Digital AI lifestyle support improves blood pressure control on top of standard dual antihypertensive therapy

Background

Early blood pressure (BP) control after treatment initiation remains challenging despite guideline-recommended fixed dual therapy. Lifestyle-related factors may influence early target attainment. AI-enabled digital behavioral support could enhance pharmacological treatment while maintaining standardized management.

Purpose

To assess whether AI-guided digital lifestyle support improves BP control when added to newly initiated fixed dual therapy with azilsartan medoxomil and indapamide.

Methods

In this prospective randomized controlled trial, 100 newly diagnosed or previously untreated hypertensive patients were enrolled at treatment initiation. All participants started fixed dual therapy with azilsartan plus indapamide according to a standardized protocol. Patients were randomized 1:1 to standard therapy alone or standard therapy combined with AI-guided digital lifestyle support for 12 weeks. Follow-up structure and titration criteria were identical in both groups.

The AI program provided personalized dietary plans, physical activity targets, sleep guidance, and medication reminders with adaptive weekly feedback.

The primary endpoint was change in systolic BP assessed by standardized 7-day home blood pressure monitoring (HBPM). Secondary endpoints included target BP <135 mmHg achievement, anthropometric and metabolic changes, adherence indices, and safety.

Results

Baseline characteristics were comparable. At 12 weeks, systolic BP reduction was greater in the AI-supported group (−13.1 ± 8.6 mmHg vs −10.2 ± 8.1 mmHg), with an adjusted between-group difference of −2.9 mmHg (p=0.04). Target HBPM systolic BP <135 mmHg was achieved in 68% vs 58% of patients (p=0.18).

Greater reductions in body weight (−2.6 ± 1.8 kg vs −1.4 ± 1.6 kg; p=0.03) and waist circumference (−4.8 ± 2.9 cm vs −2.9 ± 2.6 cm; p=0.02) were observed in the AI group. LDL cholesterol decreased by 12% vs 7% (p=0.04), triglycerides by 15% vs 9% (p=0.05), and fasting glucose by −0.34 vs −0.18 mmol/L (p=0.04). Medication adherence did not differ between groups, while lifestyle adherence improved significantly with AI support. No safety concerns were identified.

Conclusion

In newly treated hypertensive patients receiving fixed dual therapy, AI-guided digital lifestyle support provided modest additional BP reduction and improved metabolic parameters over 12 weeks. Digital behavioral augmentation may represent a scalable adjunct to early hypertension control.