An AI model supports a ‘lower is better’ blood pressure strategy in patients with diabetes

Through deep learning modeling, researchers in the UK observed a monotonic relationship between hypertension and CV events in diabetic patients, rather than a J-shaped relationship as observed in previous studies.

Using large-scale, interconnected electronic health records, our investigation using a deep learning model concluded that patients with diabetes in the United States The lowest category of systolic BP <120 mm Hg showed the lowest cardiovascular risk“,” Shishir Rao, MA, a DPhil Student in Deep Medicine at the Oxford Martin School at the University of Oxford, UK, and colleagues. This was in contrast to traditional expert-based approaches that have largely defended the paradoxical nonlinear J-shaped association between systolic blood pressure and cardiovascular endpoints in patients with diabetes. This difference is likely explained by the ability of the deep learning model to capture known and latent confounders more comprehensively in routine electronic health records, thus reducing the likelihood of missing important confounders.”

The data was published in Hypertension.

To assess the association between systolic blood pressure level and CV events in patients with diabetes, Rao and colleagues collected deep learning and associated electronic health record data for 49,000 UK diabetic patients (mean age at baseline, 65 years; 45% women; 39% smokers). current or former). CV events were defined as ischemic heart disease, HF, stroke and CV death.

Patients with lower systolic blood pressure had a higher prevalence of ischemic heart disease at baseline and lower blood pressure utilization than patients with higher systolic blood pressure.

The mean BMI was over 29 kg/m² across all BP ranges in the overall cohort.

During an average follow-up of 7.3 years, 16,378 CV events were recorded.

compared with diabetic patients Systolic blood pressure less than 120 mm Hgthe risk of CV events was greater in patients with diabetes and systolic hypertension, and increased according to the degree of hypertension:

• systolic blood pressure between 120 mmHg and 129 mmHg (adjusted RR = 1.03; 95% CI, 0.97–1.1);
• systolic blood pressure between 130 mm Hg and 139 mm Hg (ratio systolic blood pressure = 1.05; 95% confidence interval, 0.99–1.11);
• systolic blood pressure between 140 mmHg and 149 mmHg (PR = 1.08; 95% confidence interval, 1.01–1.15);
• systolic blood pressure between 150 mmHg and 159 mmHg (barometric pressure ratio = 1.12; 95% confidence interval, 1.03–1.2); And
• Systolic blood pressure of 160 mm Hg or greater (ratio systolic blood pressure = 1.19, 95% confidence interval, 1.09–1.28).

“Our analysis provides some clarity regarding the relationship between systolic blood pressure and cardiovascular events in patients with diabetes,” the researchers wrote. “Hence, while our investigation by itself is insufficient to recommend revisions of the antihypertensive guidelines, our work rather serves as independent analyzes that complement the results of the individual patient data meta-analysis of the randomized evidence. Together, they support the lower best model for systolic BP in patients with diabetes.”