Dr. Mark Gordon & Andrew Marr on the Joe Rogan Experience #1056, fact-checked

What the evidence shows: Gordon appears to be citing a real 2004 clinical trial (Mahajan et al., European Journal of Endocrinology), but reverses its population and overstates its scope. That study enrolled 25 adults with growth hormone deficiency (GHD) already confirmed to be caused by identified pituitary or hypothalamic disease; it did not start with a pool of depressed patients and screen them for GHD. Among the 18 participants with adult-onset GHD, 11 (61%) were found to have atypical depression at baseline, and their depression scores improved after growth hormone replacement in this small, double-blind, placebo-controlled crossover trial. Gordon's framing implies the reverse: that screening people with treatment-resistant depression would reveal 61% have undiagnosed GHD, a claim the study does not support. Current clinical guidance restricts GH-deficiency testing to patients with a documented history of pituitary/hypothalamic disease, cranial irradiation, or brain injury, and explicitly discourages testing patients with a low pretest probability, such as depression occurring without any such history. The cited finding rests on a single, small (n=25), two-decade-old trial in a pre-selected pituitary-disease population and has not been established as a basis for evaluating or treating depression in the general population.

What the evidence shows: The 19-times figure traces to a 2009 telephone survey commissioned by the NFL Player Care Foundation and conducted by the University of Michigan's Institute for Social Research, which found that 1.9 percent of former NFL players aged 30-49 self-reported a dementia-related diagnosis versus a cited general-population rate of 0.1 percent. That survey was never peer-reviewed, relied on self- or family-reported diagnoses rather than clinical evaluation, and measured cumulative career exposure to repeated mild traumatic brain injuries, not a single concussion. Peer-reviewed epidemiological evidence shows a much smaller association: a meta-analysis of cohort studies covering more than four million people found traumatic brain injury overall was associated with about a 17 percent higher risk of Alzheimer's disease (RR 1.17), rising to about 30 percent for moderate-to-severe TBI specifically (RR 1.30), with no comparable large single-concussion risk multiplier established in the literature. A review of former-NFL-player dementia risk likewise attributes elevated risk to repeated mild traumatic brain injuries accumulated over a career rather than to a single concussion. The claim's attribution of a 19-fold risk to "one major concussion" misrepresents both the origin (an unverified, non-peer-reviewed self-report survey of career NFL players) and the scope of the actual 19-times figure, and is far larger than risk increases found in peer-reviewed research on single or even moderate-to-severe brain injuries.

What the evidence shows: Mark Gordon claimed statin drugs have existed since 1986 and generate $36 billion a year in sales. On the timeline, Merck submitted lovastatin's (Mevacor) New Drug Application to the FDA in November 1986, but the drug was not actually approved for commercial use until September 1, 1987, according to a peer-reviewed historical review and a JAMA Cardiology retrospective. So "since 1986" is defensible if measured from the NDA submission date but is off by about a year if measured from market approval. On the sales figure, the same peer-reviewed review reports total statin-class sales of $25 billion in 2005, the closest independently sourced benchmark available and in the same general order of magnitude as Gordon's $36 billion figure. However, neither source provides a sales figure specific to 2017 (when this episode aired), so the $36 billion claim could not be independently verified for that year and its accuracy at the time of the broadcast remains unconfirmed.

What the evidence shows: Mark Gordon attributed the statistic that roughly half of all heart attacks occur in people with normal or low cholesterol to a "SmithKline-Beecham study." No such SmithKline-Beecham study could be located in the peer-reviewed literature or industry records searched. The underlying figure, however, closely matches real published data: a 2009 American Heart Journal analysis of 136,905 U.S. hospitalizations in the AHA's Get With The Guidelines registry (Sachdeva et al.) found that nearly half of patients admitted with coronary artery disease had LDL cholesterol below 100 mg/dL, the guideline-defined 'optimal' level. A separate clinical review of cholesterol-skeptic claims traces the closely related and often-repeated '52% of heart attack patients have normal cholesterol' statistic to a 2008 American Heart Journal study by Ghandehari et al., again with no connection to SmithKline Beecham. The approximate figure Gordon cites is therefore well-supported by legitimate registry data, but his specific attribution to a 'SmithKline-Beecham study' appears to be a misattribution with no identifiable primary source.

What the evidence shows: Andrew Marr cited internal, self-reported outcome figures (58 suicide-attempt veterans, 91% with a 50% improvement within three months) attributed to his own organization, Warrior Angels Foundation; no peer-reviewed publication of this specific dataset was located, and it appears to be uncontrolled clinical program data rather than a published study. The best controlled evidence on hormone replacement therapy for TBI does not support rapid, near-universal improvement of this magnitude: a Phase 2 randomized, placebo-controlled trial of recombinant human growth hormone in TBI patients (n=63) found no consistent benefit over placebo on standard functional and cognitive outcome measures at 6-12 months (a motor-score finding favoring treatment was confounded by baseline imbalances and could not be reliably attributed to the drug), and the trial did not assess suicide-specific outcomes. Reflecting persistent uncertainty about hormone therapy's benefit for TBI-related symptoms in veterans, the U.S. Department of Veterans Affairs is funding a $28 million, 20-site, placebo-controlled randomized trial specifically because existing evidence for growth hormone replacement in this population remains preliminary. No allowlisted, independently verified source corroborating Marr's specific 58-patient, 91%-improvement figures was found; the claim rests on unpublished, uncontrolled program data rather than validated clinical evidence.

What the evidence shows: Andrew Marr attributed a specific finding, that 60% of glaucoma patients fail to take their medication on time, to a study from an unnamed "glaucoma research institute." No named institute or single study reporting exactly that figure could be identified. Peer-reviewed literature does show poor adherence is common and roughly consistent with the claim: a 2023 narrative review states many studies show 60% as the average estimate of glaucoma medication adherence, with reported rates ranging from 5% to 80% depending on measurement method, and notes poor adherence results in uncontrolled intraocular pressure and thus vision loss and blindness. A pharmacy-refill study of 3,615 patients across two US chains found 63% (by MPR) to 70% (by PDC) of patients had unsatisfactory adherence (below an 80% coverage threshold) over 12 months. The approximate 60% figure and the link to blindness risk are broadly consistent with published research, but the specific attribution to a "glaucoma research institute" study appears to be a misremembered or garbled citation rather than a traceable primary source.