How Merck buried finasteride’s full impact on hormone signaling

SUMMARY: Merck explained to FDA that finasteride lowers levels of one hormone while increasing levels of its precursor. It sounds simple and narrowly targeted. In fact, finasteride disrupts hormonal pathways much more broadly. These pathways have crucial roles in the brain, reproductive system and other organs. Since approval, men who used finasteride have developed severe and lasting dysfunctions linked to the same organs that depend on hormone signaling. Merck’s selective account of the drug mechanism concealed the risks.

Figure 1. Effects of finasteride on steroid levels: Merck’s story vs. research findings. The figure incorporates findings from studies with different methodologies (see supplementary table). Left column: Merck described effects of finasteride on testosterone and dihydrotestosterone (DHT) in its application to FDA. Right column: Finasteride affects levels of at least a dozen steroids, based on general biochemistry, Merck’s own research, and other studies.
Epiandrosterone+ includes two derivatives. 
*11β-hydroxydihydroandrostenedione, produced primarily in adrenal glands.

In 1997, Merck sought approval of the U.S. Food and Drug Administration to market finasteride as a treatment for male pattern baldness. The company explained that the drug worked by reducing levels of dihydrotestosterone (DHT), believed to contribute to male pattern hair loss.1,2 This was accompanied by a modest increase in levels of testosterone. There was no mention of finasteride’s effects on other hormonal pathways.

The company repeatedly used the word “selective” to describe its affinity to a type of enzyme. An earlier publication explained that “finasteride produces a selective androgen deprivation by lowering DHT but not T with very few drug-related side effects…”3 (emphasis added). Yet general biochemistry and Merck’s own studies indicated a much broader impact on signaling across multiple organs and functions.

To this day, most researchers, physicians and patients only describe the effect of finasteride on DHT. While the drug’s effect on hair growth may be due to the suppression of DHT, the other effects are no less important from a safety perspective.

Steroids: What are they and what do they do?

The term “steroids” might bring to mind bodybuilding, but here they are a set of chemical messengers supporting a wide range of behaviors including sexuality, the stress response, sleep, thought and emotions. In a more general sense, steroid signaling helps us seize opportunities and avoid risks in an ever-changing environment.

Testosterone and estrogen are well-known steroids, but there are many others. They are created in a series of transformations enabled by different enzymes. The process starts with cholesterol, which binds with an enzyme to produce another steroid called pregnenolone (Figure 2, bottom left). The latter is then converted into two more steroids via enzymes. At the end of the pathway, steroids bind to receptors supporting the behaviors mentioned above: male sexual behavior, memory, stress response, sleep, thought and emotions.4,5

The enzyme 5-alpha reductase (5-AR), circled in Figure 2, converts testosterone to DHT, and is involved in other steroid conversions.

Figure 2. Steroid metabolism in the brain. The enzyme 5-alpha reductase is used in three reactions. Adapted from Reddy, 20136 under CC BY license.

Finasteride broadly disrupts steroid synthesis

Now we turn to how finasteride affects steroid signaling. Finasteride binds to the 5-AR enzyme, partially disabling it. Now the enzyme cannot convert steroids as efficiently, reducing levels of its downstream products. Figure 3 shows that finasteride directly reduces levels of six steroids.

Figure 3. Conversions involving 5-alpha reductase enzyme. This is an original figure based on information in Traish, 2012.7

Like putting a dam in a river, disabling an enzyme has effects farther downstream and upstream. Figure 4 shows how inhibiting 5-AR has far-reaching effects on steroid synthesis in the brain (there are similar effects in other organs such as the reproductive system, liver, kidneys, skin and eyes).

Figure 4. Steroid synthesis in the brain. Red zigzags represent finasteride’s inhibition of 5-alpha reductase. Red arrows indicate decreased steroid levels, while blue arrows indicate increased steroid levels. Finasteride not only reduces the immediate products of the enzyme, but also other steroids downstream and upstream. Supporting research findings are summarized in the supplementary table below. Figure adapted from Reddy, 20136 under CC BY license.
Merck knew

In 1990, a paper by a Merck researcher and collaborators noted (emphasis added): “Finasteride-treated subjects…have global defects that affect both hepatic and peripheral 5ɑ-reductase activities.”8 Yet when Merck submitted its Propecia application to FDA in 1997, it only pointed out reduced DHT and a modest increase in testosterone. A later paper by Merck researchers showed two conversions involving 5-AR in a figure, but the text only mentioned effects on DHT.9 Today, official prescribing information for physicians only mentions effects on DHT.10

FDA dropped the ball

In 1997, a few weeks before FDA approved Propecia, an Advisory Committee met to discuss the application. A panel of outside experts heard statements of FDA reviewers and asked questions of Merck representatives. FDA’s medical reviewer for Propecia, Dr. Hon-Sum Ko, expressed concern that finasteride affects a wide range of steroid conversions.11 His comment is worth quoting in full (emphasis added; there appear to be errors in chemical names, likely because this has been transcribed from a spoken statement).

5-alpha-Reductase [type] II is not an enzyme that is specific just for conversion of testosterone. It has other steroid hormonal substrates which are possible, and included in that would be progesterone; you can have 5-alpha-Dihydroprogesterone, and in one of the recent publications, it is known that dihydroprogesterone, the effect on the nervous system is probably the factor mediated for progesterone, because finasteride can abolish the effect of progesterone and not dihydroprogesterone, and also, you can see that dihydrotestosterone also has an effect on the expression of this particular gene in the nervous system.

So, to again reiterate the point that 5-alpha-Reductase II may have other substrates besides testosterone, and I have not the time to go into others like cortisol and aldosterone, which also have 5-alpha metabolites which may have activities.

This matter did not come up again in the meeting, nor in the drug approval packet. Today, Dr. Ko’s concerns seem prophetic. Two of the most active researchers on harms of finasteride, Dr. Abdulmaged Traish and Dr. Roberto Melcangi, have focused on disruptions to steroid levels. In a 2021 webinar, Dr. Irwin Goldstein made a comment that is strikingly similar to Dr. Ko’s comments.


Inhibiting the 5-AR enzyme directly impacts synthesis of six different steroids as well as others upstream and downstream. When seeking approval of Propecia, Merck limited their explanation to the suppression of DHT, which is believed to be the basis of finasteride’s efficacy for treating hair loss.

Merck frequently described finasteride as a “selective” inhibitor of one type of 5-AR. This was true as far as it went,A but they were silent about the effects on steroids other than DHT and testosterone. If the drug’s full biochemical impact had been disclosed earlier, or if FDA had investigated the matter, it could have raised red flags.

In the decades since the drug’s approval, a pattern has emerged in men who experienced harms linked to finasteride. The most commonly affected functions—sexuality, mental status, cognition and memory—all depend on steroid signaling.

Did Merck’s selective story break FDA rules? The agency’s guidelines as of 1997 have not been found, so this remains an open question. Current FDA guidance calls for drug makers to address toxicity related to the drug’s selectivity.12

Merck’s selective story reveals a common tactic of pharmaceutical companies: focus on effects within a strictly limited domain, while ignoring or concealing implications for the whole system. If a drug maker had to assess effects across the whole body, over the long-term, in a realistic population, safety concerns would pile up. This would reduce the chances of approval, it would cost more, and even if approved, the drugs would be less marketable.

The game, then, is to create and solve artificially constrained problems. This is a game they can win, reaping stunning profits while risks and harms are externalized—that is, borne by other parties. Patients may experience severe and irreversible harms of drugs such as finasteride. Even though patients pay for health care, they have little or no standing when things go wrong. Doctors and regulators tend to ally with the drug maker, using time-honored tactics of denying, delaying, discounting, deflecting, and blaming the patient.


A. In the late 1990s, there were two known types of 5-AR. Finasteride has a strong affinity for type 2 and a weak affinity for type 1. This is why Merck described the drug as a “selective” inhibitor. In 2008, a third type of 5-AR was reported. Finasteride was found to have a strong affinity for type 3 as well as type 2.13


💊 = conflict of interest due to Merck funding; or authorship by Merck researchers or physicians affiliated with Merck.

  1. U.S. Food and Drug Administration; Division of Dermatologic and Dental Drug Products. Review and Evaluation of Pharmacology and Toxicology Data (NDA 20-788 000-BP). Submitted November 5, 1997.
  2. 💊 [Authored by Merck researchers] Kaufman KD [details], Dawber RP. Finasteride, a type 2 5ɑ-reductase inhibitor, in the treatment of men with androgenetic alopecia. Expert Opin Investig Drugs. 1999. doi:10.1517/13543784.8.4.403 • PubMed [By Merck researchers; lead author was Director of Propecia clinical development]
  3. 💊 [Authored by a Merck researcher] Gormley GJ [details]. Finasteride: a clinical review. Biomed Pharmacother. 1995;49(7-8):319-324. doi:10.1016/0753-3322(96)82658-8 [Author is a Merck researcher]
  4. Giatti S, Diviccaro S, Serafini MM, et al. Sex differences in steroid levels and steroidogenesis in the nervous system: physiopathological role. Front Neuroendocrinol. 2020. doi:10.1016/j.yfrne.2019.100804PubMed
  5. Schiffer L, Barnard L, Baranowski ES, et al. Human steroid biosynthesis, metabolism and excretion are differentially reflected by serum and urine steroid metabolomes: a comprehensive review. J Steroid Biochem Mol Biol. 2019. doi:10.1016/j.jsbmb.2019.105439PMC full text
  6. Reddy DS. Role of hormones and neurosteroids in epileptogenesis. Front Cell Neurosci. 2013. doi:10.3389/fncel.2013.00115 • PMC full text
  7. Traish AM. 5α-reductases in human physiology: an unfolding story. Endocr Pract. 2012. doi:10.4158/EP12108.RAPubMed
  8. Imperato-McGinley J [COIs], Shackleton C, Orlic S, Stoner E 💊 [details). C19 and C21 5β/5ɑ metabolite ratios in subjects treated with the 5ɑ-reductase inhibitor finasteride: comparison of male pseudohermaphrodites with inherited 5ɑ-reductase deficiency. J Clin Endocrinol Metab. 1990. doi:10.1210/jcem-70-3-777 • PubMed [Includes a Merck researcher (ES) as co-author]
  9. 💊 [Authored by Merck researchers] Kaufman KD [details], Dawber RP. Finasteride, a type 2 5ɑ-reductase inhibitor, in the treatment of men with androgenetic alopecia. Expert Opin Investig Drugs. 1999. doi:10.1517/13543784.8.4.403 • PubMed [By Merck researchers; lead author was Director of Propecia clinical development]
  10. Merck & Co., Inc. Prescribing Information: PROPECIA (finasteride) Tablets, 1 mg. December 8, 2001.
  11. U.S. Food and Drug Administration; Center for Drug Evaluation and Research; Dermatologic and Ophthalmic Drugs Advisory Committee. Meeting No. 48 regarding Propecia, sponsored by Merck. November 13, 1997. Archived
  12. U.S. Food and Drug Administration; Center for Drug Evaluation and Research; Center for Biologics Evaluation and Research. Clinical Pharmacology Section of Labeling for Human Prescription Drug and Biological Products — Content and Format. Guidance for Industry. December 2016. Accessed March 17, 2023.
  13. Yamana K, Labrie F, Luu-The V. Human type 3 5α-reductase is expressed in peripheral tissues at higher levels than types 1 and 2 and its activity is potently inhibited by finasteride and dutasteride. Horm Mol Biol Clin Investig. 2010. doi:10.1515/HMBCI.2010.035 • PubMed
  14. 💊 [Co-authored by Merck researchers] Gormley GJ [details], Stoner E [details], Rittmaster RS [details], et al. Effects of finasteride (MK-906), a 5 alpha-reductase inhibitor, on circulating androgens in male volunteers. J Clin Endocrinol Metab. 1990. doi:10.1210/jcem-70-4-1136 • PubMed [Funded by Merck and co-authored by Merck researchers]
  15. Dušková M, Hill M, Stárka L. The influence of low dose finasteride, a type II 5α-reductase inhibitor, on circulating neuroactive steroids. Horm Mol Biol Clin Investig. 2010. doi:10.1515/HMBCI.2010.010 • PubMed
  16. Stanczyk FZ, Azen CG, Pike MC. Effect of finasteride on serum levels of androstenedione, testosterone and their 5α-reduced metabolites in men at risk for prostate cancer. J Steroid Biochem Mol Biol. 2013. doi:10.1016/j.jsbmb.2013.02.015 • PubMed
Further reading

Marchetti PM, Barth JH. Clinical biochemistry of dihydrotestosterone. Ann Clin Biochem. 2013. doi:10.1258/acb.2012.012159PubMed

Supplementary table

Studies of finasteride’s effect on steroid levels

Study and methodRelevant findings
Imperato-McGinley et al (1990)8
Included a Merck researcher as co-author and used Merck trial data.
The subgroup of interest were 29 men aged 45–75, taking finasteride doses of 0.2–80 mg per day. Steroid ratios were measured by assay of urinary metabolites at 0, 3 and 6 months.
– Elevated ratio of etiocholanolone(5β)/androsterone(5ɑ)
– Elevated 5β/5ɑ ratios of 11β-hydroxyandrostenedione, cortisol, and corticosterone
(Finasteride lowers levels of 5ɑ-reduced steriods which are in the denominator of each ratio.)
Gormley et al (1990)14
Three authors were Merck researchers, and the study was supported by grants from Merck.
30 male volunteers ages 40–77 took low doses of finasteride (0.04, 0.12, 0.2, and 1.0 mg) for 14 days. Steroid levels were measured by serum assay.
On day 14, androstanediol glucuronide and androsterone glucuronide levels were decreased in each group receiving finasteride.
Suppression was similar to that of dihydrotestosterone.
Dušková et al (2009)15
20 men with average age of 69.2 with benign prostate hyperplasia. Took finasteride 5 mg for 4 months. Steroid levels were measured by serum assay.
Androsterone, epiandrosterone, dehydroepiandrosterone, 16α-hydroxydehydroepiandrosterone, 7α-hydroxydehydroepiandrosterone and allopregnanolone levels were all significantly lower.
Allopregnanolone was down a median of over 300%. Epietiocholanolone was up by over 600%.
Stanczyk el al (2013)16
~26 men aged 57–79 years with elevated PSA levels were administered finasteride (5 mg). Blood was taken at baseline, 1, 3, 6 and 12 months. Steroid levels were measured by serum assay.
At 1 month
– 3α-androstanediol glucuronide ↓ 75.7%
– Androsterone glucuronide ↓ 43.0%

At 3 months
– Androsterone glucuronide ↓ 22.2%
– Testosterone ↑ 34.5%
– Androstenedione ↑ 18.3%
Traish (2012)7
Review article — see text for citations.
5ɑ-reductase is involved in these conversions:
– Testosterone to 5ɑ-dihydrotestosterone
– 4-Androstenedione to 5ɑ-dihydroandrostenedione
– Progesterone to 5ɑ-dihydroprogesterone
– Deoxycorticosterone to 5ɑ-dihydrodeoxycorticosterone
– Corticosterone to 5ɑ-dihydrocorticosterone
– Aldosterone to 5ɑ-dihydroaldosterone