Sleep and Mood on a GLP-1: What Improved, and What Did Not Predict It

Seventy-eight people with obesity starting semaglutide or tirzepatide were assessed on two standard questionnaires at baseline and again at three months. The share scoring as poor sleepers fell from 55.1% to 16.7%, and the share with clinically significant depressive symptoms fell from 55.1% to 11.5%.^[1] There was no control group, so the honest reading is narrower than those numbers look — but one finding inside them is worth your attention: the amount of weight lost was not what tracked with who improved.

What was measured

A prospective observational study enrolled patients with obesity starting a GLP‑1-based therapy: 52 on tirzepatide and 26 on semaglutide. Depressive symptom burden was measured with the PHQ‑9 and sleep quality with the Pittsburgh Sleep Quality Index, at baseline and at month three.^[1]

The part that is not about weight

The intuitive story is that people lost weight, so they slept better and felt better. The multivariable models point elsewhere. What was associated with a depressive-symptom response was a higher baseline PHQ‑9 and younger age. What was associated with a sleep response was a higher baseline PSQI and a greater improvement in depressive symptoms.^[1]

Percentage weight loss did not emerge as the driver of either. Mood improvement tracked with sleep improvement, which is a well-known relationship in both directions and does not need the scale to explain it.

If the benefit ran through the weight, the people who lost most would have improved most. In these models, that is not what separated them.

Two cautions on reading that as good news. Higher baseline scores predicting more improvement is exactly what regression to the mean produces, so that particular association is the least surprising thing in the paper. And an uncontrolled study cannot separate a drug effect from the effect of starting a treatment you have hopes for.

What it does not support

• No GLP-1 is approved for depression or for insomnia, and nothing here suggests taking one for either.
• Three months is short. Neither sleep nor mood was followed beyond that.
• 78 people without a control group cannot establish that the drug caused the change.
• ⚠ If you are taking medication for depression or sleep, a GLP-1 is not a substitute and changes belong with your prescriber.
• ⚠ Anyone whose mood worsens after starting one of these drugs should raise it promptly rather than wait — that direction is not what this study looked for.

For a randomized look at mood and cognition see the semaglutide depression and cognition trial; for daytime sleepiness specifically, GLP-1s and daytime sleepiness; and for the airway condition with an approved indication, Zepbound for sleep apnea.