Divergent progression pathways in male androgenetic alopecia and female pattern hair loss: Trichoscopic perspectives.

Background: Despite similarities in progressive miniaturization of hair follicles and transition of terminal hairs to vellus hairs, insufficient trichoscopic comparisons between male androgenetic alopecia (MAGA) and female pattern hair loss (FPHL) hinder our ability to select effective treatments. Aim: Our study aimed to explore gender‐specific trichoscopic characteristics of MAGA and FPHL, while formulating hypotheses regarding the progression of these conditions across clinical stages. Methods: We classified 126 male MAGA subjects using Hamilton‐Norwood Classification and 57 FPHL subjects using adopted Sinclair Scale. Subsequently, we analyzed nine trichoscopic factors divided into three categories: hair‐diameter related, hair‐number per follicular unit related, and hair density related factors. Results: Of the nine quantitative trichoscopic factors, hair‐diameter and hair‐number per follicular unit showed strong correlations with clinical stages in both genders. Hair density, a common trichoscopic factor for hair loss evaluation, weakly correlated with clinical stages in FPHL, but not at all in MAGA. In addition, MAGA was characterized by a progressive reduction in hair‐diameter, followed by a reduction in hair‐number per follicular unit. FPHL, on the contrary, showed the opposite progression. Conclusions: Trichoscopic factors vary with disease severity in a gender‐specific manner. Our research highlights that MAGA and FPHL involve two distinct streams: hair‐diameter decreasing by hair follicle miniaturization (Stream 1), and hair‐number per follicular unit decreasing by hair follicle tri‐lineage niche dysfunction (Stream 2). MAGA typically starts from Stream 1 to Stream 2, while FPHL starts from Stream 2. These diverse progression pathways underscore the importance of personalized treatment approaches.

Keywords: androgenetic alopecia; dermoscopy; hair follicles; hair loss; trichoscopy

Androgenetic alopecia (AGA), also known as pattern hair loss, is a common hair loss disorder that affects both men and women. It is characterized by the miniaturization of genetically susceptible hair follicles (HFs), the evolution of terminal hairs (thick, long hairs) to vellus hairs (thin, short hairs), the alternation of hair cycle dynamics, and age‐related progressive disorders.[[1]] Although, AGA appears to be a similar condition in both genders, there are some clear distinctions between AGA in men and women.

Male AGA (MAGA) typically shows more miniaturized HF and vellus hairs, and is more likely affected on the vertex and temples of the scalp.[3] On the contrary, female patients are more likely to experience central, frontal, and parietal diffuse hair loss.[4] In addition, the role of androgens in this type of hair loss is less clear. Because of this, female pattern hair loss (FPHL) has emerged as the preferred term over female AGA.[[1], [4]]

Genetic factors are not fully understood, and they may vary in men and women. Two major genetic risk loci (on the X‐chromosome AR/EDA2R locus and chromosome 20p11 locus) were identified in MAGA, while, the corresponding genetic factors do not account for the phenotype in FPHL.[5] Some studies showed associations between FPHL and specific SNPs such as CYP19A1 and ESR2 genes.[4] However, the role of genetic factors in FPHL is still uncertain. FPHL is multifactorial and probably both androgen‐dependent and androgen‐independent mechanisms contribute to the phenotype.

Dihydrotestosterone (DHT) which binds more tightly to androgen receptors than other androgens, is the main target for treating MAGA. DHT‐inhibitors have proven effective in MAGA treatment, although their results in treating FPHL have been less consistent.[4] In fact, FPHL can sometimes develop independently of androgens. While these differences are known, the exact causes of MAGA and FPHL remain unclear. Hence, we are exploring potential differences in trichoscopic images between the genders as a clue.

Trichoscopy is a branch of dermoscopy. In order to collect and analyze image data, it magnifies skin lesions while eliminating reflected light on the scalp surface. Trichoscopic findings in MAGA have not been distinguished from that in FPHL, and commonly measured or observed trichoscopic features in both genders include: a decrease of hair‐density (hair numbers in a certain area), hair‐diameter diversity (diversity of >20% is diagnostic), an increase in number of short vellus hair (<30 μm), an increase in single‐hair follicular units (FU), yellow dots, pinpoint white dots, peripilar sign, and scalp pigmentation (honeycombed like pattern).[[6], [8]] Among these, hair‐diameter diversity,[6] an increase in the number of short vellus hair,[8] and an increase in the number of single‐hair FUs[[7], [9]] are reported to be quantitative trichoscopic factors (QTF) correlated with clinical progression.

Clinical progression is commonly evaluated (Hamilton‐Norwood Classification for MAGA and Sinclair scales for FPHL) by global images which show clinical patterns and the extent of the disease. However, these clinical classifications do not include quantitative indicators of hair loss, so they do not accurately represent the progression of hair loss. Hence the coefficient correlation between clinical classifications and quantitative trichoscopic factors including hair diameter and hair numbers in a FU were analyzed in a previous study[9] and in this study, as well.

Hair‐diameter diversity is defined as an increase of vellus hairs with a decrease of terminal hairs. In other words, hair‐diameter diversity and increased number of vellus hair are both factors that are associated with hair‐diameter. On the contrary, single‐hair‐per follicular unit (1FU) is associated with hair‐number per FU. For a normal scalp, the FU usually contains 2–4 terminal hairs,[[1], [8]] and the hair‐number per FU decreases in relation to AGA severity.[[7], [9]] However, exactly how AGA pathogenesis affects these trichoscopic factors remains unknown.

In summary, although clear differences of clinical presentation, treatment responses for DHT‐inhibitors, and genetic factors have been documented, no clear distinctions of trichoscopic features between MAGA and FPHL have been discussed. Hence, in this report, we conducted a quantitative trichoscopic analysis to identify clues that might explain differences between MAGA and FPHL from the trichoscopic perspective. Based on our findings, we hypothesized underlying mechanisms responsible for the different progression pathways of MAGA and FPHL.

This is a retrospective observational two‐year study that included 126 Japanese MAGA subjects with vertex shedding and 57 Japanese FPHL subjects with frontal to parietal shedding, who visited the dermatology department of our clinic from March 2021 to April 2023. The subjects were screened by dermatologists and categorized by Hamilton‐Norwood Classification (H‐N C)[3] which is commonly used in clinical dermatology practice as a tool to evaluate the extent of hair loss in MAGA. In FPHL, we modified the Sinclair scale.[10] Following the original Sinclair scale (Stage 1–5), a visual evaluation was done of central hair parting line width, by dividing Sinclair scales 1 to 3 into five levels (I–V), which we called the adopted Sinclair scale (a‐Sinclair). This was done because no subjects in our study were classified as Sinclair 4 and 5. All subjects had a physical examination and blood test to exclude severe systemic diseases or hair loss diseases other than AGA. Those receiving DHT inhibitors (oral finasteride or duatasteride) or other AGA treatment for longer than 3 months were excluded.

All subjects agreed to participate in this study and the use of images for publication, and provided written informed consent. The study followed the principles outlined in the Declaration of Helsinki.

The benchmark of the measurement area was a 5 × 5 mm grid set in hair‐whorl centering areas in MAGA since vertex AGA starts from hair‐whorl areas and individual whorls never change.[11] In FPHL the designated measurement point was situated along the central axis spanning the frontal to parietal regions of the scalp, exactly 15 cm above the midpoint of the glabella. Global images of each subject and close‐contact photographs of each benchmark were taken using a dermocamera (Casio DZ‐D100, Japan). Hair details were observed, measured, and counted by image management software (D'z image Viewer, Casio, Japan).

We analyzed nine quantitative trichoscopic factors (9QTF).

• Hair density: Total hair count per 5 × 5 mm grid (THC)
• Hair‐diameter related factors: Maximum hair diameters (Max D), terminal hair count rate (TH%), vellus hair count rate (VH%), and indeterminate hair count rate (IH%)
• Max D: For predicting the ability to produce thick hair, Max D was calculated as the mean diameter of the three thickest hairs in the area.
• TH%: The definition of terminal hair (TH) is diameter > 60 μm, length >2 mm.
• VH%: The definition of vellus hair (VH) is diameter < 30 μm.
• IH%: The definition of indeterminate hair (IH) is diameter between VH and TH
• Hair‐number per FU related factors: Single‐hair per FU rate (1FU%), double‐hairs per FU rate (2FU%), triple‐hairs per FU rate (3FU%), and multi‐hairs (more than two hairs) per FU rate (MFU%). Hairs are considered to be in the same FU if they are closer than 20 μm apart.

The statistical significance of the differences of 9QTF (described above) between each clinical stage were assessed using Spearman's rank correlation coefficient. Subsequently, selected trichoscopic values were compared within stages and stage groups next to each other, and within younger, middle, and elder age groups, using the Mann–Whitney U test. All statistical values were deemed statistically significant at a p‐value of <0.05.

For men, the mean age for this study was 53.0 years (range: 28–86 years) with the highest distribution being 40–59 years (58.7%), and for women 51.9 years (range: 22–81 years) with the highest distribution being 40–59 years (57.9%). The male subjects were mostly in H‐N C III (30.2%) and female subjects in a‐Sinclair III (35.1%). Detailed information for each clinical stage is in Table 1A,B.

1 TABLE (A) Characteristics of 126 MAGA subjects. (B) Characteristics of 57 FPHL subjects.

1 Abbreviations: a‐Sinclair, adopted Sinclair scale; FPHL, female patter hair loss; H‐N C, Hamilton‐Norwood Classification; MAGA, male androgenetic alopecia; N, number of subjects; NS, non‐severe stages; S, severe stages; SD, standard deviation.

All global and trichoscopic images were ranked by severity (Figure 1A, B), and Spearman's coefficient correlation between clinical stages and 9QTF were calculated (Table 2A, B). 9QTF were divided into three categories: hair‐diameter related (green), hair‐number per FU related (blue), and hair density related (yellow). Correlation coefficients (rs‐values) greater than |0.2| with statistical significance (p < 0.05) are denoted by colors. VH%, TH%, and 1FU% had strong correlations with the clinical stages in both genders. Because of this, hair‐diameter related QTF (VH%, IH%, and TH%) and hair‐number per FU related QTF (1FU%, 2FU%, and 3FU%) are shown under the global and trichoscopic images for reference (Figure 1A, B). Max D had a stronger correlation with the clinical stages in MAGA than in FPHL, while 3FU% showed a stronger correlation with the clinical stages in FPHL than in MAGA. This suggests that Max D may be a more useful marker of clinical progression in males, while 3FU% may be a more useful marker of clinical progression in females.

2 TABLE (A) Correlation between H‐N C and quantitative trichoscopic factors in MAGA. (B) Correlation between a‐Sinclair and quantitative trichoscopic factors in FPHL.

• 2 Note: Green: hair diameter related tricoscopic factors; Blue: hair number per FU related tricoscopic factors; Yellow: hair density‐related tricoscopic factors. Correlation coefficients (rs‐values) greater than |0.2| with statistical significance (p < 0.05) are denoted by colors.
• 3 Abbreviations: 1FU%, 2FU%, 3FU%, MFU%, single‐hair, double‐hair, triple‐hair, multiple‐hair per follicular unit rate; a‐Sinclair, adopted Sinclair scale; FPHL, female pattern hair loss; FU, follicular unit; H‐N C, Hamilton‐Norwood Classification; IH%, indeterminate hair rate; MAGA, male androgenetic alopecia; Max D, maximum hair diameter; TH%, terminal hair rate; THC, total hair count; VH%, vellus hair rate.

Furthermore, we observed that the significant correlation coefficients (rs‐values) differed between the non‐severe and severe stages of the disease. Among males, VH% and IH% exhibited correlations with clinical stages in non‐severe stages (H‐N C: I‐III) while in severe stages (H‐N C IV‐VII), Max D, VH%, and TH% showed correlations. In females, 1FU% and 3FU% showed correlations with clinical stages in non‐severe stages (a‐Sinclair I, II), whereas in severe stages (a‐Sinclair III‐V), VH%, TH%, 1FU%, and MFU% demonstrated correlations. This suggests the presence of a transition point for rs‐values somewhere during disease progression. Subsequently, we conducted a detailed analysis of the differences between each stage to better understand this finding.

We performed Mann–Whitney U tests to compare adjacent stages and stage groups, focusing on presenting only those combinations that exhibited significant differences (Table 3A, B). These results helped us pinpoint the clinical stage combinations with the most pronounced differences in disease progression. We categorized the 9QTF into three groups and highlighted cells with statistical significance (p < 0.05). To enhance clarity, we extracted the color‐coded QTF in the progression of clinical stages, which are presented in Figure 2. In the case of MAGA, changes were initially observed in hair‐diameter related QTF (Max D, VH%, IH%, and TH%), followed by hair‐number per FU related QTF (1, 2, 3, MFU%). Conversely, in FPHL, QTF associated with hair‐number per FU (1FU%) exhibited a decrease from the onset, with subsequent changes observed in both hair‐diameter related and hair‐number per FU related QTF.

3 TABLE (A) Difference between each clinical stage of MAGA. (B) Difference between each clinical stage of FPHL.

• 4 Note: Green: hair diameter related tricoscopic factors; Blue: hair number per FU related tricoscopic factors; Yellow: hair density‐related tricoscopic factors. Cells with statistical significance (p < 0.05) are denoted by colors.
• 5 Abbreviations: 1FU%, 2FU%, 3FU%, MFU%, single‐hair, double‐hair, triple‐hair, multiple‐hair per follicular unit rate; 9QTF, nine quantitative trichoscopic factors; a‐Sinclair, adopted Sinclair scale; FPHL, female pattern hair loss; FU, follicular unit; H‐N C, Hamilton‐Norwood Classification; IH%, indeterminate hair rate; MAGA, male androgenetic alopecia; Max D, maximum hair diameter; TH%, terminal hair rate; THC, total hair count; VH%, vellus hair rate.

Since AGA is considered to be an age‐related progressive disease, examining 9QTF differences between younger subjects (20–39 years old) and elder subjects (60–89 years old) could provide insights into underlying mechanisms. Although we found no statistically significant differences between the two age groups, it is worth noting that hair‐diameter related QTF (colored green in Figure 3) showed a trend of inverse correlation with age (black arrows) in both MAGA and FPHL. The hair‐number per FU were similar in the age groups in MAGA, but there was a trend of age‐dependent worsening in FPHL (colored blue, red arrows in Figure 3). It suggests that early onset hair‐diameter deterioration may be caused by age‐independent factors, such as androgens and genetics, while hair‐number per FU decreases with age.

The 9QTF exhibited varying correlations with clinical progression between men and women. In general, MAGA patients showed early stages hair‐diameter deterioration, followed by a reduction in both hair‐diameter and hair‐number per FU. By contrast, the pattern was reversed in FPHL, where a decrease of hair‐number per FU was observed initially (Figure 4). Furthermore, hair‐diameter appears to be more age‐independent, likely due to the influence of androgens and genetic factors.

AGA causes progressive hair loss in both genders. Our study found that 9QTF deteriorate along with clinical progression, especially those linked to hair‐diameter (Max D, VH%, and TH%) and hair‐number per FU (1FU%, 2FU%, 3FU%, and MFU%). Notably, hair density (THC) had weak correlation in FPHL and none in MAGA. When we categorize clinical stages into non‐severe and severe stages, distinct 9QTF correlations appeared by gender. In men, hair‐diameter related QTF were significant at all stages. In women, QTF related to hair‐number per FU showed significance in non‐severe stages. However, in severe stages, all three categories, including hair‐diameter, hair‐number per FU, and hair‐density related QTF exhibited significance.

Thus, hair‐diameter related QTF are likely more important for monitoring disease progression in men, while hair‐number per FU related QTF are likely more important for monitoring disease progression in women.

Based on these results and the existing literature, we hypothesized a relationship between the disease progression and QTF that involves two distinct streams: Stream 1: a reduction in hair‐diameter caused by HF miniaturization, and Stream 2: a decrease in hair‐number per FU, which is related to dysfunction in the HF tri‐lineage system.

Stream 1, which is caused by hair‐diameter reduction due to HF miniaturization, is primarily influenced by two pathophysiological factors: shortened anagen phase[12] and dermal papilla (DP) volume loss.[13] Anagen phase significantly determines both hair‐diameter and length.[2] Shorter anagen results in thinner and shorter hair. Androgens act on epithelial cells (matrix cells) of genetically susceptible HFs in androgen‐dependent areas, suppressing hair growth, and promoting HF miniaturization by altering hair growth cycle dynamics.[5] DP volume also influences hair‐diameter. A smaller DP results in smaller HFs, as the number of dermal papilla cells (DPC) and extracellular matrix volume within the DP are closely associated with hair‐diameter.[4] DPCs play a crucial role in inducing and maintaining epithelial cell growth, and mediate the growth‐stimulating signals of androgens by releasing growth factors in a paracrine fashion on other follicle cells. Androgens act on epithelial cells via mesenchyme‐derived DP.[5]

Stream 2, which is related to hair‐number per FU, involves an expansion of the gap between individual HFs within each FU. It results in an increase in 1FU% and a decrease in MFU%. HFs in the scalp are grouped, forming well defined anatomical structures known as FUs.[[1], [8]] FUs are composed of 2–4 HFs. Each FU shares a single arrector pili muscle (APM) structure[1] that surrounds all the HFs, attaches to the bulge region where HFSCs are located.[14] In the case of MAGA and FPHL, APM degeneration, significant volume loss of the muscle, extensive fat infiltration around the degenerated APM, loss of contact between miniaturized HFs and APM, and the widening gap between HFs were observed.[[15]] Combining these findings and our results, the increase in 1FU% indicates clinical progression, possibly attributed to the weakened connection of the APM with HFs.

Research indicates that the APM‐sympathetic nerve (SN)‐HFSC tri‐lineage system plays a role in hair growth regulation. APM acts as a stable anchor for sympathetic innervation to HFs.[17] Without sympathetic signaling, HFSCs enter a deep quiescence state, slowing down their growth process.[18] In severe MAGA and some FPHL, APM degenerates, causing the loss of SN, making HFSCs more difficult to activate.[17] This disruption of the niche may prolong the catagen (regression) phase, leading to HF miniaturization and subsequent reduction of hair‐diameter. In brief, dysfunction of APM‐SN‐HFSC tri‐lineage system is associated with a decrease in both hair‐diameter (Stream 1) and hair‐number per FU (Stream 2).

How do Stream 1 and Stream 2 relate to the gender‐specific progression of AGA? We would like to discuss the relationship between the known differences (e.g., genetic factors and local hormonal levels), known common factors (e.g., cellular senescence and oxidative stress), and the results of this study. AGA is a polygenic disorder linked to many factors, including local androgen levels, cellular senescence, oxidative stress, and other unknown factors.

Genetic factors are considered a primary distinction between MAGA and FPHL with over 250 independent genetic loci identified in MAGA. However, accurate predictions for individual MAGA are still a challenge.[19] By contrast, the role of genetic factors in FPHL is less clear, with no distinct overlap of susceptibility loci between MAGA and FPHL.[20] Single nucleotide polymorphisms in genes like CYP19A1 and ESR2, involved in hormonal pathways affecting hair growth, have been investigated in FPHL. However, genetic factors may only be a part of AGA etiology.

Androgens also contribute to gender‐based differences. MAGA is strongly linked to local levels of testosterone,[2] and DHT is higher in the balding scalp than in the non‐balding scalp.[1] While most MAGA patients have normal circulating androgen levels, higher rates of testosterone and DHT production occur, and oral DHT‐inhibitors have proven effective.[2] FPHL can manifest varying degrees of a combination of androgen‐dependent and androgen‐independent patterns, making DHT‐inhibitors less consistently effective compared to MAGA.[4]

Cellular senescence is a natural aging process marked by irreversible cell cycle arrest triggered by various stressors, including DNA damage, telomere shortening, and oxidative stress. This process accelerates over time and often results from oxidative stress, which contributes to aging and age‐related disease.[21] AGA is linked to a number of factors that increase oxidative stress, such as metabolic syndrome, alcohol, smoking, and UV irradiation, further contributing to age‐related diseases.

As individuals age, hair loss increases, accompanied by structural and biological changes in HF and its surrounding environment.[14] These changes include shorter HF that failed to reach the adipose layer, increased level of matrix metalloproteinase I in the dermal papilla (DP) and dermal sheath (DS), and reduced expression of alpha smooth muscle actin mRNA in DS cells.[14] Shin et al. have demonstrated senescent‐like characteristics in hair follicle dermal stem cells (HFDSCs), and identified senescence‐associated secretory phenotypes in DPCs.[22] In healthy HFs, HFDSCs undergo self‐renewal and repopulate both DS and DP with new cells. However, when HFDSCs become functionally compromised, DPCs replace DSCs, resulting in smaller DP, subsequently smaller HF,[12] eventually results in the replacement of terminal hairs with vellus hairs.

Cellular health depends on mitochondrial energy management. In MAGA, DPCs suffer from high oxidative stress levels,[23] potentially leading to their senescence and apoptosis.[21] Healthy DPCs resist apoptosis throughout the hair cycle, ensuring hair survival.[23] However, senescent DPCs may be susceptible to apoptosis, reducing their volume and the eventual replacement of terminal hairs with vellus hairs. Notably, DPCs in MAGA patients demonstrated premature senescence in response to environmental stress.[5]

When considering the comparison of younger and elder age groups, the appearance of increasing hair‐diameter with age suggests that the impact of androgens and genetic factors is more significant than the effect of aging in MAGA. This leads to a notable reduction in hair‐diameter primarily caused by HF miniaturization (Stream 1). By contrast, in FPHL, the influence of androgens and genetic factors varies and is influenced by complex interplay of genetic, hormonal, aging, and other unknown factors, mainly related to HF tri‐lineage dysfunction (Stream 2).

Taken all together, the interplay of androgens, genetic factors, cell senescence, and oxidative stress leads to HF miniaturization, causing reduced hair‐diameter (Stream 1), primarily in MAGA. Cell senescence and oxidative stress induce APM detachment of HFs in a FU, reducing hair‐number per FU (Stream 2) primarily in FPHL (Figure 5). These factors, together with unknown factors, combine differently in individual patients. Hence, the analysis of deteriorating 9QTF allows for a detailed assessment of a patient's current hair condition.

This AGA progression model can help us to understand global, trichoscopic, and histological changes in MAGA and FPHL. However, in order to improve accuracy and reliability of hair diagnostics, we need to continue collecting data and identifying potential ethnic‐specific standards. In the future, these collected data can be used to train machine learning algorithms that can specify patterns in hair analysis results.

Our study revealed that QTF analysis can identify trichoscopic differences between genders and is a robust tool for assessing MAGA and FPHL severity in individual patients. Among 9QTF, hair‐diameter deterioration indicates HF miniaturization, while a worsening of hair‐number per FU indicates HF tri‐lineage system dysfunction. MAGA predominantly advances from hair‐diameter decline to hair‐number per FU decline, while FPHL progresses conversely. However, broader and more diverse studies are essential to confirm these findings and their applicability across all ethnicities. If we fully understand the relationship between trichoscopic changes and treatment options, we can adopt more systemic treatment approaches tailored to each patient.

T.K. conceived of the presented idea. T. K. developed the theory and performed the computations. T. K., C. H. and K. Z. Y. verified the analytical methods. J. T. encouraged T. K. to investigate, and supervised the findings of this work.

The authors appreciate Sakiko Kawano, Asami Ito for data acquisition and management, Lawrence T. Knipfing for editorial support and critical feedback.

This research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors. The authors declare no potential conflict of interest.

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

This study was approved by the Tokyo Midtown Medical Center Ethics Review Board (倫‐2021‐03/February 10, 2021), and registered with the University Hospital Medical Information Network Clinical Trial Registry (UMIN000045897/October 28, 2021). All subjects agreed to participate in this study, use for publication, of images, and provided written informed consent. The study followed the principles outlined in the Declaration of Helsinki. We state that the manuscript has not been published in any form that it is not considered for publication elsewhere.