Cardiac Surgery · Thorac Cardiovasc Surg · 2025

Stent versus Trunk: Repairing a Torn Aorta

The Scenario

Acute type A aortic dissection is a tear you survive by hours. Two repair strategies compete for the arch: the frozen elephant trunk and a newer dissection stent.

Both devices live in observational surgical cohorts, never randomized against each other. The comparison had to come from carefully pooled single-arm evidence, with the bias assessment doing as much work as the pooling.

From Question to Answer

Clinical question

Trunk or stent: which repair carries less early mortality and fewer distal complications?

The evidence

Observational cohorts per device, small and heterogeneous, with extreme proportions

The design

Transformed proportion pooling per device, one auditable script per outcome, bias probed outcome by outcome

The answer

A device-by-device outcome profile, published with a title that made the editors smile

The Decisions That Mattered

Freeman-Tukey double arcsine for every pooled proportion.

Mortality and complication rates in these cohorts sit near the edges of the 0 to 1 scale, where raw pooling misbehaves. The transform keeps small cohorts from distorting the estimate.

One script per outcome, times main, sensitivity, and publication-bias versions.

Seven outcomes, each with its own reproducible chain. Any reviewer can rerun exactly the analysis behind any number in the paper.

ROBINS-I risk of bias woven into interpretation.

Observational surgical evidence is only as strong as its confounding control. The bias assessment decided how loudly each pooled result could speak.

Leave-one-out and fixed-versus-random comparisons per outcome.

In small evidence bases a single cohort can flip a conclusion. Every headline number was checked for that fragility before it was written down.

Overview

Problem

Two surgical devices for a lethal emergency, compared nowhere except across separate observational cohorts.

Approach

Freeman-Tukey proportion meta-analysis by device, a per-outcome script suite, ROBINS-I, and full stability checks.

Outcome

Published as Stent versus Trunk: Who Wins the Aortic Tug-of-War in Type A Dissection? in The Thoracic and Cardiovascular Surgeon, December 2025, with a companion conference poster.

Figures

Illustrative figure from simulated data; the real analysis stays with the journal and the clinical team.

Reproducible R Code

Per-outcome pooling and stability

# One pooled proportion per outcome, split by device
m <- metaprop(event = events, n = total,
              studlab = cohort,
              sm = "PFT", method.tau = "REML",
              subgroup = device)

# The stability suite that ran for every outcome:
metainf(m)                   # leave-one-out
update(m, common = TRUE)     # fixed vs random comparison
funnel(m); metabias(m)       # small-study effects, k permitting

Publication

Read the paper

Read publication ↗