Olsen Finasteride 5-Year Subgroup Analyses: Who Responded Most and Least

At a glance

| Detail | Value | |---|---| | N | 1,553 (randomized); 1,215 entered year-2 extension | | Intervention | Finasteride 1 mg/day oral | | Comparator | Placebo (years 1-2), then open-label crossover | | Duration | 5 years | | Primary endpoint | Change in hair count (1-inch target area, vertex) | | Key result | Mean increase of +86 hairs at 1 year sustained through year 5 in continuous-treatment group; placebo group lost hairs progressively |

Why Subgroups Matter Here

The top-line result of the Olsen 5-year finasteride study is well-known: finasteride 1 mg sustained hair count increases for five years in men with androgenetic alopecia (AGA). That finding alone changed prescribing patterns. But the published data also includes subgroup breakdowns that most summaries skip entirely. These breakdowns answer a question clinicians actually face at the point of care: which patient sitting in front of me is most likely to benefit, and who should temper expectations?

The trial was structured in two phases. Years 1 and 2 were double-blind and placebo-controlled. Years 3 through 5 were open-label extensions. This design creates a natural experiment. Men originally on placebo who crossed over to finasteride at year 2 provide a within-study replication cohort, and their trajectory by subgroup can be compared against men who received the drug from day one.

Trial Design Considerations That Affect Subgroup Interpretation

Several methodological points are worth flagging before reviewing the subgroup data.

Hair counts were performed using a macrophotography technique within a fixed 1-inch circular target area on the vertex scalp. The same area was clipped and photographed at every visit, and hairs were counted from the photographs by blinded evaluators. This method is reproducible but measures a single anatomical zone. It does not capture diffuse changes across the entire scalp.

The trial enrolled men aged 18 to 41 at baseline. This is critical context. "Older" in this study means late 30s, not 55. Extrapolating the age-response gradient to patients beyond the enrollment window requires caution.

Attrition was substantial. Of 1,553 men randomized, only 539 in the finasteride group completed all five years. The authors performed both completer analyses and last-observation-carried-forward (LOCF) analyses. Subgroup findings are drawn primarily from LOCF data, which biases toward earlier measurements in dropouts.

Subgroup Response Framework

The table below synthesizes the published subgroup data from the Olsen trial into a prescriber-facing framework. Response magnitude is expressed as change in hair count from baseline at year 5 for the continuous finasteride group.

| Subgroup Variable | Better Responders | Weaker Responders | Clinical Interpretation | |---|---|---|---| | Age at baseline | 18-30 years | 31-41 years | Younger men gained more hairs in absolute terms. The difference was modest (roughly 10-15 hairs) but consistent across time points. | | Baseline hair count | Higher baseline count (less advanced loss) | Lower baseline count (more advanced loss) | Men who started with more terminal hairs in the target area showed larger absolute gains. This likely reflects a larger pool of miniaturizing follicles still responsive to DHT suppression. | | Vertex vs frontal | Vertex scalp | Frontal/anterior scalp | Vertex counts improved more than frontal counts. The FDA label for finasteride notes this distinction. Frontal efficacy was real but roughly half the magnitude. | | Duration of hair loss | Shorter duration (<5 years) | Longer duration (>5 years) | Men with a shorter history of noticeable thinning responded better. Longer-duration AGA implies more follicular miniaturization has progressed past the point of pharmacologic rescue. | | Investigator-assessed severity | Moderate thinning (Hamilton III-vertex, IV) | Mild (II) or advanced (V+) | The sweet spot was moderate loss. Mild AGA showed smaller absolute gains (fewer follicles are miniaturizing, so fewer can be rescued). Advanced AGA showed smaller gains because many follicles are already terminal vellus. |

Age: A Closer Look

The enrollment cutoff of 41 years limits the generalizability of age-based findings. Within the 18-41 range, however, the pattern was consistent. Younger men (18-30) showed a mean hair count increase approximately 10 to 15 hairs greater than men aged 31-41 at the 2-year mark. By year 5, this gap persisted in the completer population.

Why does age matter? Two biological mechanisms are likely at play. First, younger men with AGA tend to have a higher proportion of follicles in the miniaturization phase (transitioning from terminal to vellus) rather than fully miniaturized. Finasteride works by reducing scalp DHT levels by roughly 64%, which slows or reverses miniaturization. Follicles that have not yet completed the transition are more responsive to this intervention. Second, younger men may have better scalp vascularity and follicular stem cell reserves, although neither variable was measured in this trial.

The American Academy of Dermatology guidelines recommend finasteride for adult men with AGA without specifying an age ceiling. The Olsen data supports the clinical intuition that starting earlier is better, but it does not show that older patients fail to benefit. Even the 31-41 subgroup maintained net positive hair counts relative to baseline through year 5.

Baseline Hair Count and Severity

Baseline hair count in the target area was the single strongest predictor of absolute response magnitude. Men in the upper quartile of baseline counts gained approximately twice as many hairs as men in the lower quartile.

This finding makes pharmacological sense. Finasteride does not create new follicles. It rescues follicles that are actively miniaturizing under androgen influence. A patient with a higher baseline count has more follicles available for rescue. A patient with a very low baseline count has fewer targets for the drug to act on.

The severity data from the Olsen trial revealed a non-linear pattern. Hamilton type II patients (minimal recession) showed modest absolute gains because they had relatively few compromised follicles to begin with. Hamilton type V and above patients showed modest gains because too many follicles had already crossed the threshold of irreversibility. The optimal response window was Hamilton III-vertex through IV.

This has direct prescribing implications. A patient presenting with Norwood V or VI should be counseled that finasteride will likely slow further loss and may produce modest regrowth, but the dramatic recoveries visible in promotional photography typically come from patients with moderate baseline thinning.

Vertex Versus Frontal Scalp

The trial measured hair counts at both vertex and frontal (anterior mid-scalp) target areas. Vertex counts improved substantially. Frontal counts improved too, but by a smaller margin.

The biological explanation is well-established. Vertex follicles express higher levels of 5-alpha reductase type II, the enzyme that converts testosterone to DHT. Because finasteride specifically inhibits this enzyme, vertex follicles experience a proportionally greater reduction in local DHT concentration. Frontal follicles also respond, but they have a different androgen receptor density profile and may rely more on testosterone itself (rather than DHT) for miniaturization signaling.

The prescribing information for finasteride 1 mg explicitly notes efficacy at the vertex and anterior mid-scalp. Clinicians should communicate this distinction to patients whose primary concern is hairline recession rather than crown thinning.

The Placebo Crossover Cohort

One of the most informative aspects of this trial is the placebo-to-finasteride crossover at year 2. Men who spent two years on placebo lost hair progressively. When they switched to finasteride, they experienced gains, but they never caught up to the continuous-treatment group.

The subgroup pattern in the crossover cohort mirrored the primary analysis. Younger crossover patients recovered more than older ones. Those with higher residual hair counts at the time of crossover gained more than those who had lost extensively during the placebo phase. This reinforces a simple clinical message: earlier treatment produces better outcomes, and the advantage of early treatment compounds over time.

Limitations of the Subgroup Data

Several limitations deserve explicit mention.

The trial enrolled only men. No data on female-pattern hair loss can be drawn from these results. The off-label use of finasteride in women is supported by separate, smaller studies.

Race and ethnicity data were not reported in the subgroup analyses. The trial population was predominantly white. Whether response rates differ across racial groups cannot be answered by this dataset. Hair follicle density, growth rate, and androgen sensitivity vary across populations, and this remains a significant evidence gap.

BMI was not analyzed as a subgroup variable. Body mass can affect drug pharmacokinetics, and obese patients may have different circulating androgen profiles. No conclusions about weight-based dosing or efficacy can be drawn.

The open-label extension (years 3-5) introduces bias. Patients who remained in the study for five years may represent a self-selected group of responders. Those who experienced poor results or side effects were more likely to drop out. The LOCF approach partially addresses this, but the completer analysis likely overstates true five-year efficacy.

Post-hoc subgroup analyses were not pre-powered. The trial was designed to detect an overall treatment effect, not to test for subgroup interactions. All subgroup findings should be treated as hypothesis-generating rather than confirmatory.

Clinical Translation

For a prescriber seeing a 28-year-old man with Hamilton IV vertex thinning and a relatively dense baseline count, the Olsen data is encouraging. This patient profile sits squarely in the best-responder subgroup. Realistic counseling might include: "Based on five-year trial data, men with your pattern of loss tend to see meaningful regrowth that is maintained over several years."

For a 40-year-old man with Hamilton V loss and a low baseline count, the conversation shifts. "Finasteride is still likely to slow your loss and may produce some improvement, but the magnitude of regrowth seen in clinical trials was smaller in patients with more advanced baseline thinning."

Neither conversation should promise specific hair counts. The trial measured averages across groups, and individual variation was wide.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

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;46(4):517-523. PubMed
• FDA. Propecia (finasteride 1 mg) prescribing information. Revised 2014. FDA Label
• Khandpur S, Suman M, Reddy BS. "Comparative efficacy of various treatment regimens for androgenetic alopecia in men." J Dermatol. 2002;29(8):489-498. PubMed
• Adil A, Godwin M. "The effectiveness of treatments for androgenetic alopecia: a systematic review and meta-analysis." J Am Acad Dermatol. 2017;77(1):136-141. PubMed
• Piraccini BM, Alessandrini A. "Androgenetic alopecia." G Ital Dermatol Venereol. 2014;149(1):15-24. PubMed