Finasteride Safety Signals and FDA Actions: What the Evidence Shows

What has the FDA actually done to the finasteride label?

Proscar (finasteride 5 mg) was approved on June 19, 1992 as a new molecular entity for benign prostatic hyperplasia, under priority review. Propecia (finasteride 1 mg) was approved on December 19, 1997 as a new dosage form for male pattern hair loss. Both are prescription products, and the current NDA holder for both is Organon [4][5].

Two safety actions account for most of the label language patients ask about.

June 9, 2011. The FDA issued a Drug Safety Communication on high-grade prostate cancer with 5-alpha reductase inhibitors, and labeling supplements for both applications carry that date [2][4][5]. The current labels state that 5-alpha reductase inhibitors may increase the risk of high-grade prostate cancer and quote the trial figures for Gleason 8 to 10 disease [2].

May 31, 2022. The FDA sent a Safety Labeling Change order under section 505(o)(4) of the FD&C Act citing new safety information on the risk of suicidal ideation and behavior, additional risk of sexual dysfunction, and psychoneurocognitive adverse events [3]. The resulting supplement, S-030, was approved on August 1, 2022 and changed the postmarketing wording to 'depression, suicidal ideation and behavior' [3].

Is there a boxed warning for depression, suicide or sexual dysfunction?

No. Depression and suicidal ideation appear in the Postmarketing Experience section (6.2) rather than as numbered Warnings and Precautions subsections [1]. The Propecia label of June 2021 listed postmarketing 'Nervous System/Psychiatric: depression' [3]. The current Proscar label lists depression, suicidal ideation and male breast cancer among postmarketing reports [2]. The labels describe these reports as voluntary, from a population of uncertain size, with no reliable way to establish a causal relationship [1][2].

What the trials measured for sexual side effects

In the three 12-month controlled trials of Propecia (945 men on drug, 934 on placebo, ages 18 to 41), drug-related decreased libido was reported by 1.8% vs 1.3%, erectile dysfunction by 1.3% vs 0.7%, and ejaculation disorder by 1.2% vs 0.7%. One or more of those events was reported by 3.8% (36 of 945) on finasteride vs 2.1% (20 of 934) on placebo, p=0.04 — a 1.7 percentage-point difference. Discontinuation for drug-related sexual events was 1.2% vs 0.9%, resolution occurred in men who stopped because of these effects and in most of those who continued, and the incidence of each event fell to 0.3% or lower by year 5 [1].

A self-administered sexual function questionnaire at month 12 favoured placebo in three of four domains — sexual interest, erections and perception of sexual problems — but showed no significant difference on overall satisfaction with sex life [1].

At 5 mg in the four-year PLESS trial (1,524 on finasteride, 1,516 on placebo, men aged 45 to 78 with BPH), year-one impotence was 8.1% vs 3.7%, decreased libido 6.4% vs 3.4%, and decreased ejaculate volume 3.7% vs 0.8%. In years 2 to 4 there was no significant difference in impotence (5.1% vs 5.1%) or decreased libido (2.6% vs 2.6%) [2].

A 2019 meta-analysis of 15 randomized placebo-controlled trials in 4,495 men treated for androgenetic alopecia found a relative risk of sexual dysfunction of 1.57 (95% CI 1.19 to 2.08) for finasteride 1 mg or dutasteride 0.5 mg; for finasteride alone the relative risk was 1.66 (95% CI 1.20 to 2.30) [9].

Do sexual side effects persist after stopping?

The labels address this directly. Propecia's postmarketing section lists 'sexual dysfunction that continued after discontinuation of treatment, including erectile dysfunction, libido disorders, ejaculation disorders, and orgasm disorders', plus male infertility and poor seminal quality with reported normalization or improvement after discontinuation [1]. The Proscar label reports that persistent sexual dysfunction was reported rarely in men taking 5 mg for BPH; most were older and taking other medicines, and the independent role of Proscar is unknown [2]. The Medication Guide describes 'decrease in sex drive that continued after stopping the medication' and 'problems with ejaculation that continued after stopping medication' [1].

The trial literature has been criticised on this point. A 2015 meta-analysis of 34 finasteride trials for androgenetic alopecia found that none had adequate safety reporting, 76% evaluated safety for one year or less, funnel plots suggested systematic under-detection of sexual adverse effects, and only 31% of 5,704 men treated in real-world practice would have met the entry criteria of the pivotal trials [10].

Smaller studies recruited men who already reported persistent symptoms. A small study interviewed 71 otherwise healthy men aged 21 to 46 who reported new sexual dysfunction lasting at least three months after stopping finasteride; mean finasteride use was 28 months and mean persistence was 40 months, and 94% reported low libido [11]. That design cannot produce a frequency or a cause [11]. In a follow-up comparison, 75% (46 of 61) of former users with persistent sexual symptoms scored 14 or higher on the Beck Depression Inventory-II versus 10% (3 of 29) of never-users with hair loss, P<0.0001 — a small, self-selected, cross-sectional comparison that cannot prove cause [12].

Post-finasteride symptoms: what the small studies found

Men who report ongoing sexual, mood and cognitive symptoms after stopping finasteride have been studied in several small groups. In 16 men reporting these symptoms, 8 met criteria for major depressive disorder and all had erectile dysfunction; their cerebrospinal fluid showed altered levels of several neuroactive steroids, including lower DHT and higher testosterone [13]. A separate 26-versus-26 case-control study found higher androgen receptor expression in penile skin cells from symptomatic former users than from controls (9.961 vs 9.494, adjusted P=0.01) [14].

A controlled comparison of symptomatic former users, asymptomatic users and men who never used finasteride found no evidence of androgen deficiency, decreased peripheral androgen action or persistent peripheral inhibition of 5-alpha reductase in the symptomatic group [15]. Those men did have impaired sexual function, higher depression scores and abnormal fMRI responses to sexual and affective stimuli compared with asymptomatic users and never-users [15].

The prostate cancer signal

The Prostate Cancer Prevention Trial randomized 18,882 men aged 55 or older to finasteride 5 mg or placebo for seven years. Prostate cancer was found in 18.4% (803 of 4,368) on finasteride and 24.4% (1,147 of 4,692) on placebo, a 24.8% relative reduction (95% CI 18.6 to 30.6) [6]. Gleason 7 to 10 tumours were more common on finasteride — 280 of 757 tumours (37.0%), or 6.4% of the 4,368 men — and the labels quote Gleason 8 to 10 disease at 1.8% vs 1.1% [6][2].

The Proscar label states that it is not approved for the prevention of prostate cancer [2]. The label also says it has not been established whether the reduction in prostate volume, or study-related factors, affected the high-grade findings [2].

A modelling reanalysis of the trial addressed the volume question. Median prostate volume was 25% lower on finasteride (25.1 vs 33.5 cm3), the likelihood of detecting high-grade cancer fell as volume rose (odds ratio 0.81 per 10 cm3, 95% CI 0.74 to 0.90), and the odds ratio for high-grade cancer dropped from 1.27 (95% CI 1.05 to 1.54) to 1.03 (0.84 to 1.26) after adjusting for gland volume and number of biopsy cores [16].

In longer follow-up, 15-year survival was 78.0% with finasteride and 78.2% with placebo; the study found no statistically significant difference in overall survival [17]. That follow-up reported Gleason 7 to 10 cancer in 3.5% vs 3.0% (relative risk 1.17, 95% CI 1.00 to 1.37, P=0.05) [17]. For men using 1 mg, the label states that the clinical significance of these findings with respect to use of Propecia is unknown [1].

Depression and suicide: what the cohort data show

Two large cohorts of older men give the clearest numbers, and neither shows a completed-suicide increase. In Ontario, 93,197 men aged 66 or older starting a 5-alpha reductase inhibitor (mostly for BPH) were matched to unexposed men: suicide hazard ratio 0.88 (95% CI 0.53 to 1.45), self-harm 1.88 (1.34 to 2.64) in the first 18 months only, and depression 1.94 (1.73 to 2.16) in the first 18 months and 1.22 (1.08 to 1.37) afterwards [7]. In Sweden, among 2,236,876 men aged 50 to 90 (70,645 finasteride starters, median age 73), finasteride was associated with depression (HR 1.61, 95% CI 1.48 to 1.75) but not with suicide (HR 1.22, 95% CI 0.99 to 1.49) [8]. Neither cohort was limited to young men taking 1 mg for hair loss.

A pharmacovigilance analysis of the WHO VigiBase report database found 356 reports of suicidality and 2,926 reports of psychological adverse events in finasteride users, with a reporting odds ratio of 1.63 (95% CI 1.47 to 1.81) for suicidality and 4.33 (4.17 to 4.49) for psychological events; the signal was stronger in men aged 45 or younger (reporting odds ratio 3.47, 2.90 to 4.15) and in the alopecia subgroup (2.06, 1.81 to 2.34) [18]. A reporting odds ratio describes what gets reported, not what happens to a given patient.

What the labels tell patients to watch for

• PSA. Propecia lowered mean PSA from 0.7 to 0.5 ng/mL at month 12 in men aged 18 to 41, and Proscar lowers PSA by about 50% within six months; the Proscar label says an isolated PSA value taken after six months of treatment should be doubled for comparison with normal ranges in untreated men. The Proscar label says that any confirmed rise from the lowest PSA value during treatment may signal prostate cancer and should be evaluated, even when the PSA remains within the usual range for untreated men [1][2].
• Pregnancy and handling. Finasteride is contraindicated in women who are or may become pregnant, who should not handle crushed or broken tablets because of the risk of abnormal external genitalia in a male fetus; intact coated tablets prevent contact with the active ingredient [1].
• Breast changes. Breast tenderness, breast enlargement and male breast cancer appear among postmarketing reports, with the relationship to long-term use unknown [1][2].
• Other reported events. Hypersensitivity reactions including rash, itching, hives and angioedema, testicular pain and hematospermia [1].
• Seminal volume. After 24 weeks at 5 mg in healthy volunteers there was no clinically meaningful effect on sperm concentration, motility or morphology, but median ejaculate volume fell 0.6 mL (22.1%), reversible with an average time to return to baseline of 84 weeks [2].