Three decades of trials distilled into four steps, but much is missing in the gaps
STUDY: Lenze EJ et al, Neuropsychopharmacology 2026. doi: 10.1038/s41386-026-02338-w
STUDY TYPE: Review
FUNDING: Multiple NIH sources; Patient-Centered Outcomes Research Institute
Background
Polypharmacy, cognitive impairment, medical disorders, and altered drug metabolism complicate antidepressant therapy in older adults. This expert review synthesizes three decades of NIH-funded clinical trials into a practical, evidence-informed treatment algorithm.
The paper focuses on pharmacotherapy, but psychotherapy is an important option at each step. There’s an outdated adage that younger adults are better candidates for therapy, but new evidence finds older age predicts a better psychotherapy response.
The Algorithm
Step 1 is escitalopram or sertraline — preferred over other SSRIs because they carry fewer drug-drug interactions and QTc concerns. Sertraline is the safest in heart disease, but it does have drug interactions at doses of 150 mg and above.
The old aphorism “start low, go slow” needs an addendum: “but go all the way.” A low, inadequate dose carries risks (falls, bone loss) with none of the benefits.
Step 2. If the patient doesn’t remit, switch to an SNRI — venlafaxine or duloxetine preferred, desvenlafaxine is also reasonable. Approximately half reach remission with this step after SSRI failure (Dr. Aiken’s note: Although some data suggest a slightly higher efficacy with venlafaxine, it also has a higher risk of hypertension and withdrawal problems than other SNRIs. Duloxetine is preferable for patients with urine incontinence, but carries liver risks. Desvenlafaxine is the safest SNRI overall).
Amitriptyline is similar to SNRI but should be avoided in older adults given cardiotoxicity and anticholinergic burden (Dr. Aiken’s note: Among the tricyclics, nortriptyline and protriptyline are preferred for their lower fall risk).
Step 3 is augmentation with aripiprazole — the only strategy with a positive, adequately powered RCT specifically in older adults (the IRL-GRey trial, 44% vs. 29% remission over placebo) (dose 5-15 mg/day, usually 5 mg). Bupropion augmentation is an alternative but carries a meaningfully higher fall risk (surprisingly, it caused more side effects than aripiprazole in the IRL-GRey trial). After two monotherapy failures, augmentation outperforms switching (Dr. Aiken’s note: That is true, but TMS outperforms augmentation and is safer).
Step 4. After three adequate trials including one augmentation attempt, additional oral antidepressants are unlikely to work. At that point, IV ketamine, rTMS, or ECT should be on the table. ECT in particular remains underused — it’s safe even in the “oldest old” and is the treatment of choice for severe melancholic or psychotic depression (Dr. Aiken’s note: TMS could be considered earlier, or the new at-home device ProLivRx. For psychotic depression, ECT should be second line).
Where is Lithium?
Missing from this algorithm is lithium, which the authors skip over because it performed no better than switching meds in the OPTIMUM trial of late-life depression. However, older age predicted a better acute and long-term responses to lithium augmentation in observational studies. Lithium is a good option for highly recurrent depressions, and may have better preventative effects than antipsychotics.
Contrary to popular belief, lithium is well tolerated in the elderly, with similar renal risks and discontinuation rates in older and younger adults. The target serum levels are lower by about 30% in the elderly because the blood-brain barrier becomes more porous with age (Osterland SL et al, Acta Psychiatr Scand 2023;147(3):267-275; Flapper M et al, Int J Geriatr Psychiatry 2021;36(8):1231-1240, Buspavanich P et al, J Affect Disord. 2019;251:136-140; Lambrichts S et al, Acta Psychiatr Scand 2021;143(4):294-306; Christl J et al, 2023;56(5):e1).
Vascular Depression
Vascular depression is alarmingly common in older adults with depression, with rates approaching 100% after age 75:
Among patients with depression, the rate of vascular contributions rises with age (from Difficult to Treat Depression)
Vascular Depression was not addressed in the article, but it is a common cause of antidepressant failure. These patients are about 30% less likely to respond to antidepressants. Effective strategies include TMS, tDCS, and nimodipine augmentation. I discuss how to diagnose and treat it more in Difficult to Treat Depression.
Practice Implications
- This algorithm presents a starting place, but it leaves a lot of gaps. I’ve tried to fill some in.
- The authors stick strictly to large trials in older adults, which is both a strength and a weakness, missing nuances that help personalize the therapy, such as lithium for highly recurrent cases or specific approaches to vascular depression
- Good care involves making reasonable inferences from all the available data, not limiting our vision to the small piles of gold-standard evidence in large, randomized trials.
- If we had a well-designed trial showing TMS was safer and more effective than aripiprazole in older adults, it would quickly move to step 2. We don’t, but we have that for general adults, and we have studies showing older adults respond better to TMS.
- Older adults are twice as likely to get tardive dyskinesia (TD) on antipsychotics as younger adults, and new studies suggest a higher mortality rates from these meds (mainly from cardiovascular disease), but the algorithm overlooks these risks, which would not show up in the short-term trials it rests on.
—Chris Aiken, MD
Director, Psych Partners
Editor in Chief, Carlat Psychiatry Report
What’s Your Take? Share in Comments
- What do you turn to when antidepressants don’t work in older adults?
