Click here to return to the previous article in this series.
Click here to watch the companion YouTube video for this series.
In the previous article in this series, we discussed the metabolism of neuroactive steroids and their cognitive and behavioral impacts. If you have not read it, this article will be less valuable.
In this article, we will review the modern evidence on finasteride’s side effects, with a particular focus on its cognitive and behavioral effects.
FINASTERIDE & ITS ENZYMATIC TARGET
Finasteride is an inhibitor of the enzyme 5α-reductase (5AR). It was developed in the 1980’s for the treatment of benign prostatic hyperplasia (BPH), characterized by the growth of the prostate in response to androgenic hormones. BPH affects about 40 million men in the United States. In addition to its use for BPH, finasteride is one of only two FDA-approved medications for the treatment of androgenic alopecia, the miniaturization of hair follicles that results from androgen exposure over a man’s lifetime.
Recall that steroids are synthesized in the human body from sterols in the adrenals, gonads, and brain. Cholesterol is first converted into pregnenolone by the cholesterol side-chain cleavage enzyme (P450scc), from which all other steroids are derived through various enzymes. The target of finasteride, 5AR, converts deoxycorticosterone into dihydrodeoxycorticosterone (DHDOC), progesterone into dihydroprogesterone (DHP), and testosterone into dihydrotestosterone (DHT). Consequently, inhibiting 5AR also inhibits the synthesis of the neurosteroids androsterone, androstanediol, THDOC, allopregnanolone, and isopregnanolone.
5AR has three isoforms, the second and third of which are strongly inhibited by finasteride. Dutasteride, a 5AR inhibitor with growing popularity, inhibits all three isoforms strongly. The reader will remember that the type 1 isoform that finasteride inhibits less strongly than dutasteride is the main 5AR isoform in the brain. Consequently, dutasteride exhibits greater cognitive and behavioral side effects than finasteride. Nonetheless, neurosteroids are regularly imported into the brain through the blood-brain barrier, so any inhibition of 5AR will reduce the prevalence of neurosteroids in the brain.
Of all the natural steroids, DHT is the most powerful agonist of the androgen receptor (AR) and thus has the greatest potential to induce androgenic side effects in men and women. Androgens bind to the androgen receptor in the cell cytoplasm. The binding activates the receptor, allowing it to translocate into the nucleus, homodimerize, and act as a transcription factor, binding DNA response elements called androgen response elements (AREs) present in different target genes – essentially, changing gene expression in the body. Androgen receptors in the nuclei can modulate about 500 androgen response elements (AREs) and 200 androgen receptor-responsive genes.
Because of their effect on the androgen receptor, both testosterone and DHT cause dermal papilla cell apoptosis in a dose-dependent and time-related manner. Finasteride reduces the conversion of testosterone into the more potent androgen DHT, and can also bind lightly at the androgen receptor, reducing androgenic activity generally. At a dose of 1 mg taken daily, finasteride can prevent the progression of androgenic alopecia and benign prostate hyperplasia.
For the first two decades of its use, finasteride was thought to have mostly benign side effects, as researchers were impressed with its marginal effect on semen parameters and testosterone levels in healthy patients. Between 2010 and 2015, the first literature describing persistent side effects from finasteride were published, and in 2010 Merck included in their drug sheet a note that, rarely, persistent issues with erection had been observed in post-finasteride users. In 2012 the FDA mandated that the list of persistent side effects be expanded further.
FERTILITY, METABOLIC SYNDROME, AND AGING PHALLUSES
The major unwanted effects of finasteride involve fertility changes, metabolic syndrome, and psychological effects. Though this review will focus on finasteride’s effects on cognition and behavior, it is worthwhile to briefly discuss the other effects.
Finasteride was originally thought not to affect semen parameters. The first major randomized, double-blind, placebo-controlled study indicated that 1 mg finasteride did not affect semen parameters, but it only studied men without prior fertility problems. Prompted by case reports of men experiencing severe reductions in sperm quality, a later study examined 27 men who reported to an infertility clinic. It concluded that 1 mg/day of finasteride does worsen semen parameters in men with pre-existing fertility issues. It appears that while most men have minimal semen quality changes on finasteride, about 5% are hypersensitive to it and experience a dramatic reduction in sperm count and quality consequent to treatment.
In the last decade, finasteride has also been shown to produce symptoms of the metabolic syndrome in men and rodents. Specifically, 5AR inhibition appears to worsen insulin sensitivity as measured by glucose metrics, liver health as measured by ALT and AST enzymes, and low-density lipoprotein cholesterol levels, in the long-term. Because a lot of the metabolic effect of 5AR inhibition appears to depend on the type 1 isoform of 5AR, the effect is more pronounced with dutasteride use, though it is also evident with finasteride.
Though the role of steroids in metabolic and lipid health is less clear, physiologic levels of androgens appear necessary for the maintenance of liver health. It is established that androgen deprivation influences the development of liver disease, while androgen treatment reduces fatty liver disease. Mice lacking hepatic androgen receptors exhibit greater hepatic steatosis, and hepatic steatosis is associated with low serum testosterone and high serum DHEA in humans. On the other hand, super-physiologic doses of androgens are associated with the development of liver tumors and hepatocellular carcinoma.
Note that there are other, less well-known effects of androgen deprivation. For example, the observed reduction in erection quality in finasteride-treated patients has led academics to question whether the androgen-deprivation may be aging men prematurely. Consequent to androgen deprivation, post-finasteride patients exhibit upregulated androgen receptor density in penile skin tissue up to 8 years after discontinuation of the treatment, and androgen receptors were found to be upregulated in other tissues, including the prostate. Hypoandrogenism is thought to cause fibrosis in the corpora cavernosa of penile tissue through collagen fiber deposition and the inhibition of nitric oxide synthases. Moreover, 5AR inhibition induced smooth muscle cell apoptosis in the corpora cavernosa in rats, leading researchers to speculate that it could cause fibrosis and apoptosis in human penile tissue.
FINASTERIDE, THE PLASMA, AND CEREBROSPINAL FLUID
The most concerning side effects observed in finasteride-treated patients have been psychological and behavioral changes. Particularly, patients often exhibit an anxious or depressed phenotype and report chronic fatigue, loss of libido, sleep problems, decreased initiative, and difficulty concentrating.
Though it is obvious that downstream effects of 5AR inhibition on steroid synthesis may be the culprit behind the observed cognitive and behavioral symptoms, most studies only examine plasma levels of steroids. For good reason, patients are less willing to have their spine’s punctured to donate cerebrospinal fluid (CSF) for academics. Nonetheless, a team of Italian researchers managed to convince three groups of patients to do exactly that. This series of three studies will provide the basis for our analysis of the effects of finasteride on neuroactive steroids. Because they are necessarily small due to the invasiveness of the procedure, because of their weak study design, and because the psychological symptoms exist in complex systems, we will have to compare the results with what we know of the biochemistry of finasteride and neuroactive steroids. Here, I report only results that were statistically significant.
In 2013, Melcangi et al. began the series of studies with only 3 post-finasteride patient volunteers and compared their plasma and CSF results to 5 healthy volunteers. On average, the three patients had used finasteride for 3.6 years and ceased treatment 5.9 years before commencement of the study. Due to the small study size, the only statistically significant outcomes in plasma were undetectable DHP, raised androstanediol, and raised estrogen, while in CSF allopregnanolone and isopregnanolone were undetectable, testosterone was raised, and estrogen was raised.
In 2014, Caruso et al.  examined the plasma and CSF of 7 post-finasteride patients with persistent symptomology and compared their results to 12 healthy volunteers. On average, patients used finasteride for 2 years and discontinued treatment 4.4 years prior to the study. In this study, DHEA was unchanged in both CSF and plasma while estradiol was again raised in plasma. Androstanediol was raised in both plasma and CSF. Progesterone was reduced in CSF, its metabolite DHP was reduced in plasma and undetectable in CSF, allopregnanolone was undetectable in both CSF and plasma, and isopregnanolone was raised in plasma. Pregnenolone was raised in both CSF and plasma.
In 2017, Melcangi et al.  analyzed the plasma and CSF of 14 post-finasteride patients who reported persistent side effects and compared them to a control of 25 healthy subjects. The post-finasteride patients used finasteride for 2.8 years and had discontinued treatment 5.4 years earlier, on average. The 14 subjects exhibited increased DHEA and testosterone in both CSF and plasma and increased androstanediol in CSF. However, they exhibited lower CSF estradiol and progesterone, though plasma levels were unaffected. Of the progesterone metabolites, DHP was lowered in both CSF and plasma while allopregnanolone was undetectable in plasma. The post-finasteride patients also exhibited lower pregnenolone levels in the CSF, but higher pregnenolone levels in plasma.
Though there are major differences in the outcomes of the studies, there are also observable patterns. Unfortunately, this team of researchers did not measure THDOC, androsterone, or the sulfated forms of pregnenolone and DHEA. Nevertheless, p