How effective are SSRIs in curing depression

The short version: SSRIs produce a modest, real benefit over placebo for some people, but they do not "cure" depression for most, and the evidence is thinner than the marketing.

The pooled data from eight double-blind trials shows that about 35 in 100 people on an SSRI reach remission, compared to about 25 in 100 on placebo [2]. That is a real difference, but it means roughly 10 extra people out of 100 get well because of the drug. The rest either do not respond fully or respond to the placebo effect, which is substantial in depression trials. Another large analysis of over 5,000 patients on SSRIs found that the drug-placebo gap on the standard depression scale (HDRS-17) was about 4 to 6 points out of a possible 52, a modest separation that depended heavily on how you defined "response" [4]. The same study showed that a shorter, more sensitive scale (HDRS-6) separated drug from placebo more cleanly, which suggests the full 17-item scale dilutes the signal with items that do not change much [4].

The evidence for SSRIs in older adults is even weaker. A review of the elderly found that the odds of remission were not consistently better than placebo, and the number needed to treat for a response was about 10, meaning one extra responder for every ten people treated [1]. That same review documented real safety concerns: hyponatraemia, falls and fractures, upper GI bleeding, and dangerous drug interactions like with tamoxifen [1]. Sexual dysfunction is another major cost: up to 70 in 100 people on SSRIs report some form of sexual side effect, compared to about 7 in 100 on bupropion, a different class of drug [3]. These are not trivial. They are reasons people stop treatment.

The STAR*D study, the largest real-world effectiveness trial of an SSRI (citalopram), found that about 33 in 100 people remitted after the first treatment attempt [5]. But even among those who remitted, about 90 in 100 still had at least one residual symptom, most commonly sleep disturbance and appetite changes [6]. That is not a cure. It is a partial improvement that leaves many people still struggling. And the more residual symptoms someone had, the higher their risk of relapse over the next year [6].

The mechanism question matters here too. The serotonin hypothesis of depression, which the entire SSRI class is built on, has never been convincingly proven. Outside this retrieval, it is well documented that lowering serotonin in healthy volunteers does not reliably produce depression, and that SSRIs raise serotonin levels within hours but take weeks to show any clinical effect, which points to downstream mechanisms the field does not understand well. The evidence here does not settle that debate, but it does show that the drugs work modestly for some people through pathways that remain unclear.

My call: SSRIs offer a modest, statistically significant benefit over placebo for acute depression, with about 10 extra people in 100 achieving remission, but they do not cure most people, the evidence in the elderly is weak and carries distinct safety risks, and the side effects are common and often severe enough to stop treatment. Confidence: moderate for the acute benefit in younger adults, low for long-term cure or for efficacy in the elderly.