Scientific deep-dive

The Biggest Number Is the Weakest

A prespecified SELECT analysis reports cardiovascular reductions of 26%, 21% and 34% across three definitions of liver fibrosis risk. The 34% is the only one that was not statistically significant.

By Nora Bissett · Pricing Editor
Editorially reviewed. Not medical advice · How we verify contentLast reviewed
7 min read·2 citations

A prespecified analysis of the SELECT trial asked whether semaglutide still cut cardiovascular events in people whose blood tests suggested liver fibrosis. It reports three answers, for three definitions of high risk: 26%, 21% and 34%.[1] The 34% is the biggest, it is the one that will be quoted, and it is the only one that was not statistically significant. Understanding why is more useful than any of the three figures.

What the analysis did

SELECT enrolled people with overweight or obesity (a BMI of 27 or more) and diagnosed cardiovascular disease, none of whom had diabetes, and gave half of them weekly semaglutide. The drug cut the trial’s combined endpoint (death from cardiac causes, heart attack short of death, or stroke short of death) by a fifth.[2] This analysis takes that population and splits out those at high risk of significant liver fibrosis, using the FIB-4 score.

FIB-4 is a screening calculation from age, platelet count and two liver enzymes. It is not a biopsy and not a scan. “High risk of fibrosis” throughout this analysis means a blood-test score above a threshold, which is a reasonable way to find people worth investigating and is not a diagnosis. Whether blood and imaging tests can stand in for a biopsy is what one semaglutide trial was built to ask.

The three answers

MACE reduction with semaglutide, by definition of high fibrosis risk.[1]
DefinitionReductionHazard ratio (95% CI)P
FIB-4 ≥ 1.326%0.74 (0.63–0.88)0.0004
Age-specific (≥1.3 under 65; ≥2.0 at 65+)21%0.79 (0.63–0.98)0.035
FIB-4 > 2.67, any age34%0.66 (0.39–1.10)0.11

Look at the third row’s interval. It runs from 0.39 to 1.10. Because 1.10 is above 1, the data are compatible with semaglutide reducing events by 61%, and also compatible with it increasing them by 10%. When an interval crosses 1, the result has not distinguished benefit from no effect.

Why the biggest number is the weakest one

This is structural, not bad luck. The strictest definition of high risk selects the fewest people. The fewest people produce the fewest events. The fewest events produce the widest confidence interval. A wide interval has to be centered somewhere, so it will often land on a striking point estimate that carries almost no information.

That is why ranking these three rows by headline size inverts their ranking by reliability. The 26% figure, with the narrowest interval and a P of 0.0004, is the most trustworthy and the least dramatic. The 34% is the reverse.

A confidence interval that crosses 1 has not shown an effect, however impressive the number in front of it.

This does not mean the 34% is wrong. The true effect in that group may well be large, and the point estimate is the best single guess available. What it means is that the analysis cannot establish it, and anyone quoting “34% reduction” without the interval is transmitting confidence the data do not contain.

What the analysis does establish

The two significant rows are the substance. They show that semaglutide’s cardiovascular benefit in SELECT held up in people whose blood work suggested liver fibrosis, at roughly the magnitude seen in the trial overall. Liver risk did not appear to blunt the cardiovascular effect.

That is a modest, useful finding. It is also a subgroup analysis of a trial designed to answer a different question, which is the standing caveat on all such work. Prespecifying the subgroups, as here, is considerably better than choosing the groups after seeing the data, but it does not turn a subgroup into a trial.

One reported statistic we are not going to interpret. The liver measure itself, the fatty liver index, is described as decreasing 28% more with semaglutide, reported as a hazard ratio of 0.72 with a 95% interval of 0.71 to 0.73. A hazard ratio describes time to an event rather than a change in a continuous index, and that interval is far narrower than anything else in the analysis. We report it as published and cannot tell you what form of estimate it is.

The habit worth taking away

  • Read the interval before the headline. If it crosses 1 for a ratio, or 0 for a difference, the finding is not established.
  • Be most suspicious of the largest number in a set of subgroups. It is usually the smallest group.
  • Check whether subgroups were prespecified. Chosen in advance is a different thing from chosen after looking.
  • Ask what the risk score measures. Here, “high fibrosis risk” is a blood-test threshold, not a diagnosis.

SELECT turns up across this register in other guises: what the trial says about hospital days, in days spent in hospital, and how many years of treatment the evidence base covers, in how long is the evidence.

Frequently Asked Questions

In this prespecified analysis of SELECT, the combined endpoint of cardiac death, heart attack and stroke fell 26% among people whose FIB-4 score reached 1.3 or higher (P = 0.0004) and 21% using an age-specific threshold (P = 0.035). Both are roughly the effect seen in the trial overall.
That figure comes from the strictest definition, a FIB-4 above 2.67, and it was not statistically significant: a hazard ratio of 0.66 with a confidence interval from 0.39 to 1.10, P = 0.11. The interval crosses 1, so the analysis did not distinguish benefit from no effect in that group.
Because the strictest criterion selects the fewest people, who generate the fewest events, which produces the widest confidence interval. A wide interval often lands on a striking point estimate that carries little information.
No. FIB-4 is a screening calculation from age, platelet count and two liver enzymes. A high score identifies someone worth investigating further; it is not a diagnosis, and no biopsies or scans were involved in defining these groups.
These subgroups were prespecified, which is substantially better than choosing them after seeing the results. It remains a subgroup analysis of a trial powered to answer a different question, so it supports rather than establishes.

References

  1. 1.Meyhöfer SM, Cariou B, Cercato C, et al. Semaglutide on liver fibrosis and heart outcomes in patients at high risk of liver fibrosis: a prespecified analysis of the SELECT randomized trial Nature Medicine. 2026. PMID: 41928037.
  2. 2.Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes N Engl J Med. 2023. PMID: 37952131.

The Line Is Drawn by Surgery

A meta-regression found study arms with greater average weight loss had higher rates of fatty liver disease resolving. The authors say plainly it cannot establish cause, and a handful of surgical arms are pulling the line.

6 min read

What Changed in the Teenagers' Bloodwork

A STEP TEENS analysis found insulin resistance down 35%, liver enzymes down 18%, and lipids improved in 12-to-17-year-olds. All of it is bloodwork, and the outcomes it stands in for are decades away.

7 min read

A Trial About the Measuring Instrument

SAMARA will be reported as showing semaglutide improves fatty liver disease. It was built to ask whether non-invasive tests can replace a biopsy, and the drug was the tool used to answer it.

7 min read

GLP-1s and Fatty Liver: What the Phase 3 Trial Showed

Semaglutide resolved steatohepatitis in 62.9% of patients, against 34.3% on placebo. The difference is 28.7 points. In the one trial in compensated cirrhosis, it produced neither resolution nor fibrosis improvement.

7 min read

Can Ozempic or Mounjaro Cause Liver Damage? Liver Enzymes Evidence Review

Liver injury on semaglutide or tirzepatide is rare. Trial ALT levels fall on average, no label carries a liver warning, and about 31 suspected cases have been published. Gallstones are the likelier cause of a raised liver test, and jaundice or dark urine needs same-day care.

11 min read

Does CLA Help With Weight Loss? What the Trials and FDA Records Show

CLA (conjugated linoleic acid) supplements beat placebo by about a third of a kilogram across 70 randomized trials, and not at all for fat in the best-run ones. What the meta-analyses, insulin-resistance trials, liver reports and FDA GRAS letters show.

10 min read

Where to get GLP-1 online, safely: sellers our editors have checked

These are telehealth sellers our editors have checked. For each one we hold a price, the form the drug comes in, and the states it reaches.

No insurance needed · vetted by our editors

Some of the links on this page earn us money. If you sign up with a provider after following one, that provider may pay GLP Watchdog a commission. Learn more

9.1

Collective

Flat any-dose pricing, if you can absorb a $199 annual membership on top

From $59/mo + membership

Get started →
Read full review →
8.5

Gala

Moving between compounded and brand without changing seller

9.3

Embody

A flat compounded tirzepatide or semaglutide price at every dose, from a seller that names its pharmacies

About this analysis

GLP Watchdog publishes independently. Each seller in our dataset is checked against that seller's own website, and the dataset is updated continuously. Read our full methodology →