Page 17 - hypertension_newsletter

REFLECTIONS
Hypertension
Hypertension Global Newsletter #10 2026

Graphical abstract Hypertension

The authors conducted a systematic literature search to evaluate the current applications of AI in CV pharmacotherapy across
major therapeutic domains, including hypertension, diabetes, dyslipidaemia, thrombosis, coronary artery disease (CAD), HF,
and arrhythmias. They also assessed the emerging role of AI in in silico computational modeling and clinical trial design. The
researchers synthesized the evidence narratively, highlighting both the successes of existing ML algorithms and the critical
gaps that currently limit their safety and real-world implementation.

The review found that AI excels at personalizing treatment by predicting individual therapeutic responses. For example, ML
models successfully analysed trial data to identify specific patient clusters that derive the highest CV benefit from intensive BP
control or specific diabetes drugs (like sodium-glucose cotransporter 2 [SGLT2] inhibitors). This shifts the prescribing paradigm
from simply treating “high-risk” patients to intentionally targeting “high-benefit” patients. AI is also proving valuable in predicting
drug-related adverse events and exploring drug repurposing. AI platforms utilizing the IBM MarketScan database identified
that specific antihypertensive combinations, such as pairing ACEis with diuretics or ARBs with CCBs, are associated with a
significantly lower risk of developing Parkinson’s disease.

n the context of CAD and thrombosis, AI risk-prediction tools like the PRAISE score offer highly individualized forecasting for
ischemic and bleeding events, helping clinicians safely tailor the intensity and duration of antiplatelet therapies. ML models also
outperformed human physicians in predicting optimal warfarin discharge doses during internal validation, though they struggled
to maintain this accuracy externally. Beyond direct patient care, generative AI and deep neural networks are transforming CV

TABLE OF CONTENTS