Depression in later life can be particularly stubborn, and ECT is often effective for it. A reasonable question follows: can anyone tell in advance whether it will work for a particular person?
This review searched five research databases and screened 2,149 articles, ultimately including 19 that examined predictors of response in older adults.
The most striking finding concerns how much did not predict anything. Brain scan measures that sound as though they ought to matter, including hippocampal volume, white matter changes, amyloid load and overall brain shrinkage, showed no significant relationship with whether someone improved. Salivary cortisol showed none either.
A few smaller signals did emerge, including certain blood markers and how long people took to reorient after a treatment. Nothing has yet become a reliable test.
This matters because it sets honest expectations. If you are weighing ECT for an older parent, no one can currently scan their brain and tell you whether it will work. The decision has to rest on symptoms, severity and what has already been tried.
That uncertainty is uncomfortable, and knowing it is better than relying on a prediction that does not exist. Your treating team can talk through what is known for your situation.
What the study did
This paper is a systematic review, meaning the authors didn't run a new experiment but instead searched existing research systematically and pulled together what's already known. Pozuelo Moyano et al. searched five major research databases for studies published between 1995 and June 2023, screened 2,149 articles, and included the 19 that specifically looked at what predicts whether older adults respond to electroconvulsive therapy (ECT) for depression.
In this literature, 'response' usually means a meaningful drop in depression severity on a standard rating scale (the review references the Montgomery-Åsberg Depression Rating Scale, or MADRS), while 'remission' means depression symptoms have essentially resolved. The review asked: across these 19 studies, which patient characteristics — biological, cognitive, or clinical — reliably tracked with better outcomes?
What it found
Several biologically plausible candidates showed no significant relationship with response or remission. These included salivary cortisol (a stress hormone often studied in depression), hippocampal volume and white matter hyperintensities seen on baseline MRI scans (small areas of altered brain tissue, often linked to vascular changes), total amyloid load (a protein build-up associated with Alzheimer's disease), and global cortical atrophy (overall brain shrinkage). The review also found no significant difference in remission rates between unilateral and bilateral electrode placement during ECT — a technical choice about where the electrical stimulus is delivered.
A smaller set of factors did show associations with better outcomes. Moderately elevated levels of CRP (C-reactive protein, a marker of inflammation) and S100B (a protein associated with brain cell activity) were linked to higher remission or response. Lower baseline 'retardation' scores (a symptom of depression involving slowed movement and thinking) and, interestingly, poorer performance on a word-reading task at baseline, were also associated with better outcomes. Longer post-ictal reorientation time — how long a person takes to become mentally oriented after the brief seizure that ECT induces — was linked with higher remission or response too. Conversely, medial temporal atrophy (shrinkage in a specific brain region) was associated with a smaller decrease in MADRS scores, meaning less improvement. Finally, the review noted that elderly patients overall tend to have higher ECT response rates, and suggested this may partly be explained by retardation and psychotic features being more common in this group.
Why it matters, and its limits
Taken together, this suggests some clinical and biological markers may eventually help tailor ECT decisions for older adults — but the evidence is still preliminary. The authors are explicit that research in this specific population is scarce, and only a modest number of studies has focused on it directly. None of these findings should be read as a working predictive test; they are patterns worth investigating further, not tools ready for the clinic.
If you or someone you care about is considering ECT, these open questions are exactly the kind of thing worth discussing with a treating psychiatrist, who can weigh your specific situation rather than a general prediction rule.