Do incretins move hard outcomes in CKM, not just weight? A close read of SELECT, FLOW and SUMMIT

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, and not medical advice.

The 30-second version

• What. GLP-1 / incretin agents lowered hard cardiovascular (SELECT) and kidney (FLOW) outcomes. In SELECT’s pre-specified analysis, roughly two-thirds of the cardiovascular benefit appeared independent of weight loss (change in waist circumference mediated only about 33% of the effect). This is not full weight-independence — visceral adiposity mediates part of it. (Novo Nordisk–sponsored.)
• So what. There is a basis for reframing incretins as vascular and renal protectors rather than “weight-loss drugs.” But hard outcomes (all-cause mortality, MACE, kidney failure) must be kept separate from surrogates (albuminuria, KCCQ symptoms, decongestion). In SUMMIT (tirzepatide, HFpEF) the composite primary endpoint improved, yet cardiovascular death moved the wrong way (HR 1.58, wide confidence interval).
• Now what. The KDIGO 2026 draft places GLP-1 as a Level 1A additional pillar in T2D + CKD, layered on top of (not replacing) SGLT2 inhibitors and finerenone. The metrics to watch are tirzepatide’s completed CVOT (none exists yet) and three-pillar combination trials.

---

The five-minute read

The question is not “how much weight,” but “does it change hard outcomes?”

From a CKM (cardio-kidney-metabolic) perspective, the test of incretins is not weight but hard endpoints — cardiovascular events, kidney failure, death. The central reframe comes from SELECT’s pre-specified adiposity analysis (Lancet Diabetes & Endocrinology, October 2025): the reduction in MACE was unrelated to baseline obesity or to the amount of weight lost, and change in waist circumference mediated only about 33% of the effect (HR moving 0.80 → 0.86). In other words, roughly two-thirds of the benefit is not explained by weight, pointing to direct anti-inflammatory, haemodynamic and vascular mechanisms. This weakens the “weight-loss surrogate” narrative — though the effect is not fully weight-independent, since visceral adiposity is a partial mediator. (Novo Nordisk–sponsored.)

Do not blur hard and surrogate

In the kidney, FLOW (NEJM 2024, n = 3,533) is the strongest evidence: a 24% reduction in the primary kidney/CV composite (HR 0.76), an 18% reduction in 3-point MACE (HR 0.82), and — on top of these — a 20% reduction in all-cause mortality (HR 0.80). By contrast, SELECT’s renal signal (Nature Medicine 2024) was driven by surrogates such as albuminuria onset and eGFR decline, with hard kidney-failure events being rare. In HFpEF, SUMMIT’s (tirzepatide) primary composite was HR 0.62 (P = 0.026) but was driven by heart-failure worsening events, while cardiovascular death moved the opposite way (HR 1.58) — a decongestion signal, not a survival benefit.

Evidence-strength matrix — outcome (hard → surrogate) by drug

Domain	Character	Key figure	Agent
CV / MACE	Hard · SELECT · ~2/3 weight-independent	HR 0.80	semaglutide
All-cause mortality	Hard · FLOW / pooled	HR 0.80 / pooled 0.84	semaglutide
Kidney / CKD	FLOW = hard · SELECT = surrogate (albuminuria-driven)	FLOW HR 0.76	semaglutide
HFpEF	Decongestion / HF-event · CV death unfavourable	HR 0.62 (composite) · CV death HR 1.58	tirzepatide

A reshaping of standard care

The KDIGO 2026 draft places GLP-1 agents with proven CV/renal benefit at Level 1A for high-risk T2D + CKD patients, as an additive layer on top of SGLT2 inhibitors, RAS blockade and finerenone. Sequencing is set by eGFR (SGLT2i at eGFR ≥ 20, GLP-1 at any eGFR), but the key point is that this is a layer, not a replacement.

---

Deep dive

1. Background

Obesity, type 2 diabetes (T2D), chronic kidney disease (CKD) and heart failure are not separate diseases but a continuum captured by the CKM framework. Incretin agents (GLP-1 receptor agonists and GLP-1/GIP dual agonists) were initially assessed on glucose and weight surrogates, but recent large CVOT and renal trials have targeted hard outcomes directly, opening a serious discussion about their place in standard care. This piece splits that axis into: (a) whether each outcome is hard or surrogate, (b) whether the benefit is weight-mediated or direct, and (c) how the agents combine with SGLT2 inhibitors and finerenone.

2. What this work newly established

• Cardiovascular (SELECT, deepened): SELECT’s (n = 17,604) MACE reduction of 20% (HR 0.80) was confirmed in Part 0. Part 1 adds a layer from the pre-specified adiposity analysis (Lancet, Oct 2025): there was no linear relationship between 20-week weight loss and subsequent MACE, and waist-circumference change mediated only about 33% of the effect — quantitative evidence that roughly two-thirds of the benefit comes from pharmacology beyond weight. In the heart-failure-comorbid subgroup (4,286 patients, 53% HFpEF), MACE HR was 0.72 and CV death HR 0.76, consistent across HFrEF and HFpEF strata.
• Kidney (FLOW): FLOW (NEJM 2024, T2D + CKD, n = 3,533) showed a kidney/CV composite reduction of 24% (HR 0.76), MACE −18% (HR 0.82) and all-cause mortality −20% (HR 0.80), and was stopped early for efficacy. 79.6% had eGFR < 60, and results were consistent across KDIGO risk strata.
• Mortality (pooled): A 2026 meta-analysis (8 trials, 39,204 patients) reported all-cause mortality HR 0.84, CV death HR 0.83 and pooled renal HR 0.83 (more conservative than FLOW’s 0.76). This is a cross-trial pooling, not a head-to-head comparison.
• Guidelines: The KDIGO 2026 draft names GLP-1 as a Level 1A additional pillar.

3. Methodological strengths and limits

• Strengths: FLOW’s all-cause mortality and SELECT’s MACE are unambiguous hard endpoints. SELECT’s mediation analysis was pre-specified, which is more credible than post-hoc data dredging.
• Limit 1 — composite mixing: FLOW’s primary composite mixes hard (kidney failure, death) with surrogate (≥ 50% eGFR decline). Patients with eGFR < 25 were excluded, so G5 (end-stage kidney disease) is unaddressed.
• Limit 2 — surrogate-driven signal: SELECT’s renal composite (HR 0.78) was driven by albuminuria onset and eGFR decline, with hard kidney-failure events being rare. Merging FLOW (hard) with SELECT (surrogate) leads to over-interpretation.
• Limit 3 — subgroup concentration: FLOW’s mortality benefit was concentrated in the UACR ≥ 300 group (albuminuria interaction P =.01).
• Limit 4 — the nature of HFpEF: SUMMIT’s primary composite (HR 0.62, P = 0.026) was driven by heart-failure worsening events (HR 0.54), while cardiovascular death moved the opposite way (2.2% vs 1.4%, HR 1.58, wide CI). Decongestion markers improved (SBP −5 mmHg, estimated blood volume −0.58 L, CRP −37%), but the congestion marker NT-proBNP fell only 10.5% and was not significant (P = 0.07). In STEP-HFpEF (semaglutide), KCCQ and 6-minute walk distance are symptom / functional surrogate endpoints.

4. Neighbouring domains

Viewed as one continuum, CKM makes incretins a leading candidate for the first drug class to cross cardio, kidney and metabolic domains on hard outcomes. That said, the claim of combination heterogeneity with SGLT2 inhibitors (a null SGLT2i subgroup, HR 1.07) was underpowered (n = 550, 15.6%) and was REFUTED in verification — the combination is better read as additive. In the kidney domain, a three-pillar combination with finerenone (a non-steroidal MRA) is the next connection point.

5. Commercialization and market context (TRL, companies)

• TRL: Most of the agents are already approved and marketed (TRL 9). The novelty lies in new indications (kidney, heart failure) and guideline incorporation.
• Listed companies: semaglutide belongs to Novo Nordisk (NVO); tirzepatide to Eli Lilly (LLY). Many of the cited trials are manufacturer-sponsored, and all efficacy figures are attributed to the named trials and their journals.
• Real-world ranking (observational): incretins outperformed DPP-4 inhibitors (sitagliptin) on HF hospitalisation / death (sema HR 0.58, tirze HR 0.42), but a direct sema-vs-tirze superiority is unclear (HR 0.86, CI crossing 1). Being observational, this should be read as a signal, not causation.
• SURPASS-CVOT (tirzepatide vs dulaglutide, December 2025) was outside this verification scope; what it does and does not establish (vs placebo, not vs semaglutide) will be verified separately in Part 3.

6. The skeptic’s counterpoint

Mandatory caveats inherited from the original deep-dive’s skeptic gate (§8):

• Keep hard and surrogate separate. Merging hard outcomes (FLOW all-cause mortality, SELECT MACE) with surrogates (SELECT renal = albuminuria, STEP-HFpEF KCCQ, SUMMIT decongestion) leads to over-interpretation.
• SUMMIT cardiovascular death unfavourable. In HFpEF, cardiovascular death moved the wrong way (HR 1.58) — this must be stated alongside the composite. The basis for “it works in HFpEF” is symptoms and HF hospitalisation, not survival.
• Sponsorship attribution. Every cited trial is manufacturer-sponsored (Novo / Lilly), and quantitative claims are attributed to the respective trials and journals.
• Unverified item. SURPASS-CVOT is unverified in this part (Part 3 planned).
• Falsifiable predictions. (1) If tirzepatide demonstrates MACE superiority versus placebo in a completed CVOT, the “semaglutide CV lead” narrative weakens. (2) If mediation analysis explains most CV/renal benefit through weight loss, the “direct effect” narrative weakens.

7. What to watch

1. Tirzepatide’s completed CVOT (none exists yet) — whether it shows MACE superiority versus placebo.
2. Hard outcomes from a GLP-1RA + SGLT2i + finerenone three-pillar combination trial.
3. Whether FLOW’s mortality-benefit concentration in UACR ≥ 300 replicates in other cohorts.
4. Final adoption of the KDIGO 2026 draft and its sequencing recommendation.

---

References

• Cardiovascular Effects of Semaglutide by Baseline and Change in Adiposity (SELECT pre-specified adiposity analysis). The Lancet Diabetes & Endocrinology. October 2025. [DOI/URL per source asset — final DOI needs verification]
• Semaglutide in Patients with Type 2 Diabetes and Chronic Kidney Disease (FLOW). New England Journal of Medicine. 2024. [DOI/URL — needs verification]
• Kidney Outcomes with Semaglutide in Patients without Diabetes (SELECT renal analysis). Nature Medicine. 2024. [DOI/URL — needs verification]
• Tirzepatide for Heart Failure with Preserved Ejection Fraction and Obesity (SUMMIT). 2024–2025. [DOI/URL — needs verification]
• Pooled CKM meta-analysis (8 trials, n = 39,204). Cardiovascular Diabetology. 2026. [DOI/URL — needs verification]
• KDIGO 2026 Clinical Practice Guideline (draft), GLP-1 as a Level 1A pillar. kdigo.org. 2026. [URL — draft status needs verification]

Note: the bibliography above is inherited from the sources field of the original deep-dive asset. Individual DOIs/URLs and publication dates are to be confirmed against the primary sources. Copyrighted full text from Nature/Science/NEJM and similar is not redistributed; only summary, commentary and quotations ≤ 150 characters are included.

Disclosure

This post is for information only. It is not investment advice, and it is not medical advice — treatment decisions must be made in consultation with a qualified healthcare professional. The author holds no position in the named securities (NVO, LLY).

COI note: many of the clinical trials cited in this piece were conducted with sponsorship from the respective drug manufacturers (Novo Nordisk, Eli Lilly). The efficacy figures (for example SELECT MACE HR 0.80, FLOW all-cause mortality HR 0.80, SUMMIT composite HR 0.62) are all attributed to those manufacturer-sponsored trials and their publishing journals. This piece does not recommend the superiority of any manufacturer or drug, and notes that direct semaglutide-vs-tirzepatide comparison remains insufficiently evidenced (HR 0.86, CI including 1).