Olsen Finasteride 5-year Extension Data and What Happened After the Trial Ended

Print or save as PDF

What the five-year extension actually measured

The long-term data come from two 1-year, placebo-controlled Phase III trials that randomized 1,553 men aged 18 to 41 with mild to moderate male pattern hair loss, plus up to four further 1-year placebo-controlled extension studies that 1,215 of those men entered [2]. Hair was counted in a 1-inch diameter circle (5.1 cm²) of balding vertex scalp, where the men averaged 876 hairs at baseline [3].

The extensions did not drop the placebo arm; they were placebo-controlled 1-year studies [2]. What changed was the size of that arm. 323 men on finasteride and 23 men on placebo entered the fifth year [1]. The studies also carried two useful side groups. 543 men took placebo for the first 12 months and then started finasteride, and 290 of them were still in the study in year 5 [1]. A further 65 men took finasteride for 12 months and were then switched to placebo, a group the studies included specifically to evaluate what happens when treatment stops [1].

The two-year publication is the 1998 Journal of the American Academy of Dermatology report [3]. The five-year report is credited to the Finasteride Male Pattern Hair Loss Study Group and appeared in the European Journal of Dermatology in 2002 [2].

The hair count curve: best at two years, slow decline after

The Propecia label describes the shape of the curve directly: the maximum improvement in hair count compared with baseline was reached during the first two years, and although that was followed by a slow decline, hair count stayed above baseline throughout the five years of the studies [1]. Because hair loss in the placebo group was faster, the difference between the groups widened over time — 107 hairs at 12 months (P<0.001) [1], 138 hairs at two years (P<0.001) [3], and 277 hairs at five years (219 men on finasteride versus 15 on placebo, P<0.001) [1].

Both of the early numbers are differences from placebo, not changes from baseline. The 1998 report put the finasteride advantage at 107 hairs at one year and 138 at two years, while placebo produced progressive hair loss [3]. The five-year comparison rests on 219 finasteride and 15 placebo hair counts, so the 277-hair gap is measured against a very small untreated group, even though the direction was consistent across hair counts and across investigator and photographic assessment [1][2].

HealthRX.com offers finasteride 1 mg by prescription.

What happened when the trial ended

The question people actually ask — what happens if I stop — is answered in the label rather than in a follow-up study. The Propecia label says continued use is recommended to sustain benefit, that benefit should be re-evaluated periodically, and that withdrawal of treatment leads to reversal of effect within 12 months [1]. The trial's own version of that experiment was the 65-man group switched from finasteride to placebo after 12 months [1].

The label's postmarketing section also lists sexual dysfunction that continued after discontinuation of treatment, including erectile dysfunction, libido disorders, ejaculation disorders and orgasm disorders [1]. The label states that these reports are voluntary, come from a population of uncertain size, and that a causal relationship cannot reliably be established [1].

Five years of sexual adverse events

In the three 12-month controlled trials (945 men on finasteride, 934 on placebo), drug-related decreased libido was reported by 1.8% versus 1.3%, erectile dysfunction by 1.3% versus 0.7%, and ejaculation disorder by 1.2% versus 0.7%. One or more of these was reported by 36 of 945 men on finasteride (3.8%) versus 20 of 934 on placebo (2.1%), p=0.04, and 1.2% versus 0.9% stopped treatment because of them [1]. Resolution occurred in the men who stopped because of these effects and in most of those who continued [1].

With longer exposure the reported rates fell rather than rose: the label states that the incidence of each of these events decreased to 0.3% or less by the fifth year of treatment [1]. The five-year report concluded that finasteride was generally well tolerated and that no new safety concerns were identified [2].

A self-administered sexual function questionnaire told a slightly different story at month 12: statistically significant differences in favor of placebo appeared in three of four domains — sexual interest, erections and perception of sexual problems — with no significant difference in overall satisfaction with sex life [1].

How complete the safety picture is has itself been questioned. A 2015 meta-analysis of 34 finasteride trials for androgenetic alopecia found that none had adequate safety reporting, that 76% evaluated safety for one year or less, and that funnel plots suggested under-detection of sexual adverse effects [5].

PSA and the prostate cancer signal at 5 mg

Propecia 1 mg lowered mean PSA from 0.7 ng/mL to 0.5 ng/mL at month 12 in men aged 18 to 41 [1]. The familiar 50% reduction and the rule of doubling an isolated PSA value belong to Proscar 5 mg in older men with BPH: Proscar reduces PSA by approximately 50% within six months, and an isolated value in a man who has taken it for six months or more should be doubled for comparison with normal ranges in untreated men [6]. On Propecia, any confirmed increase from the lowest PSA value while on treatment may signal prostate cancer and should be evaluated even if the value is still within the normal range [1].

The high-grade cancer finding comes from the 7-year Prostate Cancer Prevention Trial of finasteride 5 mg, in which Gleason 8 to 10 cancer was found in 1.8% of men on finasteride versus 1.1% on placebo [1]. The label states that 5-alpha reductase inhibitors may increase the risk of high-grade prostate cancer and that the clinical significance for men taking 1 mg is unknown [1].

Where the five-year data are thin

• Vertex only. The long-term counts came from the two vertex trials [1]. The label states that efficacy in bitemporal recession has not been established, and it describes the extension studies only for the two vertex trials, not for the separate 326-man trial of anterior mid-scalp loss [1].
• The population was 88% Caucasian, aged 18 to 41, with mild to moderate loss [1].

How the five-year results fit with pooled trial data

A 2010 systematic review by Mella and colleagues in Archives of Dermatology pooled 12 randomized trials covering 3,927 men [7]. The review reported a 24.3% mean hair-count difference from placebo in trials lasting at least 24 months (95% CI 17.92 to 30.60), against 9.42% in trials of 12 months or less; the relative risk of a patient-reported improvement was 1.71 long term, with a number needed to treat of 3.4 [7]. That percentage should not be directly compared with the five-year study's 277-hair difference [1][7].