Men aged 18 to 41 years, with mild to moderately severe vertex MPHL according to a modified Norwood-Hamilton classification scale (2 vertex, 3 vertex, 4, or 5), were enrolled. (See shaded boxes in illustration.) Principal exclusion criteria included significant abnormalities on screening physical examination or laboratory evaluation, surgical correction of scalp hair loss, topical minoxidil use within 1 year, use of drugs with androgenic or antiandrogenic properties, use of finasteride or other 5αR inhibitors, or alopecia due to other causes. Men were instructed not to alter their hairstyle or dye their hair during the studies.

^a Adapted from Norword OT. Hair Transplant Surgery, 1973. Courtesy of Charles C. Thomas, Publisher; Springfield, Ill.

Two initial, 1-year, randomized, double-blind, placebo-controlled studies were initiated, and both were continued as 4 consecutive, 1-year, double-blind, placebo-controlled extension studies. The objectives of the controlled extension studies were to determine the effect of long-term (up to 5 years) treatment with finasteride 1 mg/day, the effect of withdrawing treatment after 1 year, the effect of delaying treatment by 1 year, and the progression of MPHL in men not receiving active treatment. Investigators participated at 33 sites in the United States and 27 sites in 15 countries outside the United States. Institutional review board approval and written informed consent were obtained each year before patients entered into each study.

After a screening procedure, patients entered a 2-week, single-blind, placebo run-in period. All men received Neutrogena T/Gel^® shampoo for standardization purposes and for prophylaxis of seborrheic dermatitis, which could affect scalp hair growth. Patients (N=1,553) were then randomized to finasteride 1 mg/day or placebo (1:1) for 1 year.

Every 3 months, men reported to the study clinic, where they completed a hair growth questionnaire and investigators assessed their scalp hair growth. Every 6 months, scalp hair was photographed to measure hair counts and assess hair growth. Investigators collected reports of adverse events throughout the studies.

Men completing the initial 1-year, placebo-controlled studies were eligible to enroll in 4 consecutive,
1-year, placebo-controlled extension studies. In the first extension studies, men (N=1,215) were randomly assigned (as determined at initial randomization) to treatment with either finasteride 1 mg or placebo (9:1), such that men were randomized to 1 of 4 treatment groups that allocated treatment to them during both the initial 1-year studies (year 1) and the first 1-year extension studies (year 2).

• Fin → Fin
• Fin → Pbo
• Pbo → Fin
• Pbo → Pbo

In the 3 subsequent, 1-year extension studies (years 3 to 5), men continued on the same treatment they had received during the first extension studies (year 2) except for men in the Fin Pbo crossover group; those men were switched back to finasteride (Fin Pbo Fin) during the second extension studies (year 3) and remained on that therapy throughout the subsequent extension studies (years 4 and 5).

The procedures for the first 1-year extension studies (year 2) were similar to those for the initial 1-year studies, except that hair counts were obtained only once, at month 24. In the remaining three 1-year extension studies (years 3 to 5), all efficacy evaluations were performed once yearly.

The primary efficacy end points in the first year of the study were scalp hair count and patient self-assessment. The primary efficacy end point in the extension studies was investigator assessment of hair growth and hair count. Patient self-assessment, assessment by the investigator in the clinic, and assessment of global photographs by a panel of experts were important secondary end points.

Hypothesis testing for hair counts, individual patient self-assessment questions, and investigator and global photographic assessments were performed using analysis of variance (ANOVA).

The primary efficacy analysis population was the intention-to-treat population, which included all patients with at least 1 day of randomized therapy and with both baseline and at least 1 postbaseline efficacy assessment.

For all efficacy analyses, missing data were estimated by carrying data forward from the previous visit. However, no data were carried forward from the baseline evaluation, or between the initial 1-year study and the first 1-year extension study, or between the extension studies.

Hair count was measured in a 1-inch-diameter tattooed circular area centered at the leading edge of the area of thinning vertex scalp hair. Hair-count data were obtained from macrophotographs of the area by converting all visible hairs in the macrophotographs to dot maps. The hairs in the dot map were counted using computer-assisted image analysis, which detected only "non-vellus-like/miniaturized" hairs. Thus, hair counts measured only "cosmetically significant" hairs.

Hair counts were assessed by the difference between the count at each time point vs the baseline count, and mean hair count values for each treatment group were determined using Least Squares Means.

Patients performed self-assessment by using a validated, self-administered hair-growth questionnaire that covered perception of treatment efficacy and patient satisfaction with the appearance of the hair. Each of the 7 questions was assessed separately, and the responses to each question at each time point were taken as assessments of changes from baseline.

The self-assessment consisted of 7 questions^2,a:

1. Since the start of the study, I can see my bald spot getting smaller.
   (Strongly Disagree...Strongly Agree, scored on a 5-point scale)
2. Because of the treatment I have received since the start of the study, the appearance of my hair is:
   (A lot worse...A lot better, scored on a 7-point scale)
3. Since the start of the study, how would you describe the growth of your hair?
   (Greatly decreased...Greatly increased, scored on a 7-point scale)
4. Since the start if the study, how effective do you think the treatment has been in slowing down your hair loss?
   (Not effective at all...Very effective, scored on a 4-point scale)
5. Compared with appearance at the beginning of the study, which statement best describes your satisfaction with the appearance of:
   1. the hairline at the front of your head?
   2. the hair on top of your head?
   3. your hair overall?
   (Very dissatisfied...Very Satisfied, all scored on a 5-point scale)

^a Note that number 5 comprises 3 separate questions.

Investigator assessment used a standardized 7-point rating scale as shown below. A "don't know" option was also available. The investigator was asked to rate the patient using the scale in response to the following question: "As the investigator, how would you subjectively rate the patient's hair at this time point compared to baseline?"

The investigator who performed the initial assessment also performed the assessment at year 5.

Global photographs for each patient were taken at months 0, 6, 12, 18, 24, 36, 48, and 60 before hair clipping for macrophotography. Three expert dermatologists independently evaluated the global photographs for each time point compared with the baseline global photograph. These global photographs were assessed vs baseline global photographs using the same 7-point scale that was used in the investigator assessment.

Each pair of global photographs was randomized to ensure that the dermatologists were blinded to patient, treatment, and study center. Unlike patient self-assessment and investigator assessment, there is no potential for patient interaction bias, leading to markedly decreased placebo effect.