The purpose of this study was to confirm reduced training metrics previously associated with a ketogenic low-carbohydrate (CHO) high-fat diet (LCHF) and investigate their attenuation with caffeine supplementation. At baseline, n 21 elite race-walkers followed a high CHO availability (HCHO) diet and performed a tempo hill session (14 km with a 450 m elevation gain). Athletes were then assigned to either the HCHO or LCHF in a parallel groups design for 3 weeks, where the 14 km tempo hill session was repeated each week. On weeks 2 and 3, in a randomised crossover allocation, all participants received 3 mg/kg caffeine or placebo (gum), 20 min before the session. Race-walking speed, heart rate, ratings of perceived exertion, blood metabolites and Stroop word-colour test metrics were collected. Although LCHF athletes walked faster at baseline compared with HCHO (P = 0·049), the HCHO group improved by week 2 (P = 0·009) and week 3 (P = 0·007), whereas the LCHF group was significantly slower in Week 1 (P < 0·001) and Week 2 (P = 0·026) compared with baseline. During the 14 km hill session, within-group analysis shows that athletes walked significantly faster (P = 0·010) and at a higher percentage of vVO2max (P = 0·007) when using caffeine compared with a placebo. Between-group differences remained present, with HCHO athletes walking at a higher percentage of vVO2max than those adhering to the LCHF diet (P = 0·035). No interaction between supplement treatment and dietary group occurred (P = 0·640). Caffeine supplementation partially reversed the performance impairment associated with an LCHF diet, but training quality remained lower than the combination of caffeine and high CHO availability.

The ketogenic diet, initially developed for epilepsy treatment, has gained attention as a potential intervention for neuropsychiatric disorders. A groundbreaking study by Campbell et al highlights its feasibility and potential efficacy in bipolar disorder, shedding light on shared mechanisms across neuropsychiatric disorders and the promise of metabolic treatment approaches.

Metabolic dysfunction has been long associated with severe mental illness (SMI), often viewed as a comorbidity to be managed. However, emerging evidence suggests that metabolic dysfunction, particularly at the mitochondrial level, may be a foundational element in the pathophysiology of neuropsychiatric disorders. This commentary expands on the current understanding by exploring the brain energy theory of mental illness, which posits that mitochondrial dysfunction is central to both metabolic and psychiatric conditions. The roles of insulin resistance, chronic stress and environmental factors are highlighted as shared biopsychosocial determinants that contribute to deterioration in both metabolic and mental health. The therapeutic potential of the ketogenic diet is discussed, particularly its ability to improve mitochondrial function and alleviate psychiatric symptoms. This shift in perspective, from viewing metabolic dysfunction as a secondary concern to recognising it as a root cause of SMI, has significant implications for clinical practice and research. By focusing on bioenergetic deficits and mitochondrial health, psychiatry may advance towards more effective, integrated treatment approaches that target the underlying cellular dysfunctions driving both metabolic and mental illnesses.