HoldsUp

Mark Hyman, MD · 2025-09-24 · Mark Hyman (host), Shebani Sethi

How To Use Metabolic Psychiatry To Heal Your Anxiety & Depression | Stanford Psychiatrist

31 research-tied claims examined: 1 contradicted 2 overstated 2 context 21 supported 5 unverified

21

Supported by research

0:00:00Shebani Sethisupportedmoderate

One in three people in the United States have insulin resistance.

"One in three people have insulin resistance in the United States" (said at 0:00:00)

Nationally representative surveillance data from the National Health and Nutrition Examination Survey (NHANES) support the claim that approximately one in three people in the United States have insulin resistance. Among non-diabetic US adults, the age-standardized prevalence of insulin resistance (measured by HOMA-IR) was estimated to rise from 24.8% in 1999–2000 to 38.4% in 2017–2018, averaging around one in three adults (~33%) across recent survey cycles.

0:03:24Shebani Sethisupportedmoderate

Levels of lactate were observed to be elevated in serious mental illness.

"100 years ago in psychiatry we had seen that there were levels of lactate that were elevated in serious mental illness" (said at 0:03:24)

Elevated lactate levels (particularly in the brain and cerebrospinal fluid) have been consistently documented in serious mental illnesses such as bipolar disorder and schizophrenia, reflecting abnormalities in mitochondrial function and brain energy metabolism. Systematic reviews of magnetic resonance spectroscopy (MRS) and CSF studies show increased cerebral lactate concentrations in bipolar disorder across multiple cohorts, with preliminary MRS evidence indicating similar elevations in schizophrenia.

0:03:42Shebani Sethisupportedmoderate

Glutathione levels are low in serious mental illness.

"and that there were levels of glutathione which were low—they were low and it is an antioxidant." (said at 0:03:42)

Glutathione is a primary cellular antioxidant, and meta-analyses confirm that peripheral and brain glutathione levels are reduced in patients with schizophrenia, a major serious mental illness. A 2019 systematic review and meta-analysis found lower peripheral glutathione and total glutathione levels in patients with schizophrenia compared to healthy controls. Magnetic resonance spectroscopy (MRS) studies also demonstrate small but statistically significant reductions in anterior cingulate cortex and brain glutathione in patients with established schizophrenia. However, findings can vary by specific diagnosis and region; for instance, anterior cingulate cortex glutathione levels have been reported as elevated in bipolar disorder.

0:13:41Shebani Sethisupportedhigh

Approximately 37% to 40% of people with bipolar illness have metabolic syndrome.

"So in bipolar illness about 37, almost 40% have metabolic syndrome." (said at 0:13:41)

The speaker's estimate is directly supported by published meta-analyses of individuals with bipolar disorder. A landmark 2013 meta-analysis published in The American Journal of Psychiatry evaluated 37 publications (N = 6,983) and found an overall metabolic syndrome prevalence of 37.3% (95% CI: 36.1%–39.0%) across standardized criteria. Subsequent meta-analyses have similarly reported pooled global prevalence rates in the range of ~33% to 37%, depending on geographic region, treatment status, and diagnostic criteria used.

0:14:06Mark Hyman (host)supportedmoderate

Research from Tufts shows that 93.2% of Americans have metabolic dysfunction based on criteria including blood sugar, blood pressure, cholesterol, overweight/obesity, or cardiovascular history.

"I think if you look at some of the work out of Tufts, they looked at people with what they sort of determined was metabolic dysfunction, which is either you have a high blood sugar, high blood pressure, abnormal cholesterol, you're overweight or obese, or you've had a heart attack or stroke, and if you combine all those which are all related to the fundamental biology of insulin resistance, which I want to unpack with you, that's 93.2% of Americans." (said at 0:14:06)

A 2022 study led by researchers at Tufts University (O'Hearn et al., published in the Journal of the American College of Cardiology) analyzed NHANES data from 1999 to 2018 covering 55,081 U.S. adults. The researchers evaluated cardiometabolic health across five components: adiposity (BMI and waist circumference), blood glucose, blood lipids (cholesterol/triglycerides), blood pressure, and clinical cardiovascular disease. In 2017–2018, only 6.8% of U.S. adults met the criteria for optimal cardiometabolic health across all five domains, meaning 93.2% of American adults had suboptimal cardiometabolic health (intermediate or poor levels in at least one component).

0:17:02Shebani Sethisupportedvery low

Lithium improves insulin signaling in the brain.

"And I'll give you an example is lithium. It also improves insulin signaling in the brain." (said at 0:17:02)

Preclinical and mechanistic literature supports the claim that lithium acts on and enhances downstream insulin signaling pathways in the brain. Lithium inhibits glycogen synthase kinase-3beta (GSK-3β)—a major downstream node in the canonical PI3K/Akt insulin signaling cascade—and has been shown in rodent models to restore central insulin-related kinase activation (including Akt, mTOR, and phospho-GSK3β). However, evidence is currently limited to preclinical animal models, in vitro experiments, and mechanistic reviews, with direct clinical trial evidence in human brains remaining preliminary.

0:17:25Shebani Sethisupportedlow

Peripheral insulin resistance leads to degeneration and atrophy of hippocampal neurons.

"When you have insulin resistance peripherally—so outside the brain—it leads to degeneration and atrophy of some of the hippocampal neurons as well." (said at 0:17:25)

Preclinical and clinical neuroimaging studies support the link between insulin resistance and hippocampal neurodegeneration. In human observational studies, measures of peripheral insulin resistance (such as HOMA-IR) are inversely correlated with total and regional hippocampal volume and grey matter integrity on MRI. In animal models, experimental insulin resistance directly induces dendritic atrophy and structural deficits in hippocampal CA3 pyramidal and dentate gyrus neurons. Because human evidence relies primarily on observational MRI proxies and causal cellular mechanisms are established mainly in animal models, the certainty is graded as low.

0:24:41Shebani Sethisupportedmoderate

A study showed that metformin adjunctively improved outcomes in patients with treatment-resistant bipolar depression.

"And so there was a study a colleague of mine did looking at metformin and treatment-resistant bipolar depression and that showed an improvement in those who were treated with metformin and also had psychiatric medication on board." (said at 0:24:41)

In a randomized, quadruple-masked, placebo-controlled clinical trial by Calkin et al. (the TRIO-BD study, n=45), adjunctive metformin was evaluated in patients with treatment-resistant bipolar depression and comorbid insulin resistance. The trial found that patients who reversed their insulin resistance on metformin (10 of 20 on metformin vs 1 of 25 on placebo) experienced statistically significant and sustained improvements in Montgomery-Åsberg Depression Rating Scale (MADRS) scores, functioning (GAF), anxiety (HAM-A), and global clinical impressions (CGI-BP) compared to non-converters.

0:27:44Shebani Sethisupportedmoderate

Supplementation with at least 1 gram of omega-3 containing EPA per day for 8 to 12 weeks showed modest benefit for early-phase psychosis and schizophrenia, but did not have the same effect for chronic schizophrenia.

"Give an example of research that's been done in our field by colleagues who looked at omega-3 supplementation, at least one gram including EPA per day, for I think it was a total of 8 to 12 weeks period, which significantly had a good—it was modest evidence that showed benefit for psychosis, early-phase schizophrenia, early-phase psychosis, and it's also been an adjunctive treatment for depression. Things like this are helpful for us to know about. It unfortunately didn't have the same effect for chronic schizophrenia" (said at 0:27:44)

The speaker accurately describes the findings of landmark trials and stage-specific meta-analyses. A prominent randomized controlled trial by Amminger et al. (2010) tested 1.2 g/day of omega-3 polyunsaturated fatty acids (containing 700 mg EPA) for 12 weeks in individuals at ultra-high risk for psychosis, finding a significant reduction in transition to psychotic disorder and symptom improvement. Subsequent meta-analyses (e.g., Chen et al., 2015; Fusar-Poli & Berger, 2012) confirmed stage-specific trends showing modest symptom reductions in prodromal and early-phase schizophrenia, whereas trials in chronic/established schizophrenia consistently failed to demonstrate symptomatic benefit.

0:24:38Shebani Sethisupportedhigh

Pro-inflammatory cytokines affect tryptophan metabolism and serotonin synthesis.

"And we know cytokines also are pro-inflammatory, right? And they end up also affecting serotonin synthesis. It affects tryptophan metabolism." (said at 0:24:38)

Pro-inflammatory cytokines (such as interferons, TNF-α, and interleukins) induce the enzyme indoleamine 2,3-dioxygenase (IDO). IDO catabolizes tryptophan along the kynurenine pathway, reducing the availability of tryptophan for the synthesis of serotonin (5-HT).

0:26:32Shebani Sethisupportedmoderate

Exercise has been shown to increase BDNF (brain-derived neurotrophic factor) levels and improve cognition.

"And, you know, increasing BDNF, which you and I talked about last time, I think on your show like now five years ago, the time has flown. But you described it as Miracle-Gro. And I really love the way you described it because it's exactly what it is. That was one with exercise that is shown to improve those levels and improve cognition and so forth." (said at 0:26:32)

A substantial body of randomized controlled trials and meta-analyses demonstrates that structured physical exercise (including aerobic, resistance, and combined training) significantly increases circulating brain-derived neurotrophic factor (BDNF) levels and improves various domains of cognitive performance across healthy individuals, older adults, and clinical populations.

0:31:34Shebani Sethisupportedmoderate

Depression severity and chronicity are significantly associated with the triglyceride-to-HDL ratio, which is directly related to insulin resistance.

"looking at triglyceride/HDL ratio, for example. That's been something that has been shown with a lot of good data that depression severity and chronicity is associated with that marker, and that's directly related to insulin resistance." (said at 0:31:34)

The triglyceride-to-HDL (TG/HDL) ratio is an established clinical surrogate marker of insulin resistance. Longitudinal observational studies, such as the Netherlands Study of Depression and Anxiety (NESDA), have confirmed that elevated TG/HDL ratio and other markers of insulin resistance significantly predict the development and course of major depressive disorder. Additional cohort and clinical studies have similarly observed relationships between lipid dysregulation and depression severity.

0:32:04Shebani Sethisupportedlow

In clinical research, insulin sensitivity was associated with depression severity but not chronicity, whereas the triglyceride-to-HDL ratio was associated with both severity and chronicity.

"What was interesting about that study is that insulin sensitivity wasn't associated with the chronicity of depression, but it was for the severity, but the triglyceride/HDL ratio was associated with both." (said at 0:32:04)

The speaker accurately describes the findings from a clinical cohort study (Watson et al., JAMA Psychiatry 2021) investigating metabolic endophenotypes in the Netherlands Study of Depression and Anxiety (NESDA). In that study, standard insulin resistance/sensitivity measures (such as HOMA-IR) were significantly associated with depression symptom severity but not depression chronicity, whereas dyslipidemic insulin resistance markers, specifically the triglyceride-to-HDL ratio, showed significant associations with both depression severity and chronicity/remission status.

0:32:35Shebani Sethisupportedlow

In patients with bipolar disorder, insulin resistance is associated with increased rapid cycling, greater treatment resistance, and higher suicidality.

"when we look at insulin resistance, we see that even with bipolar disorder, you have more rapid cycling, you have more treatment resistance, and you have more suicidality." (said at 0:32:35)

Published observational studies and systematic reviews indicate that insulin resistance (and broader impaired glucose metabolism) in bipolar disorder is significantly associated with an unfavorable illness course, including higher rates of rapid cycling (odds ratio ~2.9-3.1), poor response or resistance to mood stabilizers such as lithium (odds ratio ~6.7-8.4), and higher overall clinical morbidity. However, because the available evidence is largely derived from cross-sectional and observational cohort studies with relatively modest sample sizes, the GRADE certainty of the evidence is low.

0:36:07Shebani Sethisupportedhigh

The human brain constitutes approximately 2% of body weight but consumes 20% of the body's energy.

"the brain—it makes up 2% of our body, but it consumes 20% of our energy." (said at 0:36:07)

Established physiological literature confirms that the adult human brain comprises approximately 2% of total body weight while accounting for roughly 20% to 25% of the body's total basal energy (glucose and oxygen) consumption.

0:41:14Mark Hyman (host)supportedlow

In a 4-month pilot study of 23 patients with bipolar disorder and schizophrenia on a ketogenic diet, 100% experienced reversal of metabolic syndrome, average body weight decreased by 12%, visceral belly fat decreased by 36%, and insulin resistance decreased by 27%.

"in your clinical trial you did on the ketogenic diet for bipolar and schizophrenia that recently published with 23 patients, you had 100% reversal of metabolic syndrome or pre-diabetes, you had 12% reduction in body weight, you had a 36% reduction in belly fat, 27% reduction in insulin resistance" (said at 0:41:14)

A 4-month single-arm pilot trial by Sethi et al. (2024) enrolled 23 participants with bipolar disorder or schizophrenia and metabolic abnormalities on a ketogenic diet. By study completion, no participants met the criteria for metabolic syndrome (100% reversal). Among adherent participants, there was an average 12% reduction in body weight, a 36% reduction in visceral adipose tissue ('belly fat'), and a 27% reduction in homeostatic model assessment of insulin resistance (HOMA-IR). Because this was a small, open-label, single-arm pilot study without a control group, the GRADE certainty of the underlying evidence is low.

0:58:35Shebani Sethisupportedmoderate

Eating disorders involve disruptions across serotonin, dopamine, and opioid signaling pathways.

"when we look at eating disorders, whether it's anorexia or bulimia or binge eating, there are disruptions in several different pathways: there's serotonin, dopamine, and opioid pathways as well." (said at 0:58:35)

Extensive neurobiological and clinical research indicates that eating disorders (including anorexia nervosa, bulimia nervosa, and binge eating disorder) involve alterations and dysregulation across multiple neurotransmitter and neuromodulator systems, prominently including the serotonin (mood, satiety, impulse control), dopamine (reward processing, reinforcement), and endogenous opioid (hedonic evaluation, palatability) signaling pathways.

  • supports: Dysregulation of brain reward systems in eating disorders: neurochemical information from … (Neuropharmacology 2012) · cited 263x in the literature
    "The findings suggest that alterations in dopamine (DA), acetylcholine (ACh) and opioid systems in reward-related brain areas occur in response to binge eating of palatable foods. Moreover, animal models of bulimia nervosa suggest that while bingeing on palatable food releases DA, purging attenuates the release of ACh that might otherwise signal satiety. Animal models of anorexia nervosa suggest that restricted access to food enhances the reinforcing effects of DA when the animal does eat. The activity-based anorexia model suggests alterations in mesolimbic DA and serotonin occur as a result of restricted eating coupled with excessive wheel running." (abstract, results, passage verified)
    pubmedfull study (doi)
  • supports: Psychopharmacological advances in eating disorders. (Expert review of clinical pharmacology 2018) · cited 115x in the literature
    "On the basis of theoretical considerations about neurotransmitters, hormones and neural circuits, possible drug targets for the treatment of EDs may include signal molecules and receptors of the self-regulatory system such as serotonin, norepinephrine and glutamate, the hedonic system including opioids, cannabinoids and dopamine and the hypothalamic homeostatic system including histamine, ghrelin, leptin, insulin, and glucagon-like peptide-1." (abstract, results, passage verified)
    pubmedfull study (doi)
0:59:06Shebani Sethisupportedmoderate

A clinical trial evaluating Qsymia in patients with binge eating disorder and bulimia demonstrated a 63% abstinence rate in the treatment group compared to approximately 6% in the control group.

"I did a trial in the past with a colleague of mine, Deborah Safer, at Stanford, in which we looked at an FDA-approved obesity drug called Qsymia, and we looked at that in binge eating disorder and bulimia and achieved abstinence rates of 63% while on the drug versus the control group was about 6%." (said at 0:59:06)

A randomized, double-blind, placebo-controlled crossover trial led by Deborah Safer and colleagues at Stanford evaluated phentermine-topiramate extended-release (Qsymia) in 22 adults with binge-eating disorder (n=18) and bulimia nervosa (n=4). The trial found that binge abstinence rates reached 63.6% while receiving phentermine-topiramate ER compared to 9.1% while receiving placebo (p < 0.0001).

0:37:49Shebani Sethisupportedvery low

Conjugation of the ketone body beta-hydroxybutyrate with the amino acid phenylalanine mediates downstream appetite suppression and weight loss effects.

"downstream of ketone metabolism there may be, say, an amino acid called phenylalanine attached to beta-hydroxybutyrate, which is what ketones will break down into. And if you have that, the end effect of appetite reduction or weight loss is present." (said at 0:37:49)

Preclinical research demonstrates that beta-hydroxybutyrate (BHB) is enzymatically conjugated with free amino acids by the enzyme CNDP2, forming BHB-amino acids. The most abundant conjugate, BHB-phenylalanine (BHB-Phe), acts as a signaling metabolite that activates hypothalamic and brainstem neurons, suppressing food intake and reducing body weight in obese mice. Furthermore, CNDP2-knockout mice fail to produce BHB-Phe and display increased food intake and body weight despite ketosis. Subsequent human trials confirm that circulating BHB-Phe is induced during ketogenic diets and following exogenous ketone supplementation, although direct behavioral and weight-loss effects have only been causally tested in animal models (leading to a very low GRADE certainty rating).

1:08:44Shebani Sethisupportedmoderate

Acetyl-L-carnitine is low in patients with depression and increases or resolves in association with antidepressant treatment response.

"There was another biomarker, like L-acetylcarnitine, that my colleague actually at Stanford had looked at as a marker of depression. And it's low in depression, and so with antidepressants it resolved the marker, really with treatment response." (said at 1:08:44)

Clinical studies support the claim. Blood levels of acetyl-L-carnitine (ALC/LAC) are significantly lower in patients with major depressive disorder compared to healthy control subjects. Furthermore, patients who respond effectively to antidepressant treatment or achieve clinical remission demonstrate a significant increase/normalization in acetyl-L-carnitine levels, whereas treatment non-responders do not.

1:08:44Mark Hyman (host)supportedmoderate

The cost of whole genome sequencing has dropped from approximately one billion dollars historically to around $300.

"I mean, I think your whole genome sequence you can do for 300 bucks now; it used to be a billion dollars." (said at 1:08:44)

The speaker's statement accurately reflects the dramatic historical trajectory of whole-genome sequencing costs. The initial Human Genome Project, which generated the first reference human genome sequence, required billions of dollars in cumulative project funding across international consortia. With the advent and maturation of massively parallel next-generation and third-generation sequencing platforms, the per-genome sequencing cost has plummeted by orders of magnitude—passing the '$1,000 genome' milestone and reaching the range of a few hundred dollars (e.g., platforms achieving costs of ~$1/Gb or ~$100–$300 for 30x coverage).

Unverified means no publication matching the claim was located; it does not prove the claim false. Spotted an error? See the corrections policy - disputes from the people quoted are prioritized.