Is the Testosterone Crisis Real? The Numbers Behind the Headlines | Signal Ep 1

Every week there's a new headline saying men are losing testosterone. A quarter of men now start testosterone replacement therapy without ever getting their blood tested. The supplement aisle is full of boosters that either do nothing or contain undisclosed steroids. And the lab test that gets everybody to the pharmacy? Half of low results normalize on their own.

In Episode 1 of the Signal launch series, Dr. Jordan Feigenbaum and Dr. Austin Baraki (both MDs and strength coaches) walk through the three-layer problem with how testosterone gets diagnosed and treated in 2026, then take apart the "testosterone is crashing" headline with the most current data available, including a 2025 meta-analysis of more than one million men.

Pre-order our book, Signal: barbellmedicine.com/signal

Timestamps

• 0:00 Mark's story: treating the number, not the patient
• 1:18 Welcome to the Barbell Medicine Podcast
• 1:41 Problem 1: A quarter of men start TRT with no lab work
• 3:36 Problem 2: Why testosterone boosters do not work (and what is in them)
• 13:40 Problem 3: Why one low testosterone lab is not a diagnosis
• 19:19 Setup: Is the testosterone crisis headline real?
• 20:04 The MMAS data and the 1%-per-year number
• 20:52 The 2025 meta-analysis of over 1 million men
• 22:02 Why the headline is inflated: three causes
• 22:27 Cause 1: The testing method changed (immunoassay to mass spec)
• 25:58 Cause 2: BMI cannot see visceral fat
• 29:37 The Nyante study: when you fix both problems, the decline vanishes
• 33:58 What this actually means for you
• 37:05 The broken testosterone system, summarized
• 38:24 Five takeaways from this episode
• 39:14 Next week: How testosterone actually works
• 39:39 About Signal and credits
• 
  

What you'll learn in this episode:

•  Why 25% of new TRT prescriptions are written without any pre-treatment lab work (JAMA, 2015)
• What actually happens when researchers test 50+ "testosterone booster" supplements (spoiler: 12% are contaminated with undisclosed steroids)
• Why a single low testosterone reading is not a diagnosis, and the Massachusetts Male Aging Study data that proves it
• The real size of the population-level testosterone decline (much smaller than 1% per year)
• Why BMI cannot see the visceral fat that is driving most of the genuine decline
• The Nyante study that shows the decline essentially vanishes when you use an accurate test and measure waist circumference
• Five practical takeaways you can apply before your next lab draw

This is Episode 1 of a four-part series built around our upcoming book, Signal. Over the next four weeks we cover what testosterone actually is, how to tell when it is genuinely low, what is really driving population-level changes, and what the evidence says you can do about it.

Next Steps

• Check out our new book, Signal (coming soon)
• For evidence-based resistance training programs: barbellmedicine.com/training-programs
• For individualized training consultation: barbellmedicine.com/coaching
• Explore our full library of articles on health and performance: barbellmedicine.com/resources
• To consult with Drs. Baraki or Feigenbaum email us at support@barbellmedicine.com
• To support us and get ad free listening, plus special product discounts, and exclusive content, go to supercast.barbellmedicine.com

Resources

Baillargeon, J., et al. (2015). Trends in Androgen Prescribing in the United States, 2001–2011. JAMA Intern Med, 175(8), 1413–1415. — 25% no preceding lab; post-prescription monitoring gap.

Rao, P.K., et al. (2017). Trends in Testosterone Replacement Therapy Use from 2003 to 2013 among Reproductive-Age Men in the United States. J Urol, 197(4), 1121–1126. — Prescription volume growth.

Selinger, S., & Thallapureddy, A. (2024). Cross-sectional analysis of national testosterone prescribing through prescription drug monitoring programs, 2018–2022. PLoS One, 19(8), e0309160. — Recent prescribing data, 3-4 million estimate.

Vesper, H.W., et al. (2015). Serum Total Testosterone Concentrations in the US Household Population from the NHANES 2011–2012 Study Population. Clin Chem, 61(12), 1495–1504. — Population testosterone levels, NHANES data.

Clemesha, C.G., et al. (2020). "Testosterone Boosting" Supplements Composition and Claims Are Not Supported by the Academic Literature. World J Men's Health, 38(1), 115–122. — 62% no published data, 10% decreased T.

Tucker, J., et al. (2018). Unapproved Pharmaceutical Ingredients Included in Dietary Supplements Associated With US FDA Warnings. JAMA Network Open, 1(6), e183337. — 12% adulterated with undisclosed steroids.

Trost, L.W., & Mulhall, J.P. (2016). Challenges in Testosterone Measurement, Data Interpretation, and Methodological Appraisal of Interventional Trials. J Sex Med, 13(7), 1029–1046. — Half of low results normalize on repeat.

Travison, T.G., et al. (2008). The Natural History of Symptomatic Androgen Deficiency in Men: Onset, Progression, and Spontaneous Remission. JCEM. MMAS data — 50%+ spontaneous normalization.

Travison, T.G., et al. (2007). A Population-Level Decline in Serum Testosterone Levels in American Men. JCEM, 92(1), 196–202. — Original MMAS secular decline, 15–20% lower across cohorts.

Santi, D., et al. (2025). Meta-analysis of secular trend in total testosterone levels, 1971–2024. 1,256 studies, N > 1,000,000. — 0.56%/year adjusted; LH parallel decline; mass spec subgroup no significant decline.

 Methods note on the ~0.56% per year figure cited in this episode: the Santi paper does not report a single percentage rate. The headline adjusted meta-regression coefficient (−0.6 nmol/L/year) is inflated by the random-effects weighting scheme and is not a biological rate. The 0.5–0.6% per year approximation comes from the pre-2000 stratified subgroup (Fig. 5, coefficient −0.1 nmol/L/year) divided by the dataset mean of 18.5 nmol/L. The post-2000 stratum runs larger (~1.1%), and the age-stratified coefficients in Table 5 cluster in the 0.4–0.9% range. The mass spectrometry subgroup (Table 3, Group 4) showed no significant trend (p = 0.845). The episode uses the conservative end of this range as the most defensible estimate of the real population-level rate after accounting for assay drift.

Nyante, S.J., Graubard, B.I., Li, Y., McQuillan, G.M., Platz, E.A., Rohrmann, S., Bradwin, G., & McGlynn, K.A. (2012). Trends in sex hormone concentrations in US males: 1988–1991 to 1999–2004. Int J Androl, 35(3), 456–466. doi: 10.1111/j.1365-2605.2011.01230.x. — Archived NHANES samples, same platform, waist circumference added; no significant decline in total or free testosterone.