Olsen Finasteride 5-Year Subgroup Analyses: Who Responded Most and Least
At a glance
| Detail | Value |
|---|---|
| N | 1,553 randomized (reported in the trial; verify against primary source for exact extension-cohort counts) |
| Intervention | Finasteride 1 mg/day, oral |
| Comparator | Placebo (years 1-2), then open-label crossover to finasteride |
| Duration | 5 years |
| Ages enrolled | 18-41 years |
| Primary endpoint | Change in hair count in a fixed 1-inch vertex target area, by masked photographic count |
| Headline result | Hair count increased in the continuous-treatment group at year 1 and was sustained through year 5; the placebo group's hair count declined progressively before crossover |
What This Trial Actually Established
Finasteride is a synthetic type II 5-alpha-reductase inhibitor. The 1 mg oral tablet, marketed as Propecia, is FDA-approved specifically for androgenetic alopecia (male pattern hair loss) in men; a separate 5 mg formulation (Proscar) is approved for benign prostatic hyperplasia and is not the subject of this trial. The Olsen 5-year trial is the pivotal long-term randomized, placebo-controlled study behind the male AGA indication: continuous finasteride 1 mg sustained a hair-count increase in a standardized vertex target area through five years, while men on placebo lost hair count over the same period. That overall result is well established and is the basis for the FDA label.
What most summaries skip is the subgroup breakdown inside that same dataset, because it answers a more specific and more clinically useful question: given a particular patient's age, severity, and pattern, how much of that average benefit should they actually expect? This page focuses on that question, and on being explicit about which parts of the subgroup story are solid and which parts are approximate.
The useful question here is not whether finasteride works in men with AGA on average. It is which baseline profile in the trial captured most of that average effect, and where the trial's 18-to-41 age ceiling, its open-label extension design, and its substantial dropout rate limit how far that pattern can be extended to a patient sitting in front of a clinician today.
Trial Design Details That Shape How to Read the Subgroups
A few methodological points change how much weight any subgroup finding should carry.
Hair counts were made using a macrophotography technique within a fixed 1-inch circular target area on the vertex scalp, clipped and photographed at each visit, with counting performed by masked evaluators. This method is reproducible within that one anatomical zone, but it does not capture diffuse thinning across the rest of the scalp, and frontal/hairline counts were tracked as a separate measurement.
The trial enrolled men aged 18 to 41 at baseline. "Older" in this dataset means late 30s, not 55 or 65. Any statement about age and response applies inside that window; extending it to older patients is an extrapolation, not a trial finding.
Attrition over five years was substantial, and the published analyses relied heavily on last-observation-carried-forward (LOCF) methods alongside a smaller completer cohort. Patients who stayed in the study for the full five years may represent a self-selected group of responders, since those with poor results or side effects were more likely to leave. That means completer-based subgroup numbers likely present an optimistic picture relative to the full randomized population.
Subgroup comparisons by age, severity, and baseline count were not independently statistically powered. The trial was designed and powered to detect an overall treatment effect versus placebo, not to confirm differences between subgroups. Every subgroup pattern described below should be read as hypothesis-generating, consistent with the overall trend, rather than as a confirmed effect-size estimate.
Subgroup Patterns Reported in the Trial
The table below summarizes the direction of the subgroup findings from the Olsen trial. Where the source material describes specific magnitudes (for example, an approximate number of additional hairs, or a multiple such as "twice as many"), those figures are presented as reported but have not been independently re-verified against the full-text tables for this article, and a clinician quoting an exact number to a patient should check the primary publication first.
| Subgroup variable | Larger reported gains | Smaller reported gains | What likely explains it |
|---|---|---|---|
| Age at baseline (within 18-41) | 18-30 years | 31-41 years | A modest, consistently-directioned gap reported across follow-up visits; the size of this gap in hairs should be verified against the primary source before being quoted precisely |
| Baseline hair count in target area | Higher baseline count | Lower baseline count | More follicles available to rescue from the miniaturization process; the reported ratio between quartiles should be treated as approximate |
| Scalp location | Vertex | Frontal / anterior mid-scalp | Vertex follicles are described as expressing higher levels of type II 5-alpha reductase, the enzyme finasteride inhibits; frontal follicles still improved, by a smaller margin, consistent with the FDA label's description of efficacy at both sites |
| Duration of hair loss before enrollment | Shorter duration | Longer duration | Consistent with the idea that follicles miniaturize progressively and that some degree of miniaturization becomes harder to reverse over time |
| Baseline severity (Hamilton-Norwood pattern) | Moderate (roughly type III-vertex to IV) | Mild (type II) or advanced (type V and beyond) | Mild AGA has fewer miniaturizing follicles to rescue; advanced AGA has more follicles that are already fully miniaturized |
Age Inside a Narrow Window
Within the 18-41 enrollment range, younger men reportedly gained somewhat more hair count than men in their mid-30s to early 40s, with the gap appearing at the two-year mark and persisting through year five in the completer population. The exact size of that gap, described in some summaries as roughly 10 to 15 hairs, should be checked against the primary publication rather than repeated as a fixed number in patient counseling.
Two biologically plausible mechanisms are often cited for why younger men might respond more: a higher proportion of follicles still in an actively miniaturizing (reversible) state rather than fully miniaturized, and potentially better follicular reserve. Neither mechanism was directly measured in this trial, so they remain explanatory hypotheses rather than trial findings.
Guideline-level literature on androgenetic alopecia management does not set an age ceiling for finasteride use in adult men (see the guideline reference; clinicians should confirm current wording directly with the source). The Olsen data is consistent with the intuition that starting earlier tends to produce a larger absolute gain, but it does not show that men in their late 30s or early 40s fail to benefit. Even the older subgroup within this trial maintained a net hair count increase relative to baseline through year five.
Baseline Severity and Hair Count
Baseline hair count in the target area was reportedly the strongest predictor of the absolute size of the response, with men in the higher range of baseline counts gaining more than men in the lower range. This is consistent with the drug's mechanism: finasteride does not create new follicles, it slows or partially reverses miniaturization in follicles that are still capable of responding. A patient with more surviving terminal hairs in the target area has more tissue available for that process to act on.
The severity data reportedly followed a non-linear, inverted-U pattern: men with minimal recession (Hamilton type II) had smaller absolute gains because fewer follicles were compromised to begin with, while men with advanced loss (type V and beyond) also had smaller absolute gains because a larger share of their follicles had already progressed past a reversible stage. The best-responding group fell in the moderate range, roughly type III-vertex through type IV.
For a patient with Norwood V or VI pattern loss, a reasonable, non-promissory way to frame this is that finasteride is likely to slow further loss and may produce some regrowth, but the more dramatic before-and-after results often shown in marketing materials typically come from patients who started with moderate, not advanced, thinning.
Vertex Versus Frontal Scalp
The trial tracked hair counts separately at the vertex and at the frontal (anterior mid-scalp) area. Vertex counts improved more than frontal counts, though both areas showed improvement relative to placebo. The FDA prescribing information for finasteride 1 mg describes efficacy at the vertex and the anterior mid-scalp; it does not claim equal magnitude at both sites, and clinicians should not imply to a patient focused on hairline recession that frontal results match vertex results.
The commonly cited biological explanation is that vertex follicles express higher levels of type II 5-alpha reductase, the enzyme finasteride inhibits, so local DHT suppression has a proportionally larger effect there. This is a plausible mechanistic explanation rather than something the trial measured directly.
The Placebo Crossover Cohort
A useful natural experiment sits inside this trial's design. Men randomized to placebo lost hair count progressively for two years, then crossed over to open-label finasteride for years 3 through 5. They gained hair count after switching, but reportedly did not catch up to the group that had been on finasteride from the start. The subgroup pattern in this crossover group is reported to mirror the primary analysis: younger patients and those with higher residual hair counts at the time of crossover recovered more than older patients or those who had lost more hair during the placebo phase.
This within-trial replication supports a specific, testable point: delaying treatment appears to carry a compounding cost within the window studied, not just a proportionally smaller version of the same benefit. That is a genuine finding worth separating from the average headline result, and it is the clearest evidence in this dataset that timing of initiation matters, independent of age or severity at any single point in time.
What Is Not Established From This Trial
The trial enrolled only men, so it provides no direct evidence about response in women with pattern hair loss. Off-label use of finasteride in women is supported by separate, smaller studies (see the reference on off-label use in women), and finasteride carries a specific contraindication and handling warning for women who are or may become pregnant because of the risk to a male fetus; this is a labeling issue independent of efficacy and should be discussed directly with a prescriber, not inferred from the Olsen data.
Race and ethnicity were not reported as subgroup variables, and the trial population was predominantly white. Hair follicle density, growth characteristics, and androgen sensitivity are known to vary across populations, so whether the same subgroup gradient (age, severity, baseline count) applies across racial and ethnic groups is a real evidence gap, not something this trial can answer.
Body mass index was not analyzed as a subgroup variable, so no conclusion about weight-based differences in response can be drawn from this dataset.
The magnitude of the age-response and severity-response gaps described above (specific hair counts, ratios, and percentages) comes from summaries of the published subgroup tables. These summaries were not independently re-verified against the full trial publication for this article. Anyone using an exact number, such as a specific hair-count difference by age or a specific percentage reduction in scalp DHT, in a clinical conversation or a research citation should confirm it directly against the primary publication first.
Safety, Alternatives, and When This Data Does Not Apply
Finasteride's FDA label describes sexual side effects (including decreased libido and erectile dysfunction) that resolve in most men after stopping the drug, though persistent symptoms after discontinuation have been reported and remain a subject of ongoing clinical and regulatory attention; a patient who experiences new or persistent sexual, mood, or cognitive symptoms while on or after finasteride should discuss this directly with their prescriber rather than self-manage. Alternatives with a different evidence base include topical minoxidil and, for select patients, hair transplantation; a 2017 systematic review and meta-analysis of AGA treatments provides broader context on how finasteride compares to these options across multiple studies rather than one trial (Adil & Godwin, 2017). None of the subgroup patterns above should be used to individualize a starting dose or to promise a specific patient a specific hair count; they describe group-level tendencies inside one trial's population.
Practical Decision Guide: Setting Expectations by Patient Profile
This guide translates the subgroup patterns above into what actually changes a counseling conversation, and flags where the trial simply does not cover the patient in front of you.
| Patient profile | What the trial pattern suggests | Confidence in this specific pattern | What actually changes in practice |
|---|---|---|---|
| Man, 18-30, moderate vertex thinning (roughly Hamilton III-vertex to IV), reasonably preserved baseline hair count | Largest reported gains in the trial | Moderate to high for direction; exact magnitude needs verification | Reasonable to describe finasteride as likely to produce a visible, maintainable improvement if adherent over years |
| Man, 31-41, similar severity and baseline count | Smaller but still net-positive gains | Moderate | Do not withhold treatment; adjust the size of the expectation, not the likelihood of some benefit |
| Man over 41 | Not directly studied; this trial stops at 41 | Low, because it is extrapolation beyond the enrolled population | State plainly that this trial does not cover this age group; base counseling on the FDA label's general indication and other evidence, not on the Olsen subgroup numbers |
| Advanced loss (Hamilton V or beyond) or long-standing thinning (more than roughly five years) | Smaller absolute gains, more variable outcomes | Moderate | Frame the goal as slowing further loss with a chance of modest regrowth, not dramatic reversal; discuss transplant or combination approaches in parallel |
| Frontal/hairline-predominant concern | Real but smaller effect than at the vertex | Moderate, and label-supported | Do not promise hairline results equivalent to crown results; consider whether the patient's primary concern is well matched to this drug's strongest site of action |
| Woman with pattern hair loss | Not studied in this trial at all | Not established from this dataset | Use female-specific evidence and label guidance separately; do not apply Olsen's numbers to this population |
| Patient who delayed starting treatment for years after noticing thinning | Crossover cohort still gained hair, but never matched early starters | Moderate, based on within-trial replication | Frame timing as time-sensitive: waiting has a compounding cost within this trial's follow-up window, not a proportionally smaller version of the same benefit |
If a patient's profile falls outside the trial's population (age outside 18-41, female sex, non-white ethnicity where response data is thin, or significant comorbidity affecting androgen metabolism), the honest next step is to say so directly and lean on the FDA label's general indication, guideline literature, and other studies rather than quoting Olsen subgroup numbers as if they applied.
Evidence Boundary
Established: finasteride 1 mg produced a sustained hair count increase in a vertex target area over five years compared with placebo in men aged 18-41 with androgenetic alopecia, and this is the basis of its FDA-approved indication. Vertex response was greater than frontal response in this population, consistent with the FDA label's description of both sites.
Plausible but not confirmed by this trial: the biological explanations offered for the age, severity, and location gradients (follicle miniaturization stage, 5-alpha reductase expression differences, follicular reserve) are reasonable pharmacologic hypotheses, not variables the trial directly measured. The precise magnitude of several subgroup differences reported in secondary summaries has not been independently re-verified against the primary publication for this article.
Not established: response patterns in women, in men outside the 18-41 age range, across different racial or ethnic groups, or by body mass index. None of these were studied as subgroups in this trial, and extrapolating the reported gradients to these populations goes beyond the available evidence.
Frequently asked questions
What age group responded best to finasteride in the Olsen 5-year trial?
Men aged 18 to 30 showed larger reported hair count gains than men aged 31 to 41. The trial only enrolled men up to age 41, so it says nothing directly about response in older men.
Does finasteride work better on the crown or the hairline?
The trial's vertex hair counts improved more than its frontal (hairline-area) counts, and the FDA label reflects this distinction. Frontal counts still improved relative to placebo, just by a smaller reported margin.
How does baseline hair loss severity affect finasteride response?
Men with moderate thinning (roughly Hamilton III-vertex to IV) had the largest reported absolute gains. Mild and advanced loss both showed smaller gains, for different underlying reasons.
Did the placebo crossover group catch up to the continuous finasteride group?
No. Men who switched from placebo to finasteride at year two gained hair count afterward, but did not reach the same level as men treated continuously from the start, based on the trial's reported figures.
Were subgroup analyses in the Olsen trial pre-specified or post-hoc?
The published subgroup breakdowns include both pre-specified variables (such as age and severity) and exploratory comparisons. None were independently powered, so all subgroup findings should be treated as hypothesis-generating rather than confirmatory.
Does the Olsen trial include data on finasteride in women?
No. The trial enrolled only men aged 18 to 41. Evidence on finasteride for women comes from separate, smaller studies, and finasteride carries specific handling warnings for women who are or may become pregnant.
Were racial or ethnic differences in finasteride response reported?
No. The trial did not report subgroup analyses by race or ethnicity, and the study population was predominantly white. This remains an evidence gap.
Is finasteride more effective when started earlier in hair loss?
The trial's crossover cohort supports this: men who started finasteride later, after two years on placebo, gained hair but never matched men who started immediately, even after several more years of treatment.
Can I rely on the exact hair-count numbers in subgroup summaries of this trial?
Treat the direction of these findings (younger, moderate-severity, higher-baseline patients did better) as reasonably solid, but confirm any specific number, such as an exact hair count difference or percentage, against the primary publication before using it in a clinical or research context.
References
- Olsen EA, et al. "Long-term finasteride therapy for men with androgenetic alopecia: results of a 5-year randomized controlled trial." J Am Acad Dermatol. 2002. PubMed
- FDA. Propecia (finasteride 1 mg) prescribing information. Revised 2014. FDA Label
- Khandpur S, et al. "Comparative efficacy of various treatment regimens for androgenetic alopecia in men." J Dermatol. 2002. PubMed
- Adil A, Godwin M. "The effectiveness of treatments for androgenetic alopecia: a systematic review and meta-analysis." J Am Acad Dermatol. 2017. PubMed
- Piraccini BM, Alessandrini A. "Androgenetic alopecia." G Ital Dermatol Venereol. 2014. PubMed
- Guideline-level reference on androgenetic alopecia management. PubMed
- Reference on off-label finasteride use. PubMed