Why is electroconvulsive therapy for depression more effective in older age? A causal mediation analysis.
Loading...
Authors
Jelovac, Ana
Landau, Sabine
Beeley, Petros
McCaffrey, Cathal
Finnegan, Martha
Gusciute, Gabriele
Whooley, Emma
McDonogh, Sarah
Thompson, Sarah
Igoe, Anna
Issue Date
2025-04-10
Type
Journal Article
Language
en
Keywords
ECT , age of onset , bipolar disorders , depression , depressive disorders , electroconvulsive therapy , late-life depression , mediation analysis , neurostimulation , psychomotor disturbance , psychotic depression , response
Alternative Title
Abstract
Older people with depression exhibit better response to electroconvulsive therapy (ECT). We aimed to measure the total effect of age on ECT response and investigate whether this effect is mediated by psychotic features, psychomotor retardation, psychomotor agitation, age of onset, and episode duration.
We pooled data from four prospective Irish studies where ECT was administered for a major depressive episode (unipolar or bipolar) with baseline score ≥21 on the 24-item Hamilton Depression Rating Scale (HAM-D). The primary outcome was change in HAM-D between baseline and end of treatment. The estimands were total effect of age, estimated using linear regression, and the indirect effects for each putative mediator, estimated using causal mediation analyses.
A total of 256 patients (mean age 57.8 [SD = 14.6], 60.2% female) were included. For every additional 10 years of age, HAM-D was estimated to decrease by a further 1.74 points over the ECT period ( < 0.001). Age acted on all putative mediators. Mechanistic theories, whereby a mediator drives treatment response, were confirmed for all putative mediators except age of onset. Consequently, mediation of the effect of age on change in HAM-D could be demonstrated for psychotic features, psychomotor retardation, psychomotor agitation, and episode duration but not for age of onset.
A total of 43.1% of the effect of older age on increased ECT response was explained by the mediators. Treatment planning could be improved by preferentially offering ECT to older adults, especially if presenting with psychotic features, greater severity of psychomotor disturbance, and earlier in the episode.
We pooled data from four prospective Irish studies where ECT was administered for a major depressive episode (unipolar or bipolar) with baseline score ≥21 on the 24-item Hamilton Depression Rating Scale (HAM-D). The primary outcome was change in HAM-D between baseline and end of treatment. The estimands were total effect of age, estimated using linear regression, and the indirect effects for each putative mediator, estimated using causal mediation analyses.
A total of 256 patients (mean age 57.8 [SD = 14.6], 60.2% female) were included. For every additional 10 years of age, HAM-D was estimated to decrease by a further 1.74 points over the ECT period ( < 0.001). Age acted on all putative mediators. Mechanistic theories, whereby a mediator drives treatment response, were confirmed for all putative mediators except age of onset. Consequently, mediation of the effect of age on change in HAM-D could be demonstrated for psychotic features, psychomotor retardation, psychomotor agitation, and episode duration but not for age of onset.
A total of 43.1% of the effect of older age on increased ECT response was explained by the mediators. Treatment planning could be improved by preferentially offering ECT to older adults, especially if presenting with psychotic features, greater severity of psychomotor disturbance, and earlier in the episode.
Description
Citation
Jelovac, A., Landau, S., Beeley, P., McCaffrey, C., Finnegan, M., Gusciute, G., Whooley, E., McDonogh, S., Thompson, S., Igoe, A., McDonagh, K., & McLoughlin, D. M. (2025). Why is electroconvulsive therapy for depression more effective in older age? A causal mediation analysis. Psychological medicine, 55, e110. https://doi.org/10.1017/S0033291725000807
Publisher
License
Journal
Psychological medicine
Volume
55
Issue
PubMed ID
DOI
10.1017/S0033291725000807
10.1111/acps.12994
10.1016/j.jad.2010.02.131
10.1016/s0165-0327(02)00005-8
10.1111/acps.13770
10.1016/j.jad.2022.12.144
10.1017/s0033291797005436
10.1016/s0165-0327(00)00317-7
10.1016/j.eurpsy.2017.06.015
10.4088/JCP.20m13287
10.1192/bjp.2017.28
10.1016/j.jagp.2016.09.005
10.1111/acps.13631
10.1097/YCT.0000000000000560
10.1111/j.2044-8260.1967.tb00530.x
10.4088/JCP.14r09528
10.1097/JCP.0b013e3181ee0f5f
10.1016/j.jpsychires.2018.11.014
10.1192/bjp.169.1.68
10.1037/a0020761
10.1093/aje/kwv059
10.1017/S0033291716002737
10.1038/s41380-024-02803-2
10.1111/acps.13378
10.1111/j.1399-5618.2004.00137.x
10.1136/bmj.l1079
10.4088/JCP.21m14293
10.1097/00124509-200112000-00003
10.1002/gps.6133
10.1192/j.eurpsy.2024.1
10.1176/appi.ajp.2015.15030372
10.1192/bjp.2024.126
10.1111/acps.13425
10.1001/jama.2024.5756
10.1056/NEJMcp1402180
10.1176/ajp.156.12.1865
10.1016/j.jad.2016.06.013
10.4088/JCP.14r09145
10.1097/YCT.0000000000001043
10.1111/acps.12994
10.1016/j.jad.2010.02.131
10.1016/s0165-0327(02)00005-8
10.1111/acps.13770
10.1016/j.jad.2022.12.144
10.1017/s0033291797005436
10.1016/s0165-0327(00)00317-7
10.1016/j.eurpsy.2017.06.015
10.4088/JCP.20m13287
10.1192/bjp.2017.28
10.1016/j.jagp.2016.09.005
10.1111/acps.13631
10.1097/YCT.0000000000000560
10.1111/j.2044-8260.1967.tb00530.x
10.4088/JCP.14r09528
10.1097/JCP.0b013e3181ee0f5f
10.1016/j.jpsychires.2018.11.014
10.1192/bjp.169.1.68
10.1037/a0020761
10.1093/aje/kwv059
10.1017/S0033291716002737
10.1038/s41380-024-02803-2
10.1111/acps.13378
10.1111/j.1399-5618.2004.00137.x
10.1136/bmj.l1079
10.4088/JCP.21m14293
10.1097/00124509-200112000-00003
10.1002/gps.6133
10.1192/j.eurpsy.2024.1
10.1176/appi.ajp.2015.15030372
10.1192/bjp.2024.126
10.1111/acps.13425
10.1001/jama.2024.5756
10.1056/NEJMcp1402180
10.1176/ajp.156.12.1865
10.1016/j.jad.2016.06.013
10.4088/JCP.14r09145
10.1097/YCT.0000000000001043
ISSN
1469-8978