# When Pooling Studies Hides the Answer

By Nora Bissett, Pricing Editor (https://glpwatchdog.org/authors/nora-bissett) | Published: 2026-08-20 | Updated: 2026-10-27

Canonical page: https://glpwatchdog.org/research/when-pooling-studies-hides-the-answer

> A meta-analysis of 323,439 patients found no significant cardiovascular advantage for tirzepatide. The more important number sits beside it: the studies being pooled disagree with each other almost completely.

## Key points

- MACE HR 0.85 (0.70–1.04), not significant. Mortality and CV mortality were also null.
- I² = 90.4%. A common convention calls 75% high.
- Seven of eight studies were observational, which is where the disagreement comes from.
- A benchmarked claims study found 0.86 against semaglutide on a different outcome, a similar null.
- Completes the picture: tirzepatide wins decisively on weight, and shows no proven heart advantage.
- The significant results are exploratory subgroups, and are the ones that will be quoted.

A meta-analysis of 323,439 patients asked whether tirzepatide prevents more cardiovascular events than the GLP-1 drugs it competes with. The answer was no significant difference (hazard ratio 0.85, with the interval crossing 1).[1] The more important number is buried beside it: **I² = 90.4%**, which means the studies being pooled disagree with each other almost completely.

## What I² measures, and why 90% is a problem

When several studies of the same question produce different answers, some of that spread is ordinary sampling noise and some reflects real differences between the studies in their populations, definitions and confounding. I² estimates what proportion is the second kind.

A rough convention treats 25% as low, 50% as moderate and 75% as high. **Ninety percent is off the end of that scale.**

> A pooled estimate from studies that disagree this much describes none of them.

> **This is the meta-analytic equivalent of averaging a group where half the values are large and half are small, then reporting the middle as though it described anyone.** The summary hazard ratio of 0.85 is arithmetically correct, but it does not estimate a single underlying truth, because the analysis has already shown there is not one.

The likely source is visible in the study list. Seven of the eight studies were *observational*, drawn from different health systems, with different patients, different comparison groups and different residual confounding. That those seven disagree is unsurprising. Pooling them does not resolve the disagreement; it conceals it.

## The result, such as it is

*Tirzepatide against GLP-1 receptor agonists across 8 studies and 323,439 patients.[1]*

| Outcome | Hazard ratio (95% CI) | Significant? |
| --- | --- | --- |
| Major adverse cardiovascular events | 0.85 (0.70–1.04) | **No** |
| All-cause mortality | 0.90 (0.79–1.02) | **No** |
| Cardiovascular mortality | 0.88 (0.76–1.00) | **No** — the interval touches 1 |

All three point the same direction, which counts for something. None reaches significance. The third stops exactly on the boundary at 1.00, and someone will describe it as significant.

Exploratory subgroup analyses in people with type 2 diabetes did reach significance: all-cause mortality 0.85 (0.76–0.94) and heart failure 0.75 (0.58–0.97). Those are the numbers that will be quoted, and they are subgroups of an exploratory analysis inside a meta-analysis whose primary comparison was null and heterogeneous.

## It agrees with the other evidence, which matters more than the pooling

Set the meta-analysis aside and look at what else exists. A claims-based study *benchmarked its own method against two randomized trials*, then compared tirzepatide with semaglutide on heart failure hospitalization or death in people with heart failure. It found **0.86 (0.70–1.06)**, a different outcome and a narrower comparison, but a null with almost the same numbers (covered in [how to trust an observational study](https://glpwatchdog.org/research/how-to-trust-an-observational-study)).

> **Two independent analyses, using different data and different methods, both finding no demonstrated cardiovascular advantage for tirzepatide over the GLP-1 drugs.** Agreement between methods is stronger evidence than either provides alone, and it is stronger here than the pooled figure from studies that disagree with each other.

That completes a picture this register has been assembling. In [the head-to-head](https://glpwatchdog.org/research/the-head-to-head-and-what-it-settles), a randomized trial, tirzepatide beats semaglutide decisively on weight: 20.2% against 13.7%. On cardiovascular outcomes, no advantage has been demonstrated, and semaglutide is the agent with the mortality evidence across [262 trials](https://glpwatchdog.org/research/what-262-trials-say-together).

None of that means tirzepatide is worse for the heart. It means the trial that would settle it has not reported, and the authors of this analysis say exactly that: further randomized trials are needed.

## How to read the next meta-analysis you meet

- **Find I² before the summary estimate.** Above about 75%, the pooled number is describing a disagreement rather than a finding.
- **Count how many included studies were randomized.** One of eight, here.
- **Check whether the significant results are subgroups.** They usually are, and they usually get quoted.
- **Prefer agreement between methods over precision within one.** Two different approaches landing on the same answer is worth more than a tight interval around an average of contradictions.

## Frequently asked questions

### Does tirzepatide prevent more heart attacks than other GLP-1 drugs?

Not demonstrably. A meta-analysis of 323,439 patients found a hazard ratio of 0.85 with a confidence interval crossing 1, along with non-significant results for all-cause and cardiovascular mortality.

### What does I² of 90.4% mean?

That about 90% of the variation between the pooled studies is real disagreement rather than chance. A common convention treats 75% as high, so this is far beyond it. A pooled estimate from studies disagreeing that much describes none of them.

### Why do the studies disagree so much?

Seven of the eight were observational, drawn from different health systems with different patients, comparison groups and residual confounding. Pooling them does not resolve that disagreement; it averages over it.

### The subgroup results were significant. Do they count?

They are exploratory subgroups within an analysis whose primary comparison was null and highly heterogeneous. They are the numbers most likely to be quoted and the least likely to hold.

### Is there better evidence on this question?

A claims study that first reproduced two randomized trials with its own method before answering found a hazard ratio of 0.86 for tirzepatide against semaglutide on heart failure hospitalization or death: a null, and close to the pooled figure even though the outcome was different. Two methods agreeing is worth more than one pooled average of contradictory studies.

## References

1. Silva JPMRJ, Nogueira BV, Sartori DL, et al. Cardiovascular Outcomes With Tirzepatide Versus GLP-1 Receptor Agonists in Overweight or Obesity: A Systematic Review and Meta-Analysis. Diabetes, Obesity and Metabolism. 2026. PMID: 42410309. https://pubmed.ncbi.nlm.nih.gov/42410309/

---

Cite as: GLP Watchdog, Nora Bissett, "When Pooling Studies Hides the Answer", published 2026-08-20, updated 2026-10-27. https://glpwatchdog.org/research/when-pooling-studies-hides-the-answer

Canonical HTML page: https://glpwatchdog.org/research/when-pooling-studies-hides-the-answer

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.
