Bayesian Reanalyses of the Trials TOMAHAWK and COACT

Area of Science:

• Cardiology
• Clinical Trials
• Biostatistics

Background:

• Two large trials (TOMAHAWK, COACT) investigated immediate vs. delayed coronary angiography in out-of-hospital cardiac arrest (OHCA) patients without ST-elevation.
• Both trials yielded neutral results for short-term mortality, though TOMAHAWK suggested a non-significant trend towards harm with immediate angiography.
• Bayesian probabilistic analyses offer a potential method for deeper clinical interpretation of these trial findings.

Purpose of the Study:

• To reanalyze the TOMAHAWK and COACT trials using a Bayesian statistical framework.
• To explore the impact of different prior assumptions on the interpretation of immediate vs. delayed coronary angiography outcomes.

Main Methods:

• Post hoc Bayesian logistic regression analyses were conducted on data from the TOMAHAWK and COACT trials, both separately and combined.
• The primary endpoint was 30-day all-cause mortality.
• Analyses incorporated a range of priors, including "flat," "neutral," "optimistic," and "pessimistic" assumptions.

Main Results:

• Bayesian reanalysis of TOMAHAWK indicated a high posterior probability (90-97%) of increased 30-day mortality with immediate angiography across all priors.
• COACT showed odds ratios for 30-day mortality ranging from 0.98 to 1.11.
• Combined analysis of both trials demonstrated a high probability (83-95%) of increased mortality with immediate angiography, with all priors suggesting a trend towards harm.

Conclusions:

• Bayesian reanalysis suggests a high probability of increased 30-day mortality with immediate coronary angiography compared to delayed/selective angiography in OHCA patients without ST-elevation.
• These findings challenge the "neutral" interpretation of TOMAHAWK and COACT, indicating a potentially "harmful" effect of immediate angiography.
• A delayed angiography strategy may be preferable in clinical practice pending results from the DISCO trial.