Background and objectives The Mediterranean diet (MD) relies on a specific food consumption pattern that has been given pleiotropic effects on human health. However, even foods typically considered beneficial within the MD framework may vary in macronutrient composition, potentially leading to imbalanced or unhealthy dietary profiles despite overall adherence. To date, no study has evaluated the total energy intake and percentage macronutrient composition in a cohort of adult individuals stratified by their adherence to MD. We aimed to investigate the total energy intake and macronutrient composition among individuals stratified by adherence to MD, and its association with indices of adiposity and inflammation. Methods A cross-sectional study was conducted in 1342 volunteers. Participants underwent clinical examination (anthropometry, body composition, and high-sensitivity C-reactive protein assessment), and nutritional data collection, including adherence to MD (PREDIMED questionnaire) and habitual diet composition (7-day food records). The associations between PREDIMED score, macronutrient composition, and clinical outcomes were explored by Pearson’s correlation. Linear regression analysis was used to assess the macronutrient that best predicted adherence to MD. Results Participants with high adherence to MD (H-MD) presented lower intake of carbohydrate, while higher protein and fat amounts than individuals with medium (M-HD) or low (L-MD) adherence to MD. The H-MD group showed lower adiposity (body mass index, waist circumference, and fat mass) and C-reactive protein concentrations than individuals in the M-MD and L-MD groups p < 0.001 for all comparisons). Linear regression analysis showed low carbohydrate intake as the best predictor of high PREDIMED score (β: -0.875; p < 0.001). Conclusion H-MD is characterised by lower carbohydrate, higher protein and fat intakes than those usually reported in nutritional recommendations. This profile was significantly associated with lower adiposity and inflammation. A low carbohydrate amount was the best predictor of H-MD, suggesting that dietary carbohydrate should be on the lower side of the recommended range.