AIMS: To review systematically and quantify the weight loss achieved through low- (LEDs) and very-low-energy diets (VLEDs) in people with type 2 diabetes mellitus (T2DM). MATERIALS AND METHODS: Studies reporting the effects of diet-only interventions of up to 1600 kcal/d in people with T2DM were searched in MEDLINE, EMBASE and CINAHL up to July 2018. Changes in the primary (body weight and body mass index [BMI]) and secondary outcomes (glycated haemoglobin, blood lipids) according to energy restriction and duration of diet were modelled using restricted cubic splines. RESULTS: Forty-four studies (3817 participants) were included. The overall quality of the evidence was moderate and limited to short-term interventions up to 4 months. Baseline mean weight and BMI were 92.1 kg and 36.6 kg/m(2) . VLEDs of 400 kcal/d led to 5.4% weight loss at 2 weeks, increasing to 17.9% at 3 months. More modest reductions of 7.3% were observed on LEDs of 1200 kcal/d and 2.0% on 1600 kcal/d after 3 months. No clear patterns emerged for secondary outcomes. Publication bias was significant for primary outcomes. CONCLUSIONS: Through modelling, we were able to describe effective dietary deficit strategies to achieve weight reduction up to 4 months in people with T2DM. High-quality studies are required to further support clinical practice with evidence-based dietary interventions.