a.m. measurements on a 2-mg DST. Ten of the 12 subjects showed suppression, with DST cortisol levels below 3.0 pg/dl; the others were 6.1 and 13.4 pg/dl. For DST suppressors, ECT-induced cortisol release decreased by 8.5 ± 8.3 pg/dl (p = 0.010 2-tailed, t = 3.23, d.f. = 9), the same as the two DST nonsuppressors. Our initial study [Swartz and Chen. 1985] found the same pattern (t = 2.2, d.f. = 9. p = 0.028), although it was not com mented on. ECT-induced cortisol release appears to be a conse quence of serial liberation of corticotropin releasing hor-
ECT-induced Cortisol
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
line (range 20-1,075%, p = 0.000003 1-tailed, t = 8.0. d.f. = 11). ECT-induced cortisol elevation decreased 51.5% between first and last ECTs (8.5 ± 8.0 pg/dl. p = 0.015 repeated measures ANOVA, placement covaried). Each subject showed greater ECT-induced cortisol elevation with the first ECT (p = 0.00024, binomial). Likewise, ECT-induced cortisol elevation decreased (11.1 ± 8 .4 pg/dl, p = 0.0035 1-tailed, t = 3.76. d.f. = 7; p = 0.0039 binomial) in the subset of 8 unilateral ECT sub jects. This decrease exceeded (p = 0.042, Mann-Whitney for skew) that for the 4 bilateral ECT subjects (3.4 ± 4.4 pg/dl) by 225%. Correspondingly. last-ECT cortisol release was lower (p = 0.041, Mann-Whitney) for unilat eral treatment (4.7 ± 6.6 pg/dl) than bilateral (14.6 ± 5.1 pg/dl). Age of unilateral ECT patients (61 ± 7.9) was not substantially different (t = 1.28. d.f. = 10. p = 0.23) from bilateral ECT patients (66.7 ± 5.9).
mone (CRH) and ACTH [Gold and Chrousos, 1985: Smoak et al., 1991], Because ECT does not change CRHinduced cortisol release [Dored et al., 1990], the clear fall of ECT-induced cortisol release suggests that the sensitiv ity of CRH release to ECT falls with treatment. Because plasma CRH concentrations do not reliably correspond to hypothalmic CRH release [Sasaki et al., 1987], direct plasma CRH measurements are not conclusive [Widerlov et al., 1989]. Ultimately, measurements of ACTH and cortisol release sensitivities coordinated with measure
ments of ECT-induced ACTH and cortisol should pro vide direct evaluation of the sensitivity of CRH release to ECT and further elucidation ofHPA axis dysrégulation in depression.
Acknowledgment Neeta C. Saheba. MD assisted with the protocol, which was con ducted at the North Chicago Veterans Affairs Medical Center.
References Fink M. Johnson L: Monitoring the duration of electroconvulsive therapy seizure. Arch Gen Psychiatry 1982:39:1189-1191. Fink M, Gujavarty K. Greenberg L: Serial déxa méthasone suppression tests and clinical out come in ECT. Convulsive Thcr 1987:3:111-
120. Gold PW. Chrousos G P: Clinical studies with cor ticotropin releasing factor: Implications for the diagnosis and pathophysiology of depression. Cushing’s disease, and adrenal insufficiency. Psychoneurocndocrinology 1985; 10:401- 4 19. Greden JF, Gardner R, King D, Grunhaus L, Carroll BJ. Kronfol Z: Dexamethasone suppres sion tests in antidepressant treatment of melan cholia. Arch Gen Psychiatre 1983:40:493— 500. Mitchell P. Smythe G. Torda T: Effect of the anes thetic agent propofol on hormonal responses to ECT. Biol Psychiatry 1990;28:315-324. Sackeim HA, Decina P, Portnoy S, Kanzler M. Kerr B, Malitz S: Effects of electrode place ment on the efficacy of titrated low-dosage ECT. Am J Psychiatry 1987;144:1449-1455. Sasaki A, Sato S, Murakami O, Go M, Inoue M. Shimizu Y, Hanew' K, Andoh N, Sato I, Sasano N, Yoshinaga K: Immunoreactive corticotro pin-releasing hormone present in human plasma may be derived from both hypotha lamic and extrahypothalamic sources. J Clin Endocrinol Metab 1987;65:176-182. Smoak B, Deuster P, Rabin D, Chrousos G: Corti cotropin-releasing hormone is not the sole fac tor mediating exercise-induced adrenocorlicotropin release in humans. J Clin Endocrinol Metab 1991;73:302-306. Swartz CM: The time course of post-ECT prolactin levels. Convulsive Ther 1985a; 1:81 -89.
Swartz CM: Characterization of the total amount of prolactin released by electroconvulsive ther apy. Convulsive Ther 1985b: 1:252-257. Swartz CM, Abrams R: Prolactin levels after bilat eral and unilateral ECT. Br J Psychiatry 1984: 144:643-645. Swartz CM, Chen JJ: Electroconvulsive therapvinduccd cortisol release: Changes with depres sive state. Convulsive Ther 1985; 1:15-21. Swartz CM, Saheba NC: Dose effect on dexametha sone suppression testing with ECT. Ann Clin Psychiatry 1990;2:155-160. Sw artz CM, [.arson G: Generalization of the effects of unilateral and bilateral ECT. Am J Psychia try 1986:143:1040-1041. Turner TH, Ur E, Grossman A: Naloxone has no effect on hormonal responses to ECT in man. Psychiatry Res 1987;22:207-212. Weizman A, Gil-Ad 1. Grupper D. Tyano S, Laron Z: The effect of acute and repeated electrocon vulsive treatments on plasma beta-endorphin, growth hormone, prolactin and cortisol secre tion in depressed patients. Psychopharmacol ogy 1987;93:122-126. Whalley U . Eagles JM. Bowler GM, Bennie JG. Dick HR. McGuire RJ. Fink G: Selective ef fects of ECT on hypothalamic-pituitary activi ty. Psychol Med 1987;17:319-328. Widerlov E. Ekman R. Jensen L. Borglund L, Ny man K: Arginine vasopressin, but not cortico tropin releasing factor, is a potent stimulator of adrenocorticolropin hormone following elec troconvulsive therapy. J Neural Transm 1989; 75:131-109.
133
Downloaded by: Nagoya University 133.6.82.173 - 1/15/2019 1:19:15 AM
Abrams R. Swartz CM. Vcdak C: Antidepressant effects of high-dose right unilateral ECT. Arch Gen Psychiatry 1991:48:746-748. Abrams R. Swartz CN: Electroconvulsive therapy and prolactin release: Effect of stimulus param eters. Convulsive Thcr 1985; 1:115—119. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, cd 3 rev. Washington, American Psychiatric Asso ciation. 1987. American Psychiatric Association: The Practice of Electroconvulsive Therapy: Recommenda tions for Treatment. Training, and Privileging. Washington, American Psychiatric Associa tion. 1990. Arato M. Szontagh L, Tclegdy G: Effect of electro convulsive treatment on brain monoamine content, pituitary-adrenal function and plasma prolactin. Acts Physiol Acad Sci Hung 1980: 56:163-171. Carroll BJ. Feinberg M. Smouse P. Rawson SG. Grcden JF: The Carroll rating scale for depres sion. I. Development, reliability and validity. Br J Psychiatry 1981; 138:194-200. Deakin JFW. Ferrier IN, Crow TJ. Johnstone EC, Lawler P: Effects of ECT on pituitary hormone release: Relationship to seizure, clinical vari ables and outcome. Br J Psychiatry 1983:143: 618-624. d'Elia G, Raotma H: Is unilateral ECT less effective than bilateral ECT? Br J Psychiatrv 1975:126: 83-89. Dored G, Stcfansson S, d ’Elia G, Kigedal B. Karlberg E, Ekman R: Corticotropin, cortisol, and beta-endorphin responses to the human corti cotropin-releasing hormone during melan cholia and after unilateral electroconvulsive therapy. Acta Psychiatr Scand 1990:82:204209.
