Fact checked byRichard Smith

Read more

January 12, 2023
2 min read
Save

AI model backs ‘the lower the better’ blood pressure strategy in patients with diabetes

Fact checked byRichard Smith
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Via deep learning modeling, researchers in the U.K. observed a monotonic relationship between elevated BP and CV events in patients with diabetes, and not a J-shaped relationship as observed in prior studies.

“Using large-scale, linked electronic health records, our investigation utilizing a deep learning model concluded that patients with diabetes in the lowest category of systolic BP of < 120 mm Hg exhibited the lowest cardiovascular risk,” Shishir Rao, MSc, DPhil student in deep medicine at the Oxford Martin School at the University of Oxford, U.K., and colleagues wrote. “This was in contrast to expert-dependent conventional approaches that have largely defended the paradoxical nonlinear J-shaped association between systolic BP and cardiovascular endpoints in those with diabetes. This difference is likely explained by the deep learning model’s ability to more comprehensively capture known and latent confounders in routine EHR, thus reducing the potential of important confounders being missed out.”

Diabetes Words 2019
Via deep learning modeling, researchers in the U.K. observed a monotonic relationship between elevated BP and CV events in patients with diabetes, and not a J-shaped relationship as observed in prior studies.
Source: Adobe Stock

The data were published in Hypertension.

To evaluate the association between systolic BP level and CV events in patients with diabetes, Rao and colleagues combined deep learning and the linked EHR data of 49,000 patients with diabetes in the U.K. (median age at baseline, 65 years; 45% women; 39% current or former smokers). CV events were defined as ischemic heart disease, HF, stroke and CV death.

Patients with lower systolic BP had higher prevalence of ischemic heart disease at baseline and less antihypertensive use compared with patients with higher systolic BP.

Average BMI was more than 29 kg/m2 across all BP ranges in the overall cohort.

During a median follow-up of 7.3 years, 16,378 CV events were recorded.

Compared with patients with diabetes and systolic BP less than 120 mm Hg, risk for CV events was greater in those with diabetes and higher systolic BP, and rose according to degree of BP elevation:

  • systolic BP between 120 mm Hg and 129 mm Hg (adjusted RR = 1.03; 95% CI, 0.97-1.1);
  • systolic BP between 130 mm Hg and 139 mm Hg (aRR = 1.05; 95% CI, 0.99-1.11);
  • systolic BP between 140 mm Hg and 149 mm Hg (aRR = 1.08; 95% CI, 1.01-1.15);
  • systolic BP between 150 mm Hg and 159 mm Hg (aRR = 1.12; 95% CI, 1.03-1.2); and
  • systolic BP 160 mm Hg or more (aRR = 1.19; 95% CI, 1.09-1.28).

“Our analysis provides some clarity concerning the relationship between systolic BP and cardiovascular events in patients with diabetes,” the researchers wrote. “Hence, while our investigation on its own is insufficient for recommending revisions of hypertension guidelines, our work rather functions as independent analyses complementing the findings of the individualized patient data meta-analysis of randomized evidences. Together, they support the lower the better paradigm of systolic BP in patients with diabetes.”