Precision medicine approach to detect obese people who are at high risk of developing diabetes

It is well-recognized that obesity predisposes an individual to an increased risk of developing type 2 diabetes. However, a variety of factors can play a role in the risk of developing type 32 diabetes. Various strategies have been utilized to help identify obese individuals who are at high risk of developing type 2 diabetes so as appropriate intervention can be prioritized to reduce the risks of developing obesity related complications. Precision medicine approach has attracted significant amount of clinical and research interests to help predict, prevent, diagnose and manage patients with a variety of conditions.

Stratification of Obesity Phenotypes to Optimise Future Obesity Therapy (SOPHIA) is a European Union-funded innovative medicine initiative (IMI) to help develop tests and therapies which may allow the prediction of risk of obesity related co-morbidities and the prediction of response to obesity treatments. A recent publication in the journal Nature Medicine from the IMI SOPHIA consortium have reported and described a new precision prediction algorithm that distinguish subpopulations where cardiometabolic risk differs from the risk expected for their given body mass index (BMI).¹ This is important because multiple factors are in play when determining an obese person's individual risk of developing type 2 diabetes and heart disease. For example, while BMI is the common metric used by epidemiologists, health professionals and others to characterize obesity, it is insufficient for accurate classification of the disease of obesity at an individual level because people with similar BMIs often exhibit different health risks. This is partially because BMI is an imperfect measure of excess adiposity as it does not distinguish the proportion or distribution of fat mass and fat-free mass in the body.

The research was led by scientists at Lund University Diabetes Centre in Sweden, and Maastricht Centre for Systems Biology and Erasmus MC University Medical Centre in The Netherlands, in collaboration with other researchers from the IMI SOPHIA consortium. The study focused on clinical data of 170 000 adults derived from the UK Biobank, The Rotterdam Study, the Maastricht study and the Gutenberg Health study. Machine learning was then utilized to develop algorithm that would split obesity into five subtypes based on different diagnostic profiles, each with different risks of developing obesity related complications. The five phenotypic profiles consists of individuals with cardiometabolic biomarkers higher or lower than expected based on their BMI, which generally increases disease risk, representing 20% of the total population. Conversely, the study showed that 80% of people had health markers that matched their cardiometabolic risk expected for their BMI. The discordant phenotype identified for example 8% of women with higher blood pressure than expected for their weight but associated with higher protective HDL and lower waist-hip-ratio (WHR). In addition, 5% had abnormal liver enzymes and high WHR for their BMI; 4% had higher level of inflammatory markers than expected for their BMI and approximately 2.5% had higher blood sugar and lower LDL for their BMI. Furthermore, around 5% and 7% of women and men respectively had higher LDL cholesterol, triglyceride, WHR and blood pressure for their BMI. The enhanced algorithm derived from this study was reported to represents an additional net benefit of 4–15 additional correct interventions and 37–135 additional unnecessary interventions correctly avoided for every 10 000 individuals tested. The study provided evidence of the important role of precision medicine to more accurately identify cardio-metabolic risks.