Bayesian Causal Atlas · Structural Causal Model

The Cachexia Oracle

A Judea Pearl structural-causal decomposition of wasting-syndrome interventions — sourced hazard ratios and trial effect sizes, backdoor/front-door-structured mediation, and a desired-risk-reduction Pareto engine for the minimum effective set.

Endpoint: cachexia progression (weight / lean-mass / appetite preservation) Interventions: Method: SCM + noisy-OR · Monte-Carlo CI Evidence: A–D graded, web-verified

Scope & safety — read first

Cachexia is a multi-organ wasting syndrome driven by the underlying disease (advanced cancer, heart failure, COPD, CKD). This tool is a research-grade evidence aggregator, not medical advice and not a treatment protocol. Critically: almost no cachexia-directed therapy has demonstrated a mortality benefit — most move only proximal endpoints (weight, lean mass, appetite). The disease burden is a confounder that drives both the wasting and death; reductions shown here are upper-bound ceilings on the proximal endpoint, not survival guarantees. Several agents (e.g. megestrol) carry real harms. Decisions belong with the treating oncology / cardiology / palliative team.

01

The causal model

Pearl SCM · DAG

The structure below is why cachexia is hard. The disease burden node U is a backdoor confounder: it drives systemic inflammation (and thus the wasting cascade) and drives mortality directly. Cachexia drugs act on the front-door mediators (inflammation → intake/catabolism → mass → function). If U keeps progressing, the front-door path is blocked at the exit — which is exactly why anamorelin lifts lean mass but not survival.

Root confounder (disease burden) Front-door mediators (drug targets) Distal endpoint (mostly unmoved)

Adjustment strategy. Backdoor path A ← U → Mortality is closed by conditioning on disease-directed control (node R, "treat the driver"). Front-door identification along A → inflammation → mass → function isolates each agent's mediated effect, removing the spurious correlation between "patients who get drug X" and "patients whose tumour happens to be indolent."

02

Desired risk reduction

Pareto target · minimum effective set

Set your target reduction in cachexia-progression risk. The engine performs a greedy Pareto search — adding interventions by marginal Pearl-adjusted contribution (redundancy-discounted within mechanism class) — and returns the smallest set that reaches the target. Apply it to the table, then tune by hand.

A high (RCT/meta) · B cohort · C case-series/modelled · D consensus. Lower grades dropped from combined RRR.
0%realistic ceiling ≈ 85%
0%
95% CI —
Achieved (current selection)
03

Interventions

check + dose · what-if engine

Each row is a node in the SCM. Toggle interventions and drag the dose slider (dose scales the effect toward its saturation ceiling). The combined readout above updates live — this is the what-if engine. RRR = relative reduction in cachexia-progression risk (proximal endpoint), an upper-bound ceiling derived from the cited trial effect size and discounted by evidence grade.

InterventionClassProximal effect (sourced) RRR ceilingGradeDoseRef

Mechanism classes: ROOT disease-directed · INFLAM anti-inflammatory / GDF-15 axis · OREX orexigenic / appetite · ANAB anabolic / muscle · SUBSTR nutritional substrate · NEURO neurohormonal. Effects within a class are correlated and redundancy-discounted; across classes they combine near-independently.

04

Pareto frontier

cumulative RRR · diminishing returns

Minimum-effective-set curve

Greedy marginal contribution · target line tracks slider

If-not-for (necessity)

Leave-one-out drop in combined RRR

Each bar is the counterfactual loss of combined risk reduction if that single intervention were removed from the current set — its necessity. Small bars indicate redundancy (another agent in the same class is covering it).

05

Sensitivity & cross-correlation

backdoor / front-door tuning

The two parameters below control how much shared causal pathway is removed when interventions overlap. ρ-within discounts agents that hit the same mechanism (e.g. two orexigenics fighting over the same appetite circuit); ρ-across applies a residual haircut between classes for the shared inflammatory backbone.

Correlation controls

how much overlap to subtract

What % of cross-correlation is appropriate? For cachexia, the inflammatory backbone (GDF-15 / IL-6 / TNF-α) is shared by most agents, so a defensible ρ-within sits at 0.40–0.70 (we anchor 0.55). ρ-across is smaller (0.10–0.25) because nutritional substrate, anabolic stimulus and disease control act through genuinely distinct mechanisms. Setting both to 0 (full independence) over-counts and is the optimistic extreme; the tornado at right shows the swing.

Tornado — combined RRR vs assumptions

current selection across correlation extremes

06

Uncertainty

Monte-Carlo · 4,000 draws

Distribution of combined RRR

propagating each trial's confidence interval

Read-out

point estimate & dispersion

Mean RRR
Std deviation
2.5th pct
97.5th pct

Each intervention's effect is drawn from a truncated normal whose spread is set by its evidence grade (A σ≈0.20·r, B 0.28, C 0.38, D 0.50 — wider for weaker evidence). The draws are recombined through the same noisy-OR each iteration, so the interval reflects both trial imprecision and structural correlation, not a naïve sum.

07

Dose response & saturation

where more stops helping

Cachexia agents saturate — and some reverse above their optimum. The dose sliders in §03 follow these curves: effect climbs toward a ceiling, then plateaus.

Saturation curves

modelled effect vs dose fraction

Agent-specific notes

sourced dose findings

PonsegromabMonotonic 100→200→400 mg: +1.22 / +1.92 / +2.81 kg vs placebo — still ascending at 400 mg, no plateau reached in the GDF-15-high population [10].
Omega-3 (EPA)Inverted: 2 g/day showed a clinically relevant signal; 4 g/day gave no added benefit — saturation / tolerability ceiling [8].
OlanzapineLow-dose 2.5–5 mg is optimal; higher antipsychotic doses add sedation/metabolic toxicity without appetite gain [2].
Megestrol≈800 mg/day optimal; higher doses trended to weight loss and raised thromboembolism — a true reversal [9].
Anamorelin100 mg plateaus lean-mass gain (~1 kg); handgrip never responds at any dose — a ceiling on the functional axis [3].
08

Antithesis

arguing against the tool

The endpoint is soft

Weight and lean mass are surrogates. Anamorelin (Grade A for lean mass) moved neither handgrip nor survival [3]; MENAC stabilised weight but not muscle or activity [6]. A "30% RRR on cachexia progression" may buy no survival or function. The honest headline metric for patients is QoL and independence, which most agents barely touch.

Confounding may be irreducible

Front-door identification assumes the mediator (inflammation→mass) fully transmits the effect and that U doesn't have an unblocked direct path to the endpoint. In refractory cachexia the tumour drives wasting through pathways no current drug touches, so the adjustment is optimistic — the real ceiling is lower than 85%.

The biggest lever isn't a drug

Breakthrough hook: the only consistent mortality mover is controlling the driver — effective anticancer therapy or guideline-directed HF therapy (ACEi/β-blocker reduce weight-loss risk and death [11,12]). The ROOT node dominates the Pareto frontier; everything else is supportive. A tool that ranks supportive agents must not obscure that hierarchy.

09

Methods & assumptions

appendix · reproducible
EndpointComposite cachexia-progression risk: probability of continued clinically significant weight / lean-body-mass loss + anorexia persistence over a 12–16 week horizon. Not overall-survival HR (which is mostly null for these agents).
RRR ceiling derivationFor each agent, the published proximal effect size (Δ weight kg, Δ LBM kg, responder %, or weight-loss-risk reduction) is normalised against the placebo-arm natural history and capped, then multiplied by an evidence-grade factor (A 1.0, B 0.85, C 0.70, D 0.45). Result framed as an upper bound, not an expected value.
Dose modelEffect(dose) = ceiling × saturating function with agent-specific shape; EPA and megestrol use a non-monotone (peaked) curve to encode reversal at high dose.
Backdoor adjustmentDisease-burden confounder closed by conditioning on the ROOT node (disease-directed control). The ROOT node carries the only effect with a credible survival path.
Front-door combinationWithin class: rank by effect, k-th agent contributes r·(1−ρ_within)^(k−1). Class aggregate R = 1−∏(1−r_eff). Across classes: rank by R, j-th class contributes R·(1−ρ_across)^(j−1); combined = 1−∏(1−R_eff). Capped at 0.85.
Pareto searchGreedy: repeatedly add the unselected intervention with the largest marginal increase in combined RRR; stop at target or when no positive marginal remains. Ties broken by lower patient burden.
Monte-Carlo4,000 iterations; each r_i ~ TruncNormal(r_i, (grade σ)·r_i) on [0, ceiling]; recombined via the noisy-OR each draw; report mean, SD, 2.5/97.5 percentiles.
Cross-correlation defaultρ_within = 0.55, ρ_across = 0.15 (defensible range 0.40–0.70 / 0.10–0.25). Independence (0/0) is the optimistic extreme.
LimitationsEffect sizes pool heterogeneous tumour/HF populations and definitions (Fearon 2011 vs ICC vs weight-loss-only). Surrogate endpoints. No head-to-head trials for most pairings — combination effects are modelled, not measured.
10

References

web-verified · A–D landmark
  1. Prevalence & survival with cachexia (meta-analysis). Cancer cachexia associated with OS HR 1.58 (Fearon 2011 criteria) to 2.78 (other criteria); weight loss during treatment HR 2.20. Adv Nutr 2024. Grade A
  2. Olanzapine RCT (Sandhya et al). 2.5 mg/day vs placebo, n=124: >5% weight gain in 60% vs 9%; improved appetite/QoL. J Clin Oncol 2023;41:2617. ASCO-endorsed. Grade B
  3. Anamorelin (ROMANA 1 & 2). n=979: LBM +1.1/+0.75 kg, weight +2.2 kg, improved anorexia symptoms; no handgrip, no survival difference. Lancet Oncol 2016;17:519. Grade A
  4. Espindolol (ACT-ONE). Phase II, n=87: +0.54 kg/4 wk vs −0.21 placebo, ↑LBM, ↑handgrip; no survival difference. J Cachexia Sarcopenia Muscle 2016;7:355. Grade B
  5. Exercise (systematic reviews). Aerobic+resistance: positive direction on body composition 75%, strength 80%, function 64%; safe. Clin Nutr ESPEN 2023–25. Grade B
  6. Multimodal MENAC (phase III). Exercise+nutrition+EPA+NSAID: weight stabilised (Δ +1.04 kg vs SoC, p=0.04); no muscle/activity difference. J Clin Oncol 2024;42:LBA12007. Grade B
  7. Nutritional support post-discharge. Improved handgrip; mortality HR 0.69 (0.50–0.96) in one RCT. Clin Nutr 2023. Grade B
  8. Omega-3 / EPA (meta-analyses). Improves weight + QoL (NSCLC) but not lean mass; 12-RCT pooled: QoL + survival signal, no weight effect; 2 g > 4 g/day. Integr Cancer Ther 2024; Clin Nutr 2022. Grade C
  9. Megestrol acetate (Cochrane 35 trials). Weight +2.25 kg (fat, not LBM), appetite NNT≈4; no QoL/LBM/survival benefit; ↑thromboembolism/oedema; deaths RR 1.26 (0.70–2.27). Cochrane 2013/2022. Grade C
  10. Ponsegromab (anti-GDF-15, phase II). n=187, GDF-15≥1500: +1.22/+1.92/+2.81 kg (100/200/400 mg) vs placebo, 5.6% at top dose; ↑activity, muscle, appetite. N Engl J Med 2024;391:2291. Grade B
  11. ACE inhibitor in cardiac cachexia (SOLVD reanalysis). ≥6% weight loss HR 2.10 for death; enalapril cut weight-loss risk 19%; GDMT improves survival. Lancet 1999. Grade B
  12. Beta-blocker in cardiac cachexia. Carvedilol prevents weight loss, ↓mortality ~35% in HFrEF (COPERNICUS context). Cardiol Rev 2024 review. Grade B
  13. Enobosarm (SARM, phase III). Lean-mass gain inconsistent across two phase-III trials; stair-climb power not met. Conference data. Grade C
  14. Mirtazapine RCT. No better than placebo for cancer anorexia. Grade D
  15. Corticosteroids (ASCO guideline). Short-term appetite gain only; significant long-term toxicity. JCO guideline 2020/2023. Grade C
11

Custom report

print / export
Current selection & achieved RRR are captured in the printout header.