Given the prevalence of MDD among university students [40], and the potential for a WFKD to positively impact both metabolic and mental health, the primary purpose of this study was to test the feasibility of implementing a WFKD as an adjunct therapy in college students with MDD undergoing counseling and/or medication treatment. Our hypothesis was that the WFKD would be feasible in most participants and associated with improvement in measures of depression and metabolic health. Rigorous diagnostic assessments of mental health and dietary adherence were implemented. Additional assessments of body composition, neurocognitive function, and blood hormonal and inflammatory markers were conducted to inform future investigations.
This study was approved by the Institutional Review Board of Ohio State University -- Department of Human Sciences (IRB number 2022H0271), in accordance with the latest version of the Declaration of Helsinki (2013). All participants were informed of the experimental procedures, potential risks and the purpose of the study prior to enrolment and before obtaining their written and informed consent. No identifiable images of participants are included in this publication.
This was a single arm prospective study that involved a WFKD intervention that lasted 10-12 weeks depending on participant availability. Changes in variables were analyzed pre- and post-intervention. In addition, PHQ-9 and WHO-5 surveys were also assessed bi-weekly, and the HRSD, body composition, and blood markers were assessed at week 6 (i.e., mid-point). For lab-based visits, subjects arrived in the morning after an overnight fast followed by assessments for hydration, height, body mass, body composition, surveys, resting blood pressure, cognitive performance, and a blood draw.
Students enrolled at The Ohio State University (OSU) were recruited between Jan 2023 and April 2024. Students with depression who were currently engaged in counseling treatment provided through campus and community psychiatric treatment providers were made aware of the study during their therapy sessions. Interested individuals completed a pre-screening checklist by the study clinician, which included a summary of the study and a series of inclusion/exclusion criteria to which they answered 'yes' or 'no.' Individuals meeting screening criteria were then scheduled to meet with study staff to discuss the study in greater detail and sign an informed consent document approved by the IRB before participating in the study.
A total of 36 students met preliminary inclusion criteria and completed a 2-3 h baseline SCID-5 interview by a Master's level Psychology doctoral student to further determine eligibility. Participants had to be: (1) students registered at OSU (age 18-30 yr at the time of enrollment) with confirmed MDD as determined by a clinical interview and SCID-5 diagnosis conducted by a member of the study team, (2) currently engaged in counseling and/or medication treatment for depression and available for at least a 10-wk period, and (3) willing to eat a WFKD eating pattern as prescribed. Exclusion criteria included: (1) disordered eating (anorexia nervosa, bulimia nervosa, binge eating disorder, other specified/ unspecified eating disorder, avoidant restrictive eating disorder), (2) a substantial imminent risk of suicide, (3) body mass index (BMI) < 20 kg/m, (4) habitual consumption of a low-carbohydrate diet in the last 6 months, (5) gastrointestinal disorders or allergies that would prevent adherence to prescribed diets, (6) alcohol consumption more than 3 drinks daily or 14 drinks weekly, (7) diagnosed diabetes, liver, kidney, or other metabolic or endocrine dysfunction, or use of diabetic medications other than metformin, (8) an inability to access or prepare appropriate KD foods/meals, (9) pregnant, lactating, or planning on becoming pregnant during the study, or (10) unwilling to perform finger-stick blood testing.
Of the 36 participants who completed the SCID-5, 12 were excluded, and 24 completed baseline testing. A total of 16 participants completed the entire study. A CONSORT diagram is shown in Fig. 1. Baseline characteristics of the 16 participants who completed the study are shown in Table 1. Seven participants were prescribed medications including norepinephrine dopamine reuptake inhibitors (NDRI; n = 2), serotonin aganoist and reuptake inhibitor (SARI; n = 1), selective serotonin reuptake inhibitors (SSRI; n = 1), atypical antipsychotics (n = 1), tiazine anticonvusant (n = 1), and a combination of drug classes previously mentioned (n = 1).
Prior to baseline testing (BL), a dietary education session was provided to ensure participants understood the guiding principles to a WFKD and were comfortable with adhering to the diet for 10-12 weeks. The diet intervention began after the baseline testing visit. Participants received extensive personalized education and ongoing support from the dietetic team to achieve the nutritional goals of a WFKD. The intensity of coaching varied across participants depending on their baseline knowledge and individual situation. We utilized a HIPAA-compliant messaging app (Healthie, New York, NY, USA) that allowed the participant to communicate with dietetic staff throughout the intervention. We provided staple ketogenic-friendly food items to partially offset participant food costs, ensure nutrient quality, and facilitate adherence. Ten pre-packaged ketogenic meals (Factor 75, Aurora, IL, USA) were provided to each participant at baseline to provide a visual guide of how a balanced WFKD meal should be composed for the duration of the study when participants implemented the diet thereafter at home. Subsequently, participants chose food and meals on their own with the exception that we provided a few shelf-stable items such as high-quality fats (e.g., olive oil), salad dressings, salmon & sardine packets, beef jerky, Whisps cheese crips, nuts and seeds, and an oatmeal alternative.
The WFKD followed general principles with the aim to achieve blood R-BHB > 0.5 mM, which required most participants to consume < 50 g/day of carbohydrate and ~1.5 g/kg reference weight protein [41]. Fat comprised the remaining calories with an emphasis on monounsaturated and saturated sources from whole foods. A wide range of foods were encouraged including non-starchy vegetables, low glycemic fruits (berries, olives, tomatoes, lemons/limes), meats (beef, chicken, pork, fish, lamb), nuts and seeds, oils (olive, avocado, coconut), cheese, butter, cream, eggs, and fatty fish (salmon, sardines). Since a KD is associated with an enhanced natriuretic effect that may often lead to sodium and fluid losses, colloquially referred to as 'keto-flu'[42, 43], participants were provided broth/LMNT electrolyte packs (Naples, FL, USA). Consumption of magnesium and calcium rich foods were encouraged if symptoms of muscle cramping were reported.
The PHQ-9 was used to assess self-reported depressive symptoms [44] and the WHO-5 to assess global mental health data [45] using an internet-based survey and questionnaire administration program (Qualtrics, Seattle, WA, USA). The Hamilton Rating Scale of Depression (HRSD) [46] was administered by a qualified member of the study team either in-person or virtually.
Body mass was measured with an electronically-calibrated scale to the nearest ± 0.1 kg (SECA 305, Hamburg, Germany) and body composition was estimated from a single whole-body scan using dual-energy x-ray absorptiometry (iDXA, Lunar Corporation, Madison, WI, USA).
Each day of the intervention, participants recorded fasting capillary R-BHB and glucose from a fingerstick using reagent strips and a monitoring device (KetoMojo, Napa, CA, USA) in the morning before eating their first meal of the day. At study visits, participants rested in a seated position for 10 min before a phlebotomist obtained blood from a forearm vein into serum tubes. One tube was sent to a certified laboratory (Quest Diagnostics) to perform a standard chemistry panel. The remaining serum tubes were centrifuged, aliquoted into cryovial tubes, and stored at -80 °C until subsequent analyses. Fasted adipokines (BDNF, IL-1β, IL-6, IL-8, IL-10, Insulin, Leptin, MCP-1, TNF-α, β-NGF) were assessed using Meso Scale Discovery's (MSD) U-Plex Human adipokine panel (Meso Scale Discovery, Rockville, MD). Samples were run in duplicates and analyzed using MSD's Workbench 4.0 software. Calculations were run using a 4-parameter logistic with 1/y2 weighting, and intra-assay CVs for the standard and samples was 3.6%. A measure of insulin resistance (HOMA-IR) was determined from fasting serum glucose and insulin [47].
Cognitive function was assessed using the NIH Toolbox (NIH-TB) Cognition Battery (version 2) [48] on an iPad. The following tests were administered:
Fifteen words were presented orally over three consecutive learning trials. After each trial, participants were asked to recall as many words as possible. The outcome variable was the sum of the number of words recalled across all trials.
Pictures of activities were presented in a specified order in different locations on the iPad screen with an audio recording that simultaneously described each picture. Participants were asked to place the pictures in the correct location on the screen. The outcome variable was the number of pictures that were correctly placed.
Participants were asked to orally indicate the number that was paired with an abstract symbol based on a coding key. The outcome variable was the number of correct answers provided in 120 s.
Participants were asked to identity whether two pictures were the same ("yes" button) or not the same ("no" button) as quickly as possible. The outcome variable was the number of correct items (out of 130) completed in 90 s.
In this task, a target visual stimulus must be matched to one of two choice stimuli according to either shape or color. The outcome variable were percent correct and response time (seconds).
Participants were instructed to choose the direction of the central stimulus while inhibiting stimuli on the right and left of the central stimulus. On congruent trials, all stimuli faced the same direction. On incongruent trials, the central stimulus faced the opposite direction. Response time (seconds) for congruent and incongruent trials were obtained.
A series of stimuli were presented visually and orally one at a time. In the first condition, participants were instructed to sequence the stimuli in size order, from smallest to biggest. In the second condition, participants were asked to sequence the stimuli from one category (food) in size order and then stimuli from the second category (animals) in size order. The outcome measure was the total number of items recalled in the correct order.
Participant characteristics and outcome variables were summarized for completers (n = 16). The normality of continuous variables was assessed using the Shapiro-Wilk test. For variables that demonstrated normal distribution at baseline, we used a 1 (condition) x 3 (time) repeated-measures mixed-effects analysis of variance (RM ANOVA). If a significant main effect was detected, Bonferroni post-hoc corrections automatically adjusted for multiplicity were used to analyze pairwise comparisons. For variables that did not meet normality assumptions at baseline, the non-parametric Friedman test with Dunn's correction for multiple comparisons was employed. Changes in depression scores in completers were evaluated using an analysis of covariance (ANCOVA), with the baseline score included as a covariate to control for baseline variability between participants. An intent to treat (ITT) analysis was conducted on participants who completed baseline testing and at least one additional PHQ-9 assessment using a Last Observation Carried Forward to minimize bias by including all participant regardless of adherence and to provide a more generalizable estimate of the interventions effect on depression scores. To explore associations between depression scores and weight loss and ketones, we performed simple regression using Pearson correlation coefficients. The significance level for all tests was set at α ≤ 0.05.