Z
ZEUS Assumption ExplorerPublic-data scenario laboratory
Audited equations

Scenario analysis, not ZEUS trial data. No blinded or unblinded patient-level results are used. Every output is a deterministic consequence of public inputs and the assumptions you drag below.

Method 1

SGLT2i + GLP-1RA evidence → control-rate range

Independent background-therapy reconstruction

Control-arm annual MACE hazard6.66–6.99%

All four uptake corners evaluated. Low: SGLT2i 65.0% / GLP-1RA 30.0%; high: SGLT2i 36.9% / GLP-1RA 11.3%.

Midpoint scenario6.82%

A transparent midpoint of the selected uptake bands, not a probability-weighted estimate.

Expected control events by Jun 9, 2026Loading

Carried through the current regional calendar.

Primary output

Control-group event-rate range

6.66–6.99%/yr
6.4%6.6%6.8%7.0%7.2%

This route starts with an external high-risk CKD MACE anchor, splits it into cardiovascular death, myocardial infarction and stroke, then applies background-therapy effects component by component. It is a plausibility range—not an observed ZEUS control rate.

Midpoint decomposition

Where the adjusted rate comes from

Cardiovascular death2.64%
Myocardial infarction2.39%
Stroke1.80%
Cross-check

Uptake sensitivity grid

Columns: SGLT2 inhibitor use
1%25%50%75%100%1%25%50%75%100%

Rows: GLP-1 receptor agonist use. The 5 × 5 grid spans the full 1%-100% technical envelope. Click a cell to collapse both uptake bands to that point.

Component-based sensitivity analysis

Method 1: mathematical specification

A deterministic bottom-up reconstruction. The external MACE anchor is decomposed into mutually exclusive MACE-3 limbs, adjusted for background-treatment penetration and limb-specific treatment effects, then recombined into a control-arm cause-specific hazard range.

Formula stack
1
Raw limb hazardλ⁰ⱼ = a × wⱼ

Allocate the historical composite MACE hazard a to cardiovascular death, myocardial infarction and stroke using shares wⱼ.

2
Background-therapy deflatorGⱼ = [1 − pS(1 − HRS,ⱼ)] × [1 − pG(1 − HRG,ⱼ)]

Mix treated and untreated background-care strata for SGLT2 inhibitor and GLP-1RA use within each MACE limb.

3
Adjusted control limbλctrl,ⱼ = λ⁰ⱼ × Gⱼ

Apply the limb-specific background-care multiplier to each raw cause-specific hazard.

4
Composite control hazard and rangeλctrl,MACE = Σⱼ λctrl,ⱼ ; Iλ = [min λ(pS,pG), max λ(pS,pG)], (pS,pG) ∈ {L,U}²

Evaluate low-low, low-high, high-low and high-high. This remains valid when one background-therapy profile contains HRs above 1.

Analysis method
  1. Define the estimand

    Annual control-arm cause-specific hazard for first strict 3-point MACE, not cumulative incidence and not an observed placebo rate.

  2. Transport the external anchor

    Start from the selected high-risk CKD MACE anchor and split it using the selected CV-death / MI / stroke composition.

  3. Apply explicit background-care axes

    Use the low and high SGLT2i and GLP-1RA uptake handles plus the selected limb HR profiles; no benefit is hidden inside a single adjusted rate.

  4. Propagate the interval

    Evaluate all four uptake corners, retain the true minimum and maximum with their corner provenance, then recompute the midpoint and 1%-100% grid. The interval becomes Method 3's input.

Notation

aexternal annual MACE hazard anchor

wⱼshare of MACE events from limb j

pS, pGSGLT2i and GLP-1RA penetration

HRS,ⱼ / HRG,ⱼbackground-therapy limb hazard ratios

Iλreconstructed control-hazard interval

Interpretation limit. The mixture deflator is an instantaneous-hazard / rare-event approximation. Its product form assumes independent marginal SGLT2i and GLP-1RA use, expected dual-use overlap pS × pG, and multiplicative limb HRs in dual users. The component HRs are transported from external evidence, uptake can change over calendar time, and the 1%-100% envelope is technical sensitivity analysis rather than a confidence interval, credible interval or uptake forecast.