# Is It Just the Weight Loss?

By Nora Bissett, Pricing Editor (https://glpwatchdog.org/authors/nora-bissett) | Published: 2026-08-20 | Updated: 2026-10-30

Canonical page: https://glpwatchdog.org/research/is-it-just-the-weight-loss

> Every benefit reported for these drugs runs into the same question. A post hoc analysis of the SURMOUNT-OSA trials tried to separate the drug from the weight, and found the sleep apnea improvement doing work of its own.

## Key points

- Inflammation, insulin resistance and triglycerides were mediated INDEPENDENTLY by sleep apnea measures.
- An argument against 'it is all just the weight loss'.
- Mediation tests whether a hypothesis fits the data. It does not establish a causal pathway.
- The abstract reports significance but NOT the proportions mediated, so we do not quote magnitudes.
- Diastolic blood pressure was mediated by neither weight nor apnea. Unexplained.

Every benefit reported for these drugs runs into the same question: is this the medicine, or is it just the weight coming off? A post hoc mediation analysis of [the SURMOUNT-OSA trials](https://glpwatchdog.org/research/lists/top-glp1-sleep-apnea-studies) tried to separate the two, and concluded that treating **both** the sleep-disordered breathing and the obesity is likely needed to get the full cardiometabolic benefit.[1] That conclusion comes from modeling, and the modeling has limits.

## Why the question is hard

When someone with obstructive sleep apnea takes tirzepatide, several things change at once. They lose weight. Their apnea-hypopnea index falls. Their overnight oxygen levels improve. And their inflammatory markers, insulin resistance and blood pressure all move in the right direction.

Establishing which change drove which is not something a trial answers by design, because the trial randomized only the drug, not the weight loss or the apnea improvement. Mediation analysis is the statistical attempt to decompose it after the fact: given how much each intermediate changed, how much of the final outcome moves with it?

## What the analysis found

Across both 52-week studies in [the SURMOUNT-OSA program](https://glpwatchdog.org/research/trials/surmount-osa), tirzepatide improved cardiometabolic risk factors more than placebo. The mediation results then split into three distinct patterns.

*Which pathway the analysis found mediating each outcome. The published abstract reports direction and significance; it does not report the proportions mediated.[1]*

| Outcome | Mediated by sleep apnea metrics? | Mediated by weight? |
| --- | --- | --- |
| hs-CRP (inflammation) | Yes, independently | Not reported as an independent effect |
| HOMA-IR (insulin resistance) | Yes, independently | Not reported as an independent effect |
| Triglycerides | Yes, independently | Not reported as an independent effect |
| Systolic blood pressure | In combination with weight | Yes, weight alone as well |
| Diastolic blood pressure | **No** | **No** |

The first three rows are the interesting ones. Inflammation, insulin resistance and triglycerides showed an *independent* mediation effect from the sleep apnea measures. The model attributes part of their improvement to breathing better at night, over and above what the weight loss accounts for.

> In the model, part of the improvement in inflammation, insulin resistance and triglycerides ran through better breathing at night, beyond what the weight loss accounted for.

That is an argument against “it is all just the weight loss”, but a modeling argument rather than a demonstration. A far larger prespecified analysis has since reached the same conclusion for the heart, in 17,604 people against a hard endpoint: [losing weight predicted more heart attacks](https://glpwatchdog.org/research/losing-weight-predicted-more-heart-attacks). That result also undercuts a model that [projected heart benefit from risk-factor changes alone](https://glpwatchdog.org/research/a-benefit-nobody-observed).

## What mediation analysis can and cannot show

> **Mediation is not proof of a causal pathway.** It rests on assumptions that trial data cannot verify, principally that nothing unmeasured is causing both the intermediate and the outcome. If some third consequence of the drug improves both nighttime breathing and inflammation independently, mediation analysis will credit the breathing. The technique tests whether a hypothesis is *consistent* with the data. It does not establish it.

Two further cautions belong with that. Parts of this work were post hoc, including the insulin-resistance analysis and the mediation analysis itself, as the paper states. And the published abstract reports which effects were significant without reporting *how much* of each outcome was mediated, so we cannot tell you whether sleep apnea accounted for a large share of the inflammation improvement or a small one. We are not going to estimate it.

The author disclosures on this paper are dense with industry relationships, Eli Lilly among them, as declared in the publication. For a manufacturer-sponsored analysis of a manufacturer-sponsored trial, that is expected, and readers should still know it.

## The blood pressure result nobody has explained

Systolic pressure was mediated by weight, and by weight combined with the sleep apnea measures. Diastolic pressure was mediated by **neither**: not weight, not the apnea metrics, and not the combination.

Diastolic pressure did change. It simply was not explained by either intermediate the analysis examined, which leaves an open question rather than a finding: something about the treatment moved it that this decomposition did not capture. We report that as unresolved and do not offer a mechanism for it.

## What it means practically

- **Sleep apnea is worth treating on its own terms**, not only as something weight loss will eventually fix. If the modeling is right, part of the metabolic benefit runs through the breathing.
- **Stopping CPAP because the weight came off may forfeit something.** This analysis gives a reason to think the two interventions are not interchangeable, though it does not test that directly.
- **“It is just the weight loss” is too simple a story**, and so is any confident claim about exactly how much of the benefit is not.

## Frequently asked questions

### Are the benefits of these drugs just from losing weight?

This analysis argues not entirely. In the SURMOUNT-OSA trials, improvements in inflammation, insulin resistance and triglycerides showed an independent mediation effect from the sleep apnea measures, beyond what weight loss accounted for. That is a modeling result rather than a demonstration.

### What is mediation analysis?

A statistical method for asking how much of a treatment's effect on an outcome travels through an intermediate step. It rests on assumptions trial data cannot verify, chiefly that no unmeasured factor is causing both the intermediate and the outcome.

### How much of the benefit came from the sleep apnea improvement?

The published abstract reports which mediation effects were statistically significant but not the proportions mediated, so that figure is not available to us. We are not going to estimate it.

### Why was diastolic blood pressure different?

It was mediated by neither weight nor the sleep apnea measures nor the combination. The analysis does not explain what moved it, and neither will we.

### Should someone with sleep apnea keep using CPAP on a GLP-1?

That is a decision for the clinician managing the sleep apnea. What this analysis suggests is that the breathing improvement may carry metabolic benefit of its own, which is a reason not to assume weight loss makes the other treatment redundant.

## References

1. Malhotra A, Grunstein R, Azarbarzin A, et al. Tirzepatide on obstructive sleep apnea-related cardiometabolic risk: secondary outcomes of the SURMOUNT-OSA randomized trial. Nature Medicine. 2026. PMID: 41540105. https://pubmed.ncbi.nlm.nih.gov/41540105/

---

Cite as: GLP Watchdog, Nora Bissett, "Is It Just the Weight Loss?", published 2026-08-20, updated 2026-10-30. https://glpwatchdog.org/research/is-it-just-the-weight-loss

Canonical HTML page: https://glpwatchdog.org/research/is-it-just-the-weight-loss

Medical disclaimer: Nothing here is medical advice. We report on this subject; we do not treat anyone. Whether to begin a medication, supplement, or treatment, whether to adjust it, and whether to stop are questions for a clinician licensed to answer them, who knows your history.
