Case 1 is a 39-year-old single man with bipolar disorder II. He had a history of prediabetes, which was diagnosed in 2018, and presented with obesity (BMI 33.5 kg/m²). Since 2013 he had been treated with lithium (450 mg 1–0-2) and bupropion (300 mg 1–0-0). Since 2018 he has also taken metformin (1,000 mg 1–0-1). His laboratory assessment revealed a mixed dyslipidemic profile, with TG of 399 mg/dl, LDL-C of 146 mg/dl, apoB of 1.46 g/L, total cholesterol of 226 mg/dl, and HDL of 34 mg/dl. Being prediabetic, his HbA1c was elevated at 46 mmol/mol (6.4%). At the initiation of KMT 2:1, Case 1 intuitively preferred to adopt intermittent fasting and to consume only two meals per day (lunch and dinner). Describing himself as an emotional eater who frequently overeats, he was surprised to feel a remarkable decline in hunger and cravings. Within 1 week of keto-adaptation, he easily reached therapeutic nutritional ketosis with consistent serum BHB levels of between 1 and 4 mmol/L.

After the 3-month KD intervention, we found significant improvements in markers of obesity, prediabetes and inflammation, and mixed improvements in dyslipidemia. The patient lost 8 kg of body weight, BMI decreased to 30.9 kg/m², body fat mass decreased from 39.0 kg to 34.3 kg and visceral fat percentage decreased from 10.6% to 8.8%. Moreover, HbA1c decreased to 5.1%, allowing the patient to discontinue metformin. The following dyslipidemic markers improved: TG decreased from 399 mg/dl to 202 mg/dl and HDL increased from 34 mg/dl to 42 mg/dl. However, total cholesterol increased to 272 mg/dl and LDL-C increased to 211 mg/dl. Despite this increase in LDL-C, apoB did not increase, staying quasi-constant from 1.46 g/L to 1.4 g/L. Accordingly, despite improvements in dyslipidemic markers, KMT did not reverse dyslipidemia, with LDL-C, TC and apoB still in the elevated range. Lp(a) slightly increased from 2.0 mg/dl to 3.1 mg/dl, but remained in a very low range. MDA-LDL, an important indicator of oxidative stress, decreased from 111 U/L to 70.4 U/L, now in the normal range. Inflammatory markers also improved with a significant reduction in hsCRP from 4.42 mg/L to 0.91 mg/L and a reduction in TNF-α from 10.6 to 9.5 pg/ml – albeit still in the elevated range. Nitrotyrosine remained in the standard range, despite an increase from <200 nmol/L to 572 nmol/L. Finally, laboratory assessments showed a concerning increase of homocysteine from 10.2 umol/L to 16 umol/L and a concerning increase of AGE from 36.1 ug/ml to 66.5 ug/ml.

The psychiatric outcomes of KMT were remarkable for Case 1. At the end of the intervention, he experienced profound improvements in emotional resilience, increased energy, better mental focus and stronger willpower. He went from a part-time job to working full-time. This was reflected in improvements in psychiatric assessments. His GAF, reflecting social, occupational and psychological functioning, improved from 60 to 100. CGI-BD-S, measuring symptom severity and treatment response in bipolar disorder, improved from 5 to 2 (1 = clinical remission, 2 = minimal symptoms). PHQ-9 and BDI-II, which are self-reported assessments measuring severity of depression, improved from 13 to 1 (PHQ-9) and from 20 to 3 (BDI-II), thereby achieving a “not depressed range.” Finally, HAM-D and MADRS, which are used by mental health professionals to measure severity of depression and treatment changes, also improved notably. HAM-D improved from 14 (mild depression) to 2 (clinically unremarkable) and MADRS improved from 22 (middle-grade depression) to 1, again ending up in the non-depressed range.

Case 2 is a 50-year-old married woman with rapid-cycling bipolar disorder II, which was diagnosed at age 30. Her medical history revealed sleep apnea, chronic pain syndrome, tinnitus and ADHD. She also had a history of irritable bowel syndrome, but she indicated that she was not troubled by this at present. Further complaints included itching, skin rashes, psoriasis, eczema, hives, acne and frequent headaches. She was treated with sertraline (50 mg 1–0-0) from 2023, lamotrigine (200 mg 1–0-1) from 2018 and quetiapine (50 mg 0–0–1-2) from 2023. At the initiation of KMT 1.5:1, Case 2 had difficulties reaching consistent serum BHB levels above 0.5 mmol/L. Upon shifting her KD ratio to 2:1, she still hardly reached 1 mmol/L. We surmise this may have been due to the impact of quetiapine, an antipsychotic that may lead to increased insulin secretion and plasma glucose levels, which both hinder ketogenesis. Throughout the 12-week-program, her blood BHB levels mostly fluctuated between 0.2 mmol/L and 1.0 mmol/L.

Metabolically, she was obese (BMI 34.7 kg/m²), with a body fat mass of 57.5 kg and visceral fat percentage of 11.2%. After 12 weeks of KMT, she had lost 14.5 kg of body weight, BMI decreased to 30.3 kg/m², body fat mass decreased to 46.7 kg and visceral fat was reduced to 8.7%. Labs revealed a dyslipidemic profile, which improved significantly after the intervention. TG dropped from 245 mg/dl to 133 mg/dl, LDL-C decreased from 218 mg/dl to 203 mg/dl, apoB decreased from 1.86 g/L to 1.35 g/L, and total cholesterol decreased from 317 mg/dl to 270 mg/dl. However, Lp(a) was somewhat increased, going from 34.3 mg/dl to 38.7 mg/dl. In turn, MDA-LDL was reduced significantly, from 155 U/L to 72.1 U/L. Inflammatory markers improved slightly, with hsCRP going from 4.83 mg/L to 3.78 mg/L and TNF-α staying quasi-unchanged with an evolution from 9.8 pg/ml to 10 pg/ml – albeit still in the elevated range. Nitrotyrosine decreased from 326 nmol/L to <200 nmol/L, both being in the standard range. Finally, laboratory assessments again showed a concerning increase of homocysteine from 11 umol/L to 13.1 umol/L and a concerning increase of AGE from 25.7 ug/ml to 43.7 ug/ml.

At the end of the intervention, Case 2 also saw remarkable improvements in all psychiatric assessments. GAF improved from 50 to 80. CGI-BD-S improved from 5 to 2 (1 = clinical remission, 2 = minimal symptoms). PHQ-9 improved from 18 (indicating moderately severe depression) to 3 (indicating no depression). BDI-II improved from 42 (indicating severe depression) to 7 (indicating no depression). HAM-D improved from 21 (indicating moderately severe depression) to 5 (clinically unremarkable) and MADRS improved from 26 (indicating moderately severe depression) to 7 (indicating mild depression).

Case 3 is a 46-year-old married man with bipolar disorder II. Due to unwanted side-effects such as lethargy and fatigue, he had discontinued all medications under the supervision of his treating psychiatrist, 1 year prior to the present intervention. Six months after discontinuing his medications, and 4 months prior to the present intervention, he had a first 2.5 month experience with KMT. Since the shift to KMT, he had felt more energized, more emotionally balanced and resilient, and less irritable. During this first 2.5 month KMT, he had lost 11 kg. He decided to pursue KMT in the present intervention to optimize his KD and in the hope KMT would allow him to stay off medications. Until the time of this writing, he has been able to do so successfully.

Upon initiating the present KMT (1.5:1) intervention, Case 3 quickly reached therapeutic nutritional ketosis with consistent serum BHB levels between 1.0 and 3.0 mmol/L. Metabolically, he started the intervention with a BMI of 25.0 kg/m², a fat mass of 29.3 kg and visceral fat percentage of 7.2%. Having already lost 11 kg in the previous 2.5-month KD phase in 2023, he lost less in this 3-month period, namely 3 kg. His BMI further decreased to 24.2 kg/m², body fat mass further decreased to 26.8 kg, and visceral fat percentage decreased to 6.9%. While initial labs revealed a dyslipidemic profile, all elevated markers improved after the intervention. TG decreased from 60.3 mg/dl to 50.2 mg/dl, LDL-C decreased from 178 mg/dl to 145 mg/dl, apoB decreased from 1.42 g/L to 1.03 g/L and total cholesterol decreased from 215 mg/dl to 183 mg/dl. HDL remained unchanged at 35 mg/dl. Of note was also a decrease in Lp(a), from 10.4 mg/dl to 8.1 mg/dl, and of MDA-LDL, from 77.4 U/L to 38.7 U/L. In turn, inflammatory markers also improved, with hsCRP decreasing from 2.27 mg/L to 1.01 mg/L and TNF-α decreasing from 7.7 pg/ml to 6.2 pg/ml. Baseline nitrotyrosine was strongly elevated at 3043 nmol/L and greatly decreased to 1463 nmol/L, albeit still in the elevated range. Finally, we found a decrease in homocysteine from 17 umol/L to 14 umol/L, albeit both in the elevated range, and a concerning increase of AGE from 28.7 ug/ml to 45.7 ug/ml, albeit both in the standard range.

After the intervention, Case 3 showed improvements across all psychiatric assessments. His GAF improved from 95 to 100. CGI-BD-S improved from 4 to 2. PHQ-9 and BDI-II improved from 6 to 2 (PHQ-9) and from 6 to 2 (BDI-II). HAM-D improved from 5 to 3 and MADRS improved from 7 to 1. Case 3 continues to follow KMT to this day, considering it a life-changing intervention, his friends and family attesting to its favorable therapeutic effects.

Case 4 is a 47-year-old woman with bipolar disorder II, which was diagnosed in 2023, at 46 years of age. She has been treated with lamotrogine (50 mg 1–0-1) since her diagnosis in 2023 and escitalopram (15 mg) since having had long covid in February 2022. At the initiation of KMT 1.5:1, Case 4 quickly reached consistent serum BHB levels of 1.0 mmol/L. Throughout the 12-week-program, her blood BHB levels mostly fluctuated between 0.5 and 1.0 mmol/L.

Metabolically, she began the intervention with a BMI of 22.8 kg/m², a body fat mass of 18.6 kg and visceral fat percentage of 4.5%. After the intervention, her weight decreased to 54.5 kg, BMI decreased to 21.8 kg/m², her fat mass decreased to 17.5 kg, and her visceral fat percentage decreased to 4.1%. Labs revealed a dyslipidemic profile with elevated LDL-C of 123 mg/dl and total cholesterol of 240 mg/dl. After the intervention, both markers further increased to 141 mg/dl (LDL-C) and 240 mg/dl (total cholesterol). However, her baseline apoB of 0.92 g/L remained quasi unchanged at 0.91 g/L, again marking discordance between LDL-C and apoB. HDL increased from 105 mg/dl to 113 mg/dl, and triglycerides slightly increased from 63.1 mg/dl to 66.6 mg/dl. Also, Lp(a) decreased from an elevated 92.8 mg/dl to 56.8 mg/dl, albeit still in the elevated range. MDA-LDL and nitrotyrosine, both markers of oxidative stress, decreased significantly from 73.3 U/L to 42.8 U/L (MDA-LDL) and from 926 nmol/L to 432 nmol/L (nitrotyrosine). Inflammatory markers also improved after the intervention: baseline hsCRP of <0.3 mg/L remained unchanged, while TNF-α decreased from 5.6 pg/ml to 4.3 pg/ml. Finally, both homocysteine and AGE again showed concerning changes with homocysteine increasing from 6.1 umol/L to 12.4 umol/L and AGE increasing from 28.5 ug/ml to 64.1 ug/ml.

After the intervention, Case 4 showed improvements across all psychiatric assessments. Her GAF improved from 60 to 100. CGI-BD-S improved from 3 to 2. PHQ-9 and BDI-II improved from 9 to 0 (PHQ-9) and from 13 to 1 (BDI-II). HAM-D improved from 9 to 0 and MADRS improved from 8 to 0.

Case 5 is a 34-year-old married woman with bipolar disorder I, which was diagnosed at 32 years of age, and comorbid anxiety disorder, which was diagnosed at 19 years of age. Since 2022, she has been treated with lithium (450 mg 1–0-1.5) and quetiapine (150 mg 0–0-1). Metabolically, she had a BMI of 26.6 kg/m², with a body fat mass of 30 kg and visceral fat percentage of 4.1%. After the intervention, she lost 2.9 kg, her BMI decreased to 25.7 kg/m², body fat mass decreased to 27.9 kg and visceral fat decreased to 3.5%. She too presented with dyslipidemia at the outset, but her elevated markers evolved heterogeneously. While total cholesterol increased from 219 mg/dl to 234 mg/dl, HDL slightly decreased from 64 mg/dl to 63 mg/dl, and LDL-C increased from 151 mg/dl to 179 mg/dl, her apoB remained unchanged at 1.01 g/L, again marking discordance between LDL-C and apoB. Favorable changes included a decrease in Lp(a) from 7.3 mg/dl to 4.9 mg/dl, and a decrease in TG from 51.1 mg/dl to 33.8 mg/dl. Both markers of oxidative stress also improved with MDA-LDL decreasing from 67.2 U/L to 54.9 U/L and nitrotyrosine decreasing from 438 nmol/L to 315 nmol/L. Inflammatory markers also improved with hsCRP remaining unchanged at <0.3 mg/L and TNF-α decreasing from 13.1 pg/ml to 8.9 pg/ml. Finally, both homocysteine and AGE again increased unfavorably with homocysteine rising from 6.6 umol/L to 11.3 umol/L and AGE increasing from 18.6 ug/ml to 53.5 ug/ml.

After the intervention, Case 5 also showed improvements across all psychiatric assessments. Her GAF improved from 60 to 100. CGI-BD-S improved from 4 to 1. PHQ-9 and BDI-II improved from 3 to 2 (PHQ-9) and from 8 to 3 (BDI-II). HAM-D improved from 7 to 0 and MADRS improved from 9 (mild depression) to 1 (no depression).

Case 6 is a 29-year-old man with bipolar disorder I, which was diagnosed in 2015, at 20 years of age. Medical history was notable for a severe case of Lyme disease immediately preceding the onset of his first depressive episode at the age of 18, with the formal diagnosis of bipolar disorder following 3 years later. He was treated with valproate (1,000 mg 1–0-1) and sertraline (200 mg 1–0-0). At the initiation of KMT 1.5:1, Case 6 quickly reached consistent serum BHB levels between 1.0–2.0 mmol/L. Throughout the 12-week-program, his blood BHB levels mostly fluctuated between 1.0 and 3.0 mmol/L.

Metabolically, he had a BMI of 26.5 kg/m², a body fat mass of 25.2 kg and visceral fat percentage of 4%. After 12 weeks of KMT, he had lost 1.9 kg of bodyweight, BMI decreased to 25.9 kg/m², body fat mass decreased to 19.2 kg and visceral fat decreased to 3.5%. Labs revealed a dyslipidemic profile, with elevated levels of total cholesterol (237 mg/dl) and LDL-C (178 mg/dl). After the intervention, both dyslipidemic markers increased, with total cholesterol reaching 242 mg/dl and LDL-C reaching 192 mg/dl. However, apoB decreased from 1.33 g/L to 1.17 g/L, again showing remarkable discordance between LDL-C and apoB. Other favorable lipid changes include an increase of HDL from 52 mg/dl to 54 mg/dl, a lowering of TG from 70 mg/dl to 59.1 mg/dl and a decrease in Lp(a) from 24.3 mg/dl to 16 mg/dl. MDA-LDL also improved, decreasing from 108 U/L to 79 U/L. On the other hand, nitrotyrosine increased from 425 nmol/L to 799 nmol/L, now in the elevated range. Inflammatory markers evolved positively with hsCRP lowering from 1.2 mg/L to 0.3 mg/L and TNF-α slightly decreasing from 8.7 pg/ml to 8.4 pg/ml, albeit still in the elevated range. Baseline homocysteine of 14.7 umol/L remained elevated at 14.7 umol/L, and baseline AGE again increased unfavorably from 27.6 ug/ml to 56.3 ug/ml.

After the intervention, Case 6 saw mixed changes in psychiatric assessments. While some assessments improved, others remained unchanged and some slightly worsened. For this reason, we do not consider Case 6 an unequivocal respondent to this intervention. While GAF improved from 80 to 100, CGI-BD-S slightly worsened from 3 to 4. PHQ-9 slightly worsened from 1 to 2, but this is still in the “healthy” range. BDI-II improved from 9 (minimal depression) to 5 (no depression). HAM-D slightly worsened from 2 to 3, both being in the “clinically unremarkable” range, while MADRS slightly improved from 3 to 2, both indicating “no depression.”

Case 7 is a 50-year-old married man with bipolar disorder I, which was diagnosed in 2004, at 30 years of age. His medical history was notable for ankylosing spondylitis (Morbus Bechterew) and recurrent migraines. He was treated with quetiapine (275 mg 0–0-1) and valproate (200 mg 0–0-1). At the beginning of our KMT intervention, Case 7 chose to adopt a carnivore KD, with virtually no carbohydrate intake. We agreed to this, on the condition his serum BHB levels would consistently reach between 1.0 mmol/L and 5.0 mmol/L. Throughout the intervention, this was consistently the case.

Metabolically, he had a BMI of 22.3 kg/m², a body fat mass of 16.2 kg and visceral fat percentage of 5.6%. After 12 weeks of KMT, he had gained 4.4 kg (3.8 kg of which was muscle mass), BMI increased to 23.6 kg/m², body fat mass increased to 17.5 kg and visceral fat percentage also increased to 6.9%. Labs revealed a dyslipidemic profile, with baseline elevated markers including a total cholesterol of 224 mg/dl and LDL-C of 162 mg/dl. At the end of the intervention, total cholesterol increased to 258 mg/dl, and LDL-C increased to 185 mg/dl. Further concerning changes included an increase of baseline TG from 51.1 mg/dl to 111 mg/dl and an increase in baseline apoB from 1.18 g/L to 1.35 g/L. Conversely, baseline HDL increased from 54 mg/dl to 61 mg/dl and Lp(a) decreased from 46.5 mg/dl to 44 mg/dl. MDA-LDL also decreased from 84.3 U/L to 52.4 U/L, while baseline nitrotyrosine of <200 nmol/L remained unchanged at <200 nmol/L. Effects on inflammatory markers were also mixed, with hsCRP slightly increasing from 1.14 mg/L to 1.3 mg/L and TNF-α slightly decreasing from 70.9 pg/ml to 69.9 pg/ml, being still markedly elevated. Finally, homocysteine decreased from 10.9 umol/L to 8.6 umol/L and AGE increased from 38.6 ug/ml to 89.9 ug/ml.

Psychiatric assessments indicated that Case 7 was a non-respondent to KMT. While some assessments either improved minimally and some assessments remained unchanged, others worsened. Initial GAF score of 60 improved to 70, but initial CGI-BD-S score remained unchanged (4). The PHQ score slightly worsened from 8 to 9, but this remained within the “clinically unremarkable” range. BDI-II slightly improved from 8 to 7, both indicating “no depression.” HAM-D remained unchanged, both initial and final assessments scoring 10, indicating “mild depression.” Finally, the MADRS assessment gave a worse result at the end of the intervention, going from 6 (“no depression”) to 12 (“mild depression”).

Following the KD intervention, mean (SD) effects on lipid metabolism showed largely favorable changes. While most of these were predictable, our findings also showed an unexpected discordance between LDL-C and apoB (Figure 2).

First of all, in line with the literature to date, one would expect KMT to have the following positive effects on lipid metabolism: a reduction in TG, an increase in HDL-C, and possibly a reduction in Lp(a) (47, 55, 56). Our mean results were in line with these expectations. Mean TG was lowered by 40.56 mg/dl, which is a 30% reduction over the span of 3 months. Mean HDL-C was slightly increased by 3 mg/dl, which is a 5% increase. Of note, mean Lp(a) was lowered by 6.57 mg/dl, which is a 21% reduction.

Second, based on the current literature, we know KMT may have a variable effect on LDL-C and total cholesterol, potentially improving or worsening hyperlipidemia (47, 48, 55). If an animal-based KD is adopted, one may reasonably expect an increase in both markers. This was also the case in our findings, with both LDL-C and TC increased in 5 out of 7 patients. We found a mean increase of LDL-C of 14.29 mg/dl, which is a 8.6% increase, and a mean increase of total cholesterol of 7.14 mg/dl, which is a 3% increase. It should be noted that all participants adopted a KD high in saturated fat, against our dietary recommendation to consume a Mediterranean ketogenic diet, consisting of 50% MUFA, 25% PUFA and 25% SFA.

Third, based on the current literature, one would expect to find a high correlation between changes in LDL-C and apoB. As the 2019 European Atherosclerosis Society/European Society of Cardiology Guidelines state, it is generally estimated that only in 20% of patients there may be discordance between measured LDL-C and apoB levels (57). This high correlation between LDL-C and apoB is one of the reasons why apoB is often not measured in routine clinical care. When recommended by guidelines and consensus groups, measurement of apoB has been prioritized or restricted to patients with hypertriglyceridaemia, the presumption being cholesterol-depleted apoB particles are only common in such individuals.

In contrast to the high correlation between apoB and LDL-C, our case series highlights a remarkable discordance between changes in LDL-C and apoB. In fact, while mean LDL-C increased by 14 mg/dl (a 8.6% increase), mean apoB was decreased by 14 mg/dl (0.14 g/L, as indicated in the table above), which is a reduction of 10.46%. Indeed, whereas 5 out of 7 patients saw their LDL-C increase, only 1 patient saw his apoB increase, 2 patients saw their apoB decrease, and 4 saw their apoB unchanged. Of the 5 patients who had increased LDL-C, 3 saw their apoB remain unchanged, 1 saw his apoB decrease and only 1 saw his apoB increase.

The observed increase in LDL-C without a change in apoB, or with a decrease in apoB, may indicate a shift in the composition of LDL particles, from small-dense LDL (sdLDL) particles towards large, buoyant (lbLDL) LDL particles. Given such a shift, LDL particles would contain more cholesterol per particle (hence the increase in LDL-C) but the total number of LDL particles (and hence apoB) would not have increased, or even have decreased. This suggests that in at least 4 patients, our intervention led to a shift in LDL particle size, which would increase LDL concentration but not apoB count. Whether or not such a shift reduces overall cardiovascular risk, has been a heated topic of debate (58–65). As it stands, the markers of apoB and LDL particle number are considered stronger predictors of cardiovascular risk than LDL-C. Thus, the observed mean increase in LDL-C should be seen in light of the observed mean decrease in apoB. Potential discordance between LDL-C and apoB underlines the importance of measuring both markers to accurately evaluate individual cardiometabolic effects of KMT.

Finally, it is noteworthy that triglycerides did not add to the predictive value of changes in apoB. Contrary to current expectations, discordance between LDL-C and apoB did not correlate with elevated TG. Of the 4 patients who saw discordance between both markers, only one had elevated TG, whereas all others had TG below 66.6 mg/dl.

Since mitochondria and metabolism are fundamental regulators of immune function, it should come as no surprise that metabolic interventions can have a profound impact on immunity and inflammation (66). KDs in particular have consistently been shown to lower markers of inflammation such as C-reactive protein (CRP), tumor necrosis factor-alpha (TNF-α) and interleukin (IL)-6 (67). After our intervention, mean hsCRP was reduced by 0.94 mg/L, which is a reduction of 45%, and mean TNF-α was reduced by 1.31 pg/ml, which is a reduction of 7% (Figure 3).

Chronic oxidative stress is found in most chronic diseases, ranging from heart disease and type 2 diabetes to schizophrenia and bipolar disorder (68). Our case series looked at two key biomarkers for systemic oxidative stress, namely malondialdehyde-modified LDL (MDA-LDL) and nitrotyrosine. Malondialdehyde is a result of oxidative fatty acid degeneration. It has cytotoxic effects and attacks LDL particles, with MDA-LDL as a result. While LDL-C is an important risk factor for atherosclerosis, its oxidized forms, such as MDA-LDL, are especially atherogenic. Some studies have even suggested MDA-LDL is a superior marker of ASCVD compared to other commonly used lipid markers (69). We found a mean reduction of MDA-LDL of 36.77 U/L, which is decrease of 38% (Figure 4).

Persistent oxidative stress can lead to excessive formation of reactive nitrogen species (RNS), also referred to as nitrosative stress (NSS). RNS can lead to cell damage through the generation of nitrotyrosine, causing toxic effects on cells, much like oxidative stress. Nitrosative stress is a highly oxidizing, pathological condition that is found to participate in many cardiometabolic diseases such as atherosclerosis, diabetes, hypertension and endothelial dysfunction (70). After our KMT intervention, we have found a mean reduction in serum nitrotyrosine levels of 225 nmoL/L, a reduction of 28%.

As noted in the case presentations above, laboratory assessments showed concerning increases in homocysteine and advanced glycation end products (AGE) throughout our cases. Given the impact of both markers on cardiovascular risk, these findings are remarkable. Mean baseline homocysteine was already elevated, with 10.9 umol/L. After the intervention, mean homocysteine increased to 12.9 umol/L, an 18% increase. This phenomenon may be attributed to an increased dietary intake of methionine without a corresponding and sufficient supply of essential cofactors such as vitamin B6, vitamin B9, vitamin B2, choline, zinc and/or magnesium, which are required for proper function of the methylation cycle. These co-factors can be lacking in a non-supplemented ketogenic diet without organ meats. Mean baseline AGE was in the standard range, with 29 ug/ml. However, after the intervention mean AGE was increased to 60 ug/ml, a 106% increase. This may have been caused by maillard reactions of fried, roasted and grilled meats, a depleted glyoxalase system caused by a relative deficiency of vitamin B1 (often lacking in organ-free ketogenic diets), general anti-oxidant deficiency, or physiological insulin resistance often seen in ketogenic diets, potentially leading to higher blood glucose and AGE formation.

Mean weight reduction was 4 kg (−4.6%), mean visceral fat reduction was 0.69% (−10%) and mean fat mass reduction was 3.7 kg (−12%).