did a meta-analysis find trazodone improved sleep quality versus placebo by an smd of 0.827?
“The Trasadone group demonstrated significant improvements in sleep quality compared to placebo with a standardized mean difference of 0.827.”
a guest, Dr. Paul Sarkowski (clinician presenting the meta-analysis) said on Psychopharmacology and Psychiatry Updates, 2026-09-29, at 04:15. hear it in context → the speaker cited: a new meta-analysis.
the episode: Trazodone for Insomnia and Comorbid Depression (at 04:15). we link the show's own page; we do not re-host audio.
first, the part that holds
the number is right. a 2025 meta-analysis in psychopharmacology reports that trazodone improved pittsburgh sleep quality index scores with a standardized mean difference of -0.827 (95% ci -1.331 to -0.323, p = 0.001), the same magnitude the speaker quoted, and the minus sign just means lower, better psqi scores.
what we found when we went and looked
we found the paper: a systematic review and meta-analysis of 10 randomized trials (1,029 participants) in people with depression and sleep problems, published online in psychopharmacology in october 2025 (pmid 41081841). its abstract gives psqi smd = -0.827, 95% ci -1.331 to -0.323, p = 0.001, which matches the quoted 0.827 in size and direction. the one clause we could not check is "compared to placebo": the word placebo does not appear anywhere in the pubmed abstract, which says only that included trials were rcts "assessing trazodone's effects on sleep and depression in depressive patients", and the full text, where each trial's comparator would be listed, is paywalled and we could not read it. a larger 2024 meta-analysis limited to placebo-controlled trials (44 rcts, 3,935 participants) also found better sleep quality with trazodone, with a smaller pooled effect (smd -0.58).
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| a meta-analysis found trazodone significantly improved psqi sleep quality with a standardized mean difference of 0.827 | confirmed | the abstract reports psqi smd = -0.827 (95% ci -1.331 to -0.323, p = 0.001), pooled from trials in depressive patients. the negative sign means lower, better psqi scores, so the speaker's unsigned 0.827 is the same result. the abstract does not say how many of the 10 trials fed this pooled figure. the confidence interval is wide, running from a small-to-moderate effect to a very large one. [1][2] |
| the comparison was against placebo | unverifiable | 'placebo' does not appear anywhere in the pubmed abstract. the methods section says only that rcts 'assessing trazodone's effects on sleep and depression in depressive patients' were included. the full text, which would list each trial's control arm, is paywalled, and we could not read it. the psqi pool may be all placebo-controlled, or it may mix in other control arms; the abstract doesn't say which. [1][2] |
| the finding applies to people with depression and sleep problems, not insomnia in general | confirmed | the review's stated objective was trazodone 'for managing sleep disturbances in patients with depression'. so the 0.827 figure is not an estimate for primary insomnia. [1][2] |
| across the wider placebo-controlled record, trazodone improves self-rated sleep quality | confirmed | a 2024 cns drugs meta-analysis restricted to placebo comparisons (44 rcts, 3,935 participants) found improved sleep quality, smd -0.58 (95% ci -0.87 to -0.28). it found no significant change in self-reported total sleep time (wmd 0.73 min). the direction is the same as the 0.827 figure, and the pooled size is smaller. [3] |
| the same meta-analysis found more adverse effects with trazodone | confirmed | the abstract reports higher odds of blurred vision (or 17.50, 95% ci 2.28-134.02), somnolence (or 7.34, 2.91-18.50) and sedation (or 6.53, 3.59-11.87) with trazodone. the very wide interval on blurred vision means few events. [2] |
both directions
for the speaker: the quoted figure matches the published abstract exactly in magnitude, and the speaker described it correctly as statistically significant (p = 0.001). the direction holds up in a separate, larger 2024 meta-analysis of 44 placebo-controlled trials, which also found better self-rated sleep quality with trazodone (smd -0.58). the 2025 paper also found a smaller but significant drop in hamilton depression scores (smd -0.365).
against: the abstract never says "placebo", so the "versus placebo" framing can't be confirmed from what is publicly readable. the 0.827 estimate is imprecise: the 95% interval runs from -0.323 to -1.331. the larger placebo-only pool put the sleep quality effect lower, at about 0.58, so 0.827 is likely toward the high end of current estimates and not a settled number. the 2024 review found no gain in self-reported total sleep time. the 2025 review reported markedly higher odds of somnolence, sedation and blurred vision, and its conclusion says the benefits "must be weighed against a higher risk of adverse effects". clinical global impression scores did not differ significantly (smd -0.209, p = 0.118).
what this cannot say
we read the abstract and pubmed record only. the full text is paywalled, so we could not see the trial-by-trial comparators, how many trials fed the psqi pool, heterogeneity (i²), or the grade certainty ratings the title advertises. psqi is self-reported, and standardized mean differences don't translate directly into points on the psqi scale. the two meta-analyses differ in population (depressive patients vs. mixed sleep populations) and in how they define sleep quality, so the 0.827 vs 0.58 comparison is a rough one. we checked the published paper, not the episode audio, so the quote is as supplied to us.
sources, each one fetched and read
- the efficacy and safety of trazodone for sleep problems in depressive patients: a GRADE-assessed systematic review and meta-analysis of clinical trials. Psychopharmacology (Berl) 2026;243:457-473 (online october 2025). PMID 41081841, doi:10.1007/s00213-025-06910-y (pubmed record) What we read there: methods: 'randomized controlled trials (RCTs) assessing trazodone's effects on sleep and depression in depressive patients were included.' the word placebo does not appear in the abstract; no pmc full-text id is listed. Fetched 2026-10-12 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/41081841/
- same article, publisher abstract page, Psychopharmacology, Springer, doi:10.1007/s00213-025-06910-y What we read there: 'Ten RCTs (1,029 participants) were included. Trazodone significantly improved sleep quality (Pittsburgh Sleep Quality Index: SMD = -0.827, 95% CI: -1.331 to -0.323, p = 0.001) and reduced depression severity (Hamilton Depression Rating Scale: SMD = -0.365...).' adverse effects: blurred vision OR 17.50, somnolence OR 7.34, sedation OR 6.53. Fetched 2026-10-12 (HTTP 200). https://link.springer.com/article/10.1007/s00213-025-06910-y?error=cookies_not_supported&code=1cb0a08f-f5be-4554-8f99-6895a8f3a3f9
- effects of trazodone on sleep: a systematic review and meta-analysis. CNS Drugs 2024. PMID 39123094, doi:10.1007/s40263-024-01110-2 (pubmed record, fetched with related records) What we read there: 44 rcts (3,935 participants) versus placebo; trazodone did not significantly change subjective total sleep time but improved sleep quality (SMD = -0.58, 95% CI -0.87; -0.28, p < 0.01); more dropouts owing to adverse effects (RR = 2.30). Fetched 2026-10-12 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/42787467/
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam