Updated risk equations for type 2 diabetes complications published
medwireNews: Researchers have developed and validated new risk equations to help predict macrovascular and microvascular complications in patients with type 2 diabetes.
Sanjay Basu (Stanford University, Palo Alto, California, USA) and co-developers have made their Risk Equations for Complications Of type 2 Diabetes (RECODe) available as an online calculator. The equations predict the 10-year risk for atherosclerotic cardiovascular (CV) disease, myocardial infarction, stroke, congestive heart failure, and CV and all-cause mortality, as well as renal endpoints, including albuminuria and end-stage renal stage, and retinopathy endpoints, including cataract extraction, vision loss, and need for photocoagulation or vitrectomy.
The team used individual patient data from more than 15,000 participants of ACCORD, Look AHEAD, and DPPOS to develop and externally validate RECODe. The equations proved more accurate than existing models, correctly reclassifying a large number of patients who were incorrectly classified by the UKPDS Outcomes Model 2 and ACC/AHA Pooled Cohort Equations.
This improved classification of patients’ risk was “particularly driven by the fact that older equations tended to overestimate risk of people who were actually low risk, whereas RECODe correctly identified people at low risk,” Basu and colleagues write in The Lancet Diabetes & Endocrinology.
In a linked commentary, Coen Stehouwer (Maastricht University Medical Centre, the Netherlands) notes that current risk models are based on relatively old data, since when medical advances have resulted in reduced CV event rates in diabetes patients. He therefore describes the research as “an important advance.”
However, he highlights the challenge of making “risk estimation for patients with type 2 diabetes more accurate, more complete, and more balanced.”
Despite being based on a large number of variables and successfully reclassifying many patients, RECODe had average C-statistics of around 0.6 to 0.7 (where 1.0 is perfect prediction and 0.5 is no better than chance), which Stehouwer says is “far from ideal.” He suggests that improved accuracy based on baseline measurements “might simply be too much to ask,” and calls for study of equations that incorporate time-varying factors.
He also emphasizes the need to include outcomes such as neuropathy and amputation, account for the influence of factors such as cognitive impairment and depression, and consider adverse medication effects.
In addition, Stehouwer notes the large number of predictive variables included in RECODe, encompassing demographics, clinical features, medication use, and biomarkers. He says: “Another question which remains to be addressed is whether simpler models perform equally well, which is of particular importance in health-care systems with limited resources.”
medwireNews is an independent medical news service provided by Springer Healthcare. © 2017 Springer Healthcare part of the Springer Nature group