Prognostic performance of clinical scores and biomarkers in nonagenarians after TAVR requiring ICU admission

In this large binational retrospective cohort study using administrative database, we examined the prognostic value of clinical scores and routine biochemical markers in 587 nonagenarian and centenarian patients admitted to intensive care units across Australia and New Zealand following transcatheter aortic valve replacement (TAVR) between 2017 and 2024. ICU mortality was 1.5% and hospital mortality was 4.6%. We assessed clinical scores (APACHE III, SOFA) and biochemical markers (e.g., lactate, glucose, urea, albumin) for their ability to predict these outcomes, as well as length of stay. Both APACHE III and SOFA demonstrated excellent discriminative performance for short-term mortality (AUROCs > 0.90), though predictive capacity declined at longer follow-up. Elevated lactate, glucose, and urea were significantly associated with higher mortality, including glucose (adjusted OR 4.83, 95% CI 1.40 to 19.10, p = 0.015), lactate (4.38, 1.28 to 17.20, p = 0.001), and urea (3.67, 1.21 to 12.30, p = 0.028). Higher bicarbonate levels were protective (adjusted OR 0.30, 95% CI 0.09 to 0.98, p = 0.044). In adjusted spline models, several biomarkers exhibited non-linear risk patterns, indicating threshold effects. Each one-point increase in APACHE III predicted a 1% increase in ICU and hospital stay, whereas SOFA was associated only with ICU stay. These findings suggest that early clinical and biochemical profiling have potential to support prognostication, risk stratification, and resource planning for very elderly TAVR patients requiring postoperative intensive care.