medwireNews: A novel scoring system that could be used in routine clinical care can predict remission from type 2 diabetes 5 years after gastric bypass surgery, researchers report.
The 5-year Advanced-Diabetes Remission (5y-Ad-DiaRem) score uses machine learning algorithms to integrate medical history, bioclinical data, and diabetes treatments from patients who undergo Roux-en-Y gastric bypass surgery.
It had a negative predictive value of 90%, a positive predictive value of 80%, accuracy of 85%, and sensitivity of 90% in the retrospective study of 175 patients with type 2 diabetes who were severely obese at baseline (average BMI=46–48 kg/m2), and was further validated in three other cohorts.
“We propose to use the 5y-Ad-DiaRem score as an easily integrated tool to identify patients at risk for longer-term relapse during their 1-year follow-up assessments to propose strategies targeting optimized weight reduction and maintenance to maximize the length of their remission,” the researchers report in Diabetes Care.
“For example, increasing physical activity should be tested to evaluate whether it could reduce [type 2 diabetes] relapse as already seen in the reduced incidence of [type 2 diabetes] in glucose-intolerant patients,” they add.
The European team led by Judith Aron-Wisnewsky (Sorbonne Université, Paris, France) created the score by identifying the most easily accessible clinical variables that significantly differed at baseline and at 1 year after surgery in patients that experienced diabetes relapse, no diabetes remission, or diabetes remission (fasting plasma glucose <7 mmol/L and glycated hemoglobin [HbA1c] <6% without any glucose-lowering medication).
The score includes preoperative factors such as diabetes duration, number of glucose-lowering agents, and HbA1c levels as well as post-operative factors assessed at 1 year, including remission status, fasting plasma glucose, number of glucose-lowering agents, and the weight loss percentage from baseline to 1 year.
Among the 175 patients studied, who were aged an average of 46–52 years at baseline and had a mean diabetes duration of 4–12 years, the 5y-Ad-DiaRem accurately reclassified 13 out of 39 patients misclassified with the DiaRem scoring method and 12 of 38 misclassified with the advanced DiaRem method.
Importantly, 96% of patients with a score of 11 or less on the 5y-Ad-DiaRem – the threshold for 5-year diabetes remission – were free of antidiabetic medication at 5 years compared with just 47% of patients with a score above this.
The researchers confirmed the predictive ability of the 5y-Ad-DiaRem score in three independent cohorts from France, Italy, and Germany, which included patients with a wide range of age and diabetes severity before gastric bypass surgery, “thus suggesting its robustness,” they say.
“The use of machine learning methods together with the selection of features by physicians has again displayed its superiority compared with clinician experience,” conclude Aron-Wisnewsky and colleagues.
By Anita Chakraverty
medwireNews is an independent medical news service provided by Springer Healthcare. © 2018 Springer Healthcare part of the Springer Nature group
See also: