What does STEP 2 actually establish about semaglutide?

The particulars: STEP 2 · semaglutide.

I have the document in front of me and I can read the numbers. What I cannot do is interpret them.

I am reasonably comfortable with statistics and completely uncomfortable with chromatography, or vice versa.

Which parts of this are informative and which are decoration?

Put another way, read the estimand before the effect size. Almost every apparent contradiction between two published figures from the same trial resolves once you notice that one is a trial-product estimand and the other is a treatment-policy estimand.

Creatinine is a muscle-derived metabolite, so a substantial loss of lean mass lowers serum creatinine and mathematically raises estimated GFR without anything happening to the kidney. If you have lost twenty kilograms, your creatinine-based eGFR is flattering you. Cystatin C is not muscle-dependent and is the measure to use when the two disagree.

Stated carefully, estimated average glucose from HbA1c: eAG in mg/dL = 28.7 × A1c − 46.7, or in mmol/L, 1.59 × A1c − 2.59. An A1c of 6.5 per cent is therefore about 140 mg/dL or 7.8 mmol/L. The relationship is a population regression, so an individual can sit well off the line.

SURMOUNT-4 randomised participants after an open-label lead-in to continued tirzepatide or placebo, and the withdrawal arm regained a substantial proportion of the lost weight over the following year[1].

The caveat is the population. Trial participants were screened, monitored and supported; the effect size in an unmonitored setting is not the trial effect size, and it is not obvious in which direction the difference runs.

Read the confidence interval, read the estimand, and compute the absolute effect yourself. It takes two minutes and it changes how the result feels.

edited 20 Jun 2025 by swab_and_wait — reworded for clarity after a comment

Start with the population. The inclusion criteria of the trial determine what its result can be extrapolated to, and the extrapolation people want is usually to a population the trial excluded.

Absolute risk reduction, worked: if the control-arm event rate is 8.0 per cent over the follow-up period and the hazard ratio is 0.80, the treated rate is approximately 6.4 per cent, the absolute risk reduction is 1.6 percentage points, and the number needed to treat is 1 ÷ 0.016 ≈ 63 over that period. A 20 per cent relative reduction and a number needed to treat of 63 are the same finding stated two ways, and only one of them sounds impressive.

A fasting lipid panel drawn during rapid weight loss reads oddly for a mechanical reason: mobilised adipose tissue delivers free fatty acids to the liver, and hepatic triglyceride export rises. Triglycerides can transiently increase while the person is doing exactly the right thing. Draw the panel when weight has been stable for a few weeks if you want an interpretable number.

STEP 1 reported a mean weight change of approximately −14.9 per cent with semaglutide 2.4 mg versus −2.4 per cent with placebo at 68 weeks[1]; the difference between the figures quoted from this trial in different places is an estimand difference.

One qualification: a trial that demonstrates an endpoint at a given dose has demonstrated it at that dose. Extrapolating the endpoint down the dose ladder is an assumption, not a finding.

Convert everything to an absolute effect before you compare two interventions. Relative effects are not comparable across different baseline risks.

It helps to be literal here: this is a question about what the trial was designed to answer, and the honest response is that it was not designed to answer this.

Liver enzymes are a poor surrogate for hepatic histology in both directions: substantial steatohepatitis with normal transaminases is common, and modest enzyme elevation with minimal fibrosis is common. If the question is fibrosis, the answer comes from a non-invasive score such as FIB-4 or a stiffness measurement, not from ALT.

The part that matters: the early fall in estimated glomerular filtration rate on treatment is haemodynamic rather than structural. Reduced intraglomerular pressure lowers the filtration rate acutely and preserves the glomerulus chronically — the same pattern seen with renin-angiotensin blockade and with SGLT2 inhibition. A dip of a few millilitres per minute in the first weeks, followed by a shallower long-term slope, is the desired trajectory, not a warning sign.

SURMOUNT-1 reported mean weight reductions of approximately 15, 19 and 21 per cent at tirzepatide 5, 10 and 15 mg respectively at 72 weeks[1].

The papers are readable. Read the paper rather than the summary of the paper, especially where the summary is enthusiastic.

In practice, the confidence interval is the informative part. A point estimate with an interval spanning no effect is a different object from the same point estimate with a tight interval, and the abstract presents them identically.

ApoB and LDL-C disagree because they measure different things: LDL-C is the cholesterol mass carried in the LDL fraction, ApoB is a count of atherogenic particles. Small dense particles carry less cholesterol each, so a person with many small particles has a concordantly higher ApoB than their LDL-C suggests. When they disagree, ApoB is the better risk marker.

Worth being explicit that this is interpretation of published data and not medical advice. Laboratory results belong in a conversation with whoever ordered them.

None of this replaces a clinician who can see the whole picture, and the whole picture is usually where the answer is.

It helps to be literal here: a single laboratory value is a point on a noisy curve. What you want is a trend across at least three draws under comparable conditions, and "comparable" is doing a lot of work in that sentence.

A network meta-analysis can rank agents that were never compared directly, but only under a transitivity assumption — that the trials being linked are similar enough in population, duration and endpoint definition for the indirect comparison to hold. In this field that assumption is often visibly violated, which is why indirect rankings should be read as hypotheses.

I would resist reading a subgroup finding as a result. Subgroups in these trials were not powered, and a striking subgroup in a large trial is the expected consequence of multiplicity.

If the trend across three draws is flat, the difference between draws one and two was noise. Most of what people react to is noise.