Aims/Introduction: The pathophysiology of diabetes differs between Asian and Western patients in many ways, and diet is a primary contributor. The present study examined the effect of diet on the efficacy of 25% insulin lispro/75% insulin lispro protamine suspension (LM25) and 50% insulin lispro/50% insulin lispro protamine suspension (LM50) as starter insulin in Chinese and Japanese patients with type 2 diabetes and inadequate glycemic control with oral antidiabetic medication. Materials and Methods: This was a predefined subgroup analysis of a phase 4, openlabel, 26-week, parallel-arm, randomized (computer-generated random sequence) trial (21 January 2013 to 22 August 2014). Nutritional intake was assessed from food records kept by participants before study drug administration. Outcomes assessed were changes from baseline in self-monitored blood glucose, 1,5-anhydroglucitol and glycated hemoglobin. Results: In total, 328 participants were randomized to receive twice-daily LM25 (n = 168) or LM50 (n = 160). Median daily nutritional intake (by weight and percentage of total energy) was 230.8 g of carbohydrate (54%), 56.5 g of fat (31%) and 66 g of protein (15%). Improvements in self-monitored blood glucose were significantly greater (P <= 0.028) in the LM50 group than in the LM25 group, regardless of nutritional intake. When carbohydrate (by weight or percentage energy) or fat (by weight) intake exceeded median levels, LM50 was significantly more efficacious than LM25 (P <= 0.026) in improving 1,5-anhydroglucitol and glycated hemoglobin. Conclusions: Glycemic control improved in both LM25 and LM50 groups, but LM50 was significantly more efficacious under certain dietary conditions, particularly with increased carbohydrate intake.