Cassy Dingena¹, Elisabeth den Brok¹, Nefeli Dimitropoulou¹, Stephan Proennecke², Stavroula Mougiakakou³, Peter Mertens⁴, Andriani Vazeou⁵, Konstantinos Makrilakis⁶, Ulrik Pedersen-Bjergaard^7,8, Bastiaan de Galan^1,9,10

(1) CARIM School for Cardiovascular Disease, Maastricht University, Maastricht, Netherlands.

(2) Debiotech SA, Lausane, Switzerland.

(3) ARTORG Center for Biomedical Engineering Research, University of Bern, Bern, Switzerland.

(4) Department of Kidney and Hypertension Diseases, Diabetology and Endocrinology, Otto-Von-Guericke-Univeristat Magdeburg, Magdeburg, Germany.

(5) P & A Kyriakou´ Children’s Hospital, Athens, Greece.

(6) Diabetes Center, National and Kapodistrian University of Athens, Athens, Greece.

(7) Department of Endocrinology and Nephrology, Copenhagen University Hospital – North Zealand, Hillerod, Denmark.

(8) Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.

(9) Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, Netherlands.

(10) Department of Internal Medicine, Radboud University Medical Centre, The Netherlands.

Aims

Despite considerable progress in diabetes technology, maintaining optimal glycaemic control remains challenging for people with diabetes receiving intensive insulin treatment. The MELISSA trial aims to clinically validate an artificial intelligence (AI)-driven smartphone application to support people on multiple daily injections (MDI) with personalised insulin dose recommendations and a novel approach for carbohydrate estimation.

Methods

The MELISSA trial, a 22-week European multi-centre randomised open-label blinded endpoint trial, aims to enrol people with type 1 (n=402) and type 2 diabetes (n=90) on MDI. This trial investigates the MELISSA application, which integrates two AI-driven features: an adaptive basal-bolus advisor (ABBA) and an automated nutrition assessment system (goFOOD^TM). ABBA provides personalised insulin dose recommendations, and goFOOD^TM converts food images into carbohydrate content estimations used by ABBA. Participants will be randomized 1:1 to the MELISSA application or to continuation of usual care.

Outcomes

The primary outcome is the change in time spent in glucose target range (3.9-10.0mmol/l) between trial arms, assessed from baseline to study completion. Secondary outcomes include other sensor-derived glucometrics (e.g., times below and above range, and glycaemic variability), biomarkers in blood (e.g., haemoglobin A1c, oxidative stress and inflammation) and urine (e.g., albumin-creatinine ratio and Advanced Glycation End-products), patient-reported-outcomes (e.g., diabetes distress and fear of hypoglycaemia), and safety parameters (e.g. hypoglycaemia). Approval was granted by the azM/UM Medical Ethics Research Committee (NL-009099), and recruitment has started.

Conclusion

MELISSA may improve glycaemic control quality-of-life and reduce complications in people with diabetes on MDI. Trial data will be used to obtain Conformite Europeenne certification.