Modeled treatment preference between observation and coiling for unruptured intracranial aneurysms in octogenarians was conditional on endpoint definition and key structural assumptions, and these model outputs do not establish treatment recommendations for individual patients.
Key Findings
Results
Under the primary composite endpoint, observation was favored in every age-90 panel and in all male panels at age 85 or older.
The primary outcome was the undiscounted lifetime cumulative incidence of aneurysm- or treatment-attributable events as a first-event composite.
The four composite components were treatment morbidity, treatment death, rupture death, and rupture poor outcome, carrying equal weight regardless of severity, duration, or reversibility.
Median differences favored observation in every age-90 panel and in all male panels at age ≥ 85.
The probability that coiling was favored ranged from 0.000 to 0.984 across all panels.
Results
Under the primary composite endpoint, coiling was favored in selected age-80 panels and selected age-85 female panels.
The probability that coiling was favored ranged from 0.000 to 0.984 across all panels.
Coiling benefit was concentrated in younger octogenarians (index age 80) and in female patients at age 85.
The model compared endovascular coiling with observation, with microsurgical clipping and stent-assisted coiling or flow diversion as additional scenarios.
Results
When a severity-weighted quality-adjusted analysis with 3% discounting was applied, 21 of 30 endpoint-direction comparisons were concordant with the original endpoint results.
A 24-panel, 3%-discounted severity-weighted quality-adjusted analysis was conducted alongside the primary analysis.
Across 30 endpoint-direction comparisons, 21 were concordant between the two endpoint definitions.
All nine discordant comparisons shifted from observation toward coiling relative to the primary endpoint.
Eight of the nine discordant comparisons had 95% quality-adjusted uncertainty intervals spanning zero, indicating statistical uncertainty.
Results
Structural sensitivity analyses produced median-direction changes confined to modified Rankin Scale-based panels with reference quality-adjusted differences near zero and uncertainty intervals spanning zero.
Structural sensitivities were executed for treatment-morbidity recovery, residual rupture risk, and coiling morbidity.
Median-direction changes were confined to modified Rankin Scale-based panels.
These panels had reference quality-adjusted differences near zero and uncertainty intervals spanning zero, indicating minimal clinical certainty in those scenarios.
Residual post-treatment rupture risk was fixed at 10% of the natural-history rate in the base case.
Methods
Periprocedural coiling morbidity and mortality inputs were anchored to a U.S. National Inpatient Sample analysis covering 2001–2008, representing historically derived and potentially outdated treatment-risk estimates.
The U.S. National Inpatient Sample data spanned 2001–2008.
The authors identified this as a key structural assumption and limitation affecting model outputs.
The model was applied to Korean patients at index ages 80, 85, and 90 years.
The authors noted that 'historical treatment-risk estimates and uncalibrated residual post-treatment rupture risk' were among the key conditional assumptions.
Methods
The competing-risk Markov decision model was constructed to account for the interaction of treatment morbidity, rupture risk, post-rupture outcome, and competing mortality in octogenarians.
Management of unruptured intracranial aneurysms in patients aged 80 years and older remains uncertain due to these interacting factors.
The model compared observation with endovascular coiling as primary strategies.
Microsurgical clipping and stent-assisted coiling or flow diversion served as additional scenario comparisons.
The primary outcome was the undiscounted lifetime cumulative incidence of the composite first-event endpoint.
What This Means
This research used a mathematical modeling approach to compare two strategies for managing brain aneurysms that haven't ruptured yet in patients aged 80 and older: watchful waiting (observation) versus endovascular coiling, a minimally invasive procedure to seal off the aneurysm. The model was applied to hypothetical Korean patients at ages 80, 85, and 90, and it tracked outcomes including death or disability from treatment itself, death or poor recovery from a rupture, and natural death from other causes. The study found that whether observation or coiling appeared better depended heavily on which outcome measure was used, the patient's age and sex, and key assumptions built into the model—such as how risky the coiling procedure is and how much it reduces future rupture risk.
In general, the model suggested observation was more favorable for patients aged 90 and for men aged 85 or older, while coiling appeared more favorable for some patients aged 80 and some women aged 85. However, when the researchers applied a more nuanced analysis that weighted outcomes by their severity and adjusted for time, most of the cases where the two approaches shifted in favor of coiling still had results that were statistically uncertain (confidence intervals crossing zero). The treatment-risk data used in the model came from U.S. hospital records from 2001–2008, which may not reflect current procedural safety or outcomes in contemporary practice.
This research suggests that there is no single clear answer about whether to treat or observe unruptured brain aneurysms in very elderly patients—the best approach appears to depend on individual patient factors, the specific outcomes that matter most, and assumptions about procedural risk that may vary across centers and over time. The authors explicitly state that these model outputs are not intended to serve as treatment recommendations for individual patients, but rather to illustrate how sensitive conclusions are to the choice of outcome measure and model assumptions.
Yoo H, Kim J, Chung S. (2026). Observation versus coiling for unruptured intracranial aneurysms in octogenarians: A decision-analytic modeling study.. PloS one. https://doi.org/10.1371/journal.pone.0357374