To determine the current state of knowledge and provide evidence-based recommendations that could be valid for all specialists taking care of female pattern hair loss (FPHL), a common form of hair loss in women that is characterized by the reduction of hair density in the central area of the scalp, whereas the frontal hairline is generally well conserved.
An expert task force appointed by the Androgen Excess and PCOS Society, which included specialists from dermatology, endocrinology, and reproductive endocrinology.
Levels of evidence were assessed and graded from A to D. Peer-reviewed studies evaluating FPHL published through December 2017 were reviewed. Criteria for inclusion/exclusion of the published papers were agreed on by at least two reviewers in each area and arbitrated by a third when necessary.
(i) The term “female pattern hair loss” should be used, avoiding the previous terms of alopecia or androgenetic alopecia. (ii) The two typical patterns of hair loss in FPHL are centrifugal expansion in the mid scalp, and a frontal accentuation or Christmas tree pattern. (iii) Isolated FPHL should not be considered a sign of hyperandrogenism when androgen levels are normal. (iv) The assessment of patients with FPHL is primarily clinical. (v) In all patients with FPHL, assessment of a possible androgen excess is mandatory. Measurement of vitamin D, iron, zinc, thyroid hormones, and prolactin are optional but recommended. (vi) Treatment of FPHL should start with minoxidil (5%), adding 5α-reductase inhibitors or antiandrogens when there is severe hair loss or hyperandrogenism.
Hair loss in women is common and may have different causes. A specific and particularly common subtype is characterized by reduction of hair density in the central area of the scalp whereas the frontal hairline is generally well conserved. In the 1950s, Hamilton (1) noted that this form of hair loss shares certain characteristics with male pattern baldness [also called male pattern hair loss (MPHL)] and considered it androgen-dependent. Further studies specifically related MPHL to increased DHT production in skin (2). However, similar evidence for a relationship between DHT overproduction and this pattern of hair loss in women has not been produced.
This phenotype of hair loss in women has been given various names, although in the endocrine and gynecologic literature it is most frequently referred to as androgenetic alopecia, under the assumption that is androgen related (3). Recent guidelines on the diagnosis of polycystic ovary syndrome (PCOS) continue using the more general term “alopecia” and consider the finding diagnostic for the presence of hyperandrogenism (4–7).
Overall, the relationship of hair loss and androgen excess in women is neither consistent nor clear. Most affected women with the frontal–central pattern of hair loss have normal circulating androgen concentrations (8, 9) and an absence of other androgen-related signs or symptoms, such as hirsutism or irregular periods/ovulation (10). The presence of this phenotype of hair loss has been reported also in women with total absence of circulating androgens, deficiency of postpubertal androgenization (11), or lacking an androgen receptor (AR) (12). Because of this, the term female pattern hair loss (FPHL) has become the most common name used in the dermatologic literature for this common condition in women (10, 13–15), without immediately stressing a link to androgen excess. Consequently, we use the term FPHL in the present publication.
As many women with hyperandrogenism also exhibit or complain of scalp hair loss, the Androgen Excess and PCOS (AE-PCOS) Society appointed an expert task force to determine the current state of knowledge concerning the relationship of these two conditions. This publication provides evidence-based recommendations for the evaluation, diagnosis, and treatment of FPHL, as well as its association with hyperandrogenism for use by the many different specialties of health professionals that are called upon to evaluate and treat these women.
AE-PCOS Society task force
The AE-PCOS Society Board appointed an expert task force of varying medical disciplines to determine the current state of knowledge concerning FPHL and its relationship to androgen excess, and to provide evidence-based recommendations. All members of the task force declared no conflicts of interest or financial interests that might interfere with their objectivity and duty.
Each section of this review was prepared by at least two investigators and then reviewed by all members of the task force. Analyzed papers included individual studies, systematic reviews, and abstracts published in the English scientific literature. Criteria for inclusion/exclusion of the published papers relating to each section were agreed on by at least two reviewers in each area and arbitrated by a third when necessary. Levels of evidence were assessed and graded from A to D (16, 17).
The final manuscript was reviewed and approved by the AE-PCOS Society Board of Directors. Institutional Review Board approval was not obtained because the study reviewed publicly available medical literature.
Systematic review of the literature and meta-analysis
Peer-reviewed studies evaluating FPHL published through December 2017 were reviewed. Multiple databases were searched, including MEDLINE, Embase, Cochrane, ERIC, EBSCO, Dissertation Abstracts International, and Current Contents. This review focuses on the epidemiology, pathophysiology, diagnosis, and treatment of the disorder. Some studies were eliminated because the data were either not related to the focus of the systematic review, insufficient for epidemiological analysis, or reported in previous publications. All data sources were analyzed while recognizing positive publication bias.
A meta-analysis was also performed to obtain pooled prevalence estimates on FPHL in subjects with PCOS. The terms of the MEDLINE search for this meta-analysis were: (alopecia OR androgenetic OR female pattern hair loss OR mid scalp OR Ludwig OR frontal accentuation OR Olsen OR male pattern OR Hamilton) AND prevalence AND (PCOS OR polycystic ovary OR polycystic ovarian OR polycystic ovaries). The search was completed by examining the references listed in the articles identified.
After study identification, screening, selection, and inclusion, a quality-effects model was applied considering the heterogeneity of the studies in terms of age, race, and ethnicity of the populations being described. Double arcsine transformations were applied to stabilize the variance (18). Publication bias was assessed by a funnel plot representing the double arcsine transformation of the prevalence against the SE (19). MetaXL 3.0 software was used for the meta-analysis (18).
Results and Discussion
Clinical presentation of FPHL
A shortened anagen phase and miniaturization of hair follicles in the affected scalp are key features of pattern hair loss, a condition with distinct patterns of hair loss in men and women (14). However, hair loss in women with FPHL exhibits several important differences from that of men with MPHL, including patterns and density in affected areas.
There are typically two patterns of hair loss in FPHL: centrifugal expansion in mid scalp with preservation of the frontal hair line (Ludwig pattern) (20) and a frontal accentuation or Christmas tree pattern (Olsen pattern) (21). The latter pattern is distinctive for FPHL and helps to distinguish FPHL from other hair loss conditions that favor the central scalp. Hamilton (1) concurred that women with pattern hair loss generally do not develop the vertex baldness seen in men but may have bitemporal recession; this recession is usually not baldness but rather thinner, shorter, less dense hair. There are, however, women with significant androgenization that may develop typical MPHL (Hamilton pattern) that includes vertex thinning. In Fig. 1, the two main patterns of FPHL are presented and compared with MPHL (22).
Women with FPHL do not experience the degree of baldness that men with MPHL do, largely because the miniaturization process is not as profound and not all hairs are affected equally in the involved areas. Many women with FPHL may also display parietal thinning, but this should not dissuade against making the diagnosis of FPHL.
FPHL may first present in adolescence/early adulthood and/or in the perimenopause or postmenopause age range. Considering the peak age-related incidence of this type of hair loss and the possible role of androgens in at least some of these patients, the following classification of FPHL has been suggested (14):
FPHL: early onset
With androgen excess
Without androgen excess
FPHL: late onset/postmenopausal
With androgen excess
Without androgen excess
However, because androgen secretion decreases during the late reproductive age in normal and hyperandrogenic women (23, 24), patients with androgen excess who develop FPHL (i.e., hyperandrogenic FPHL) usually present during young adulthood. The role of hyperandrogenism in postmenopausal women with FPHL remains unclear, and further studies are needed to understand the hormonal mechanisms involved in FPHL during this period of life.
Hair loss in women presents differently than that in males and three typical patterns (Ludwig, Olsen, and Hamilton) may be distinguished (level of evidence B).
Psychological consequences of FPHL
Hair loss in women has a substantial psychological impact, and several studies have noted that it may result in depression symptoms and reduced quality of life (25, 26). Interestingly, hyperandrogenic disorders, but mainly PCOS, may be associated with mood disorders and depression symptoms per se (27, 28), but to our knowledge no study has specifically addressed the impact of hyperandrogenism on depression symptoms and quality of life in women with FPHL.
Pattern hair loss in women has significant impact on psychological well-being and quality of life (level of evidence C).
Pathophysiology of FPHL
The maintenance of a healthy hair follicle cycle requires well-coordinated progression and timing through the different phases of growth (anagen), regression (catagen), quiescence (telogen), and regeneration. Multiple genes, growth factors, and hormones govern this cycle.
Molecular processes that regulate the different phases have been extensively studied in animals (29–31). The major pathway influencing the normal sequence of events is the activation of the wingless (Wnt) ligand system. Through Wnt and the production of β-catenin, a downstream effector of the Wnt pathway, a regular sequence of events is occasioned. Activation of the Wnt system in the dermal papilla promotes hair follicle growth and induces the switch from telogen to anagen by stimulating the proliferation and differentiation of bulge stem cells. The vitamin D receptor, independent of circulating vitamin D3, also plays an important role in anagen initiation. Data from animal models demonstrate that the vitamin D receptor and β-catenin act together, forming a complex important for anagen activation. The catagen phase, alternatively, seems to be linked to increased expression of bone morphogenetic proteins during the late anagen phase. Many other factors participate in the process of hair regeneration, including several growth factors. Follicular growth factors (Fgf7 and Fgf10) and bone morphogenetic protein inhibitors produced from the dermal papilla activate stem cells to regrow a new follicle during hair regeneration (30, 31).
In both MPHL and FPHL, the same pathologic process appears to be present and is characterized by a reduction in the duration of the anagen phase and an increase in percentage of hairs in the telogen phase. There is an increasing delay in the onset of the next anagen phase, leading to prolongation of the latent phase (also termed kenogen) that follows the telogen phase in pattern hair loss (32). Follicles then also undergo a process of miniaturization that will determine a progressive scalp hair loss (33).
Polygenic susceptibility and increased androgen action in the scalp seem to be the main drivers of scalp hair loss in both MPHL and FPHL. The possible role of chronic low-grade inflammation in the scalp may represent an additional or different driver of hair loss in FPHL.
Genetic susceptibility in FPHL
In MPHL, large genome-wide association studies have been performed and several genetic loci linked to scalp hair loss have been identified (34). However, the candidate genes that have been identified in MPHL [genes encoding the AR, the histone deacetylases 4 and 9, and the Wnt molecule Wnt10A (34)] have not shown any association with FPHL (35). Although not all genetic loci associated with MPHL [>60 (36)] have been studied in FPHL, these data suggest that the genetic components of MPHL and FPHL may be different (35).
Possible role of androgens in FPHL
In MPHL, the role of androgen action on the scalp has been extensively studied demonstrating elevated expression of ARs and/or exaggerated formation of DHT in the scalp (37) with a clear effect of androgens in altering Wnt expression (38). However, data in FPHL are less clear. In fact, many women with FPHL present normal androgen levels (see below), and it has been shown that FPHL may present also in subjects with no ARs (8). Additionally, in the scalp, AR content is higher in the frontal vs occipital scalp in both men and women, higher in men than in women (39), and higher in the balding vs nonbalding scalp (40).
In women with FPHL, enhanced androgen action in the scalp may occur due to alterations in 5α-reductase and enhanced DHT formation locally, or in androgen binding to the AR (39). In fact, the 5α-reductase mRNA levels of all three isozymes (I, II, and III) were observed to be increased in the hair follicles of some, albeit not all, women with hair loss (41). Consistent with an increased formation of DHT in the scalp, elevated circulating levels of 5α-reductase–derived C19 androgen conjugates have been found in women with hyperandrogenic hair loss (42, 43). Legro et al. (43) observed a shift favoring peripheral formation of C19 androgen sulfate conjugates vs glucuronide conjugates.
Recently, using more sensitive mass spectrometry/chromatographic analyses, 11-oxo androgens of adrenal origin have been found to be quantitatively greater in the circulation than levels of the usual androgens assessed (44). These 11-oxo testosterone (T) and DHT metabolites have great affinity for the AR. It is possible that these steroids or their converted conjugates could be involved in the pathogenesis of FPHL, but no data in this regard are yet available.
The role of androgen excess in the FPHL of postmenopausal women is more difficult to establish. Androgen levels start decreasing several years before menopause, and in normal women plasma T decreases by 50% moving from the age of 20 to the age of 40 (45). In hyperandrogenic women, this decline in plasma T is also present but the degree of the decrease appears to be somewhat less (23), which may reflect continued ovarian androgen excess after menopause. Postmenopausal women with FHPL who have had evidence of PCOS may have relatively higher circulating androgens and a low estrogen/androgen ratio, but no data support this hypothesis.
A possible protective effect of estrogens on human hair growth has been suggested, and it is supported by indirect evidence, including increased prevalence of FPHL following menopause, prolongation of anagen during pregnancy (46), and reports of hair loss in women taking tamoxifen or aromatase inhibitors for treatment of breast cancer (47). Additionally, genome-wide association studies have shown an association between polymorphisms of the aromatase gene and FPHL (48). Alternatively, there is no clear effect of estrogen supplementation on human hair growth, and in many animal species estrogens have an inhibitory effect on hair growth (49).
Chronic low-grade inflammation in FPHL
Several data suggest that chronic low-grade inflammation in the scalp may be an important factor favoring hair loss in women (34, 50, 51). This hypothesis is supported by the finding that many inflammatory peptides, including TGF-β1, TNF-α, and IL-1, are involved in the catagen phase (34, 52). Several studies have also found that inflammation is more common in biopsy samples from patients with pattern hair loss compared with those of control subjects (53–55). Additionally, minoxidil downregulates the genetic expression of IL-1 in human keratinocytes (56).
The possible role of chronic inflammation in the pathogenesis of hair loss is reinforced by data indicating that prostaglandin D2 is elevated in bald scalp and inhibits hair growth in explanted human hair follicles (57). Transgenic mice, who have elevated prostaglandin D2 in the skin, develop hair loss and miniaturization of hair follicles (57). Because prostaglandins are not only able to induce acute inflammation but also convert short-lived inflammatory responses to long-lasting chronic inflammation (58), it is possible that their greater concentration in the scalp is one of the mechanisms involved in the inflammatory process that may be associated with hair loss. However, no data on prostaglandins in FPHL are available, and their role in FPHL remains unclear. Additionally, treatment with classic anti-inflammatory agents or immunosuppressive drugs does not seem to improve FPHL, and the relationships between scalp inflammation and other possible causal factors of FPHL have not been clarified (34).
FPHL seems to be related to several factors, including genetic influences (level of evidence B), androgen sensitivity and concentrations (level of evidence C), and possibly inflammation in the scalp (level of evidence C). In many patients, FPHL may be present in the absence of any detectable signs of androgen excess (level of evidence B). Overall, further research on the pathogenesis of FPHL is critically needed.
Epidemiology of FPHL in the general population
In the general population, the prevalence of FPHL increases with age and is higher in older women. Although there are different prevalence figures due to varying study designs and differences in ethnicity, in white women the prevalence of FPHL ranges from 3% to 12% in young women in their third to fourth decade, rising to 14% to 28% in postmenopausal women in their 50s and further increasing to 29% to 56% in women aged ≥70 years of age (59–61).
In a study conducted in the United Kingdom, Birch et al. (59) examined 377 women presenting to a general dermatology clinic with complaints not related to hair and found that FPHL was present in 6% of women <50 years of age and 38% of those ≥70 years of age. In another study, investigators examined a large sample of white women (n = 1006), including patients in a dermatology clinic as well as individuals from the surrounding community, none of which specifically sought help for hair loss and found a rate of FPHL of 19% for the overall population, ranging from 3% for women in their 20s, 17% for women in their 30s, and 23% for those in their 50s (60).
A 2005 community-based study of women in Australia assessed the self-reporting of hair loss using a mailed survey with a 34% response rate followed by examination of 193 women (61). Using a photographic scale with a progressive decrease in density in the central scalp and an assignment of FPHL for those with score ≥2, the investigators found an overall prevalence of FPHL of 32%. Notably, the survey also observed underreporting of hair loss by many of the respondents with FPHL, even those with most pronounced hair loss.
The prevalence of pattern hair loss in East Asian women is also related to age but seems to be overall lower than in white women (62–64). The largest study was conducted by Su et al. (62) in Taiwan, where it was observed that there was a prevalence of FPHL of 11.8% overall, ranging from 7% to 15% with aging. A study in Korean women (n = 4601) found an overall prevalence of 5.6% (63), similar to the rate observed in a 2010 study of Chinese women (64).
Differences in the prevalence of FPHL between ethnic groups may depend on genetic differences in molecular action and biochemistry of several factors that regulate hair growth. Consistent with this hypothesis, in East Asia populations, hirsutism is less common. For example, in hyperandrogenic Japanese women, despite similarly elevated T levels, hirsutism scores and serum levels of 3α-hydroxy-androstanediol-glucuronide, a marker of 5α-reductase activity, are significantly lower than in white populations (65).
Available data suggest a relatively low prevalence of FPHL in young women aged 20 to 30 years, increasing with age to one-third to one-half of women ≥70 years old (level of evidence C). In East Asian women a similar age-related trend is observed, although the overall prevalence of FPHL is lower than in white women (level of evidence C).
Prevalence of hyperandrogenism in women with FPHL
In studies of consecutive women with FPHL, the prevalence of clinical (hirsutism) or biochemical (increased circulating androgen levels) hyperandrogenism has varied widely. In one retrospective study of 187 women with FPHL and 21 healthy control women, 67% of the patients with hair loss alone (n = 110) and 84% of patients with hair loss plus other symptoms of androgen excess had hyperandrogenemia (66). In another retrospective study of 53 consecutive female patients with isolated FPHL (without hirsutism or other clinical sign of endocrine disturbance), only 26% demonstrated hyperandrogenemia (67). In a retrospective study of 166 female patients diagnosed with FPHL, 8.2% of patients had increased total T, 2% increased free T, and 8.9% increased dehydroepiandrosterone (DHEA) sulfate (DHEAS) concentrations (68). In a prospective study, Futterweit et al. (9) assessed 109 women with FPHL and observed that 38.5% had increased androgen levels; 26.1% of those with hyperandrogenemia had no associated hirsutism or menstrual dysfunction.
Although hyperandrogenism appears to be associated with FHPL, the strength and nature of this association remains unclear (level of evidence C).
Prevalence of FPHL in women with PCOS: results of a systematic review and meta-analysis
In retrospective studies including female patients with different forms of androgen excess, the prevalence of hair loss has been reported to range from 1.9% to 8% (69–71). To address the relationship of the most common androgen excess disorder in women, namely PCOS, with FPHL, the task force undertook a systematic review and a meta-analysis to obtain pooled prevalence estimates of FPHL in patients with PCOS.
The initial search identified 86 articles (Fig. 2, left panel). Review of article abstracts led to the exclusion of 62 articles, with 24 studies remaining, whose text was fully assessed for eligibility (Fig. 2, left panel). Of these, 15 were subsequently excluded because of several reasons (Figure 2, left panel). Consequently, nine articles (72–80) were included in the quantitative synthesis and meta-analysis of the prevalence of FPHL in patients with PCOS (Fig. 2, left panel). The characteristics of the patients included in these studies are summarized in Table 1. Most of the patients in these studies were in their third decade of life, important when comparing the data derived from these studies with those obtained in the general population.
|Study||Ascertainment of Alopecia by a Dermatologist||Ethnicity||Age of Patients With PCOS (y)||Sample Size (n)||Patients With FPHL (n)|
|Özdemir et al., 2010 (72)||Yes||Turkish||23 (16–40)||115||40|
|Wang et al., 2013 (73)||Yes||White plus Asian||28 ± 6||123||25|
|Quinn et al., 2014 (74)||Yes||Multiethnic||28 ± 6||254||56|
|Gowri et al., 2015 (75)||Yes||Indian||(21–30)||40||12|
|Hong et al., 2015 (76)||Yes||Korean||25 ± 7||40||5|
|Leerasiri et al., 2016 (77)||No||Thai||26 ± 7||145||63|
|Schmidt et al., 2016 (78)||Yes||Multiethnic||28 ± 6||237||53|
|Feng et al., 2018 (79)||Yes||Chinese||30 ± 6||186||61|
|Keen et al., 2017 (80)||Yes||Indian||25 ± 4||100||31|
All of these studies applied the European Society of Human Reproduction and Embryology/American Society of Reproductive Medicine criteria for the diagnosis of PCOS. Data are counts, means ± SD, or medians (range).
Applying a quality-effects model, the pooled prevalence of FPHL in patients with PCOS was 28% (95% CI, 22% to 34%) (Fig. 2, top right panel). A funnel plot (Fig. 2, bottom right panel) representing the double arcsine transformation of the prevalence against the SE demonstrated minimal asymmetry, thereby ruling out publication bias. Of note, only one study (79) compared PCOS to non-PCOS controls, finding a prevalence of FPHL of 23.1% vs 8.8% in age-matched controls in China.
Although the available data are few and the results highly variable, it may be concluded that evidence of FPHL is found in 20% to 30% of patients with PCOS (level of evidence C).
Diagnosis of FPHL
The diagnosis of FPHL is primarily made clinically, although laboratory analyses play an important role in identifying specific underlying causes that may contribute to the development of hair loss in affected women.
Women with potential FPHL should be queried about age of onset and evolution of their hair loss. Sharing with patients the Sinclair scale for FPHL (81, 82) or the modified Hamilton–Norwood scale for MPHL (83) may be helpful for determining family history of MPHL or FPHL. Symptoms related to hyperandrogenism such as hirsutism, menstrual irregularities, difficulty conceiving, or severe cystic acne should be sought. Use of medications that are related to hyperandrogenism, such as valproic acid and androgen supplementation or prior use of 13-cis retinoic acid, should be also investigated.
The following clinical features are characteristic of FPHL (10, 14, 21).
Pattern of hair loss/decreased hair density on the top of the scalp
Usually scalp hair loss will be either central loss or frontal accentuation, rarely the vertex/male pattern of hair loss (10). A decrease in hair follicle density can be confirmed by comparing the density in the occiput, which is not under androgen control and not involved in FPHL, with that of the top of the scalp. This is best demonstrated by parting the hair vertically in both areas and comparing the part widths (84). The parietal scalp may be involved.
Miniaturization of hairs in the affected areas of the scalp
In FPHL, the degree of miniaturization is never as profound as in men, so true baldness is rare (10, 85). Instead, there will be hairs having a variety of diameters and lengths in affected areas. This can be appreciated using a piece of black or white fabric (depending on hair color) placed behind the hairs or by dermatoscopy.
Absence of diffuse shedding
Although by histological examination there is an increased percentage of hairs in the telogen phase in affected areas of FPHL compared with normal (85), a history of increased shedding is inconstant. Hair shedding should not be global, but only in affected areas, if it does occur at all. An increase in the percentage of hairs in the telogen phase can be determined by performing several hair pulls in the affected area of FPHL and three to four other areas of the scalp, including at least one from the occiput. A hair pull is a simple bedside test performed by gently pulling on a group of 50 to 60 hairs from the scalp surface to the ends of the hair: it should normally generate 0 to 2 hairs per pull (86). If there is an increase in the number of hairs removed with each pull in multiple areas of the scalp, this suggests another pathophysiologic process in addition to FPHL such as telogen effluvium (84).
Preservation of follicular ostia (follicular opening) on the scalp in affected areas
The presence of ostia implies potential regrowth. Loss of ostia implies loss of the follicular unit and sebaceous gland and a scarring or cicatricial process (84).
Presence of focal atrichia
Focal atrichia are small (∼4 mm) areas of the central scalp devoid of hair. They are observed in 44% of women with FPHL compared with 2% of subjects with other hair disorders, and they are more common in late onset (67%) than early onset (15%) FPHL (87).
A dermatoscopic examination offers a means of confirming the presence of follicular ostia and the variety of diameter hairs involved in the miniaturization process in FPHL (88). Also seen in FPHL is an increase in single hair units in the frontal scalp compared with normal, yellow dots indicative of empty follicles and peripilar hyperpigmentation (89), with the latter seen in 86% of women with FPHL (90).
To be most useful, a 4-mm punch scalp biopsy involving the dermis should be performed, followed by horizontal sections to establish the number of hairs per unit area (91). Much of our knowledge about histological findings in FPHL and potential explanations for the clinical findings have been extrapolated from biopsies of men with MPHL. In one study, biopsies of MPHL demonstrated a terminal-to-vellus hair ratio of <1.7:1 (<3:1 considered diagnostic) and an increased percentage of telogen hairs (>10%) (91). Vellus hairs are defined as <0.03 mm in diameter and have no attached arrector pili muscle (92). For comparison, in a sample of control male and female patients without hair loss, a terminal-to-vellus hair ratio of 7:1 and 7% telogen hairs were observed (91). There are ethnic variabilities in the density of hair follicles. The total number of hairs reported in 4-mm punch biopsies in normal white (91, 93), African American (93), and Asian subjects (94) was 36 to 40, 22, and 16, respectively.
In FPHL, there is a greater variability in the number of miniaturized hairs in each follicular unit in an area of hair loss, supporting the concept that each hair follicle may have a different susceptibility to the responsible factors. One of the most notable changes in pattern hair loss is the histopathologic presence of concomitant inflammation, highlighted by the presence of lymphocytic infiltrates. Mild inflammation is observed in about one-third of controls and in men with MPHL, but moderate inflammation is far more common in patients with MPHL than in controls (36% to 9%) (94).
Stelae (or streamers) are the collapsed connective tissue sheath that is left behind as a telogen hair moves upward in the scalp to the level of the bulge area. In MPHL, and possibly FPHL, the stelae are hypertrophic, excessively vascularized, and have a mixed inflammatory infiltrate of lymphocytes, mast cells, macrophages, and large Arao bodies (dense aggregates of elastin fibers) (33). In late MPHL, streamers may present as fibrotic tracts with some follicles obliterated by the inflammatory process (95).
Some common causes of female hair loss must be distinguished from FPHL, including chronic telogen effluvium (CTE), central centrifugal cicatricial alopecia (CCCA), frontal fibrosing alopecia (FFA), and fibrosing alopecia in a pattern distribution (FAPD), as detailed below.
The most common condition requiring differentiation from FPHL is CTE. This condition is defined as diffuse scalp hair thinning, generally in middle-aged women, which is accompanied by generalized persistent shedding (53). There is no frontoparietal or central pattern of hair loss as in FPHL, but frequently there may be bitemporal recession. A hair pull is usually positive from multiple areas of the scalp, including the occiput. A biopsy may be helpful in distinguishing CTE from FPHL, as the terminal-to-vellus hair ratio in CTE is generally 9:1, indicating a lack of miniaturization. An algorithm using the amount of total and ≤3-cm hairs (considered telogen vellus hairs) shed per day, alone or in combination, has been suggested as a noninvasive method to help distinguish the two conditions (96).
CCCA is a scarring type of hair loss, with lymphocytic infiltration predominant on biopsy, primarily seen in African American women who present with central, frontal, or vertex predominant loss (97–99). With time, the hair loss leads to frank baldness and lack of follicular ostia in the involved areas. Many of these women have signs and symptoms of hyperandrogenism (98), but no large-scale evaluation of their androgen status has been done. Olsen (99) has hypothesized that these women may have underlying FPHL that becomes inflammatory with the use of some hair care practices, such as hot combs, relaxers, or occlusive ointments, leading to a cicatricial type of hair loss.
FFA and FAPD
FFA was only first described in 1994 (100) but has become one of the most frequent causes of cicatricial alopecia. Most subjects with FFA are women, primarily white and primarily postmenopausal, although men, premenopausal women, and affected patients from all races have been reported with FFA. Key clinical features are a receding hairline with perifollicular erythema and eyebrow loss. A scalp biopsy shows a lymphocytic infiltrate typical of the patchy scarring alopecia called lichen planopilaris. Many women have central scalp hair loss as well, and a biopsy is critical to do in this area in addition to that in the frontal scalp.
FAPD presents with central hair loss in a Ludwig pattern, albeit typically with perifollicular erythema. FAPD also demonstrates on biopsy lichen planopilaris (101). Both FFA and FAPD, similar to FPHL, show some improvement with 5α-reductase inhibitors.
Laboratory evaluation in women with FPHL
Circulating androgen measures
The most common androgens measured in the circulation include total and free T, androstenedione, and the DHEA metabolite DHEAS. When considering total T measures, it is generally best to avoid those immunoassays that do not use extracted and otherwise purified blood samples, because of their recognized unreliability (102). Consequently, many commercial diagnostic laboratories in the United States have begun to measure total T (and other related steroids) using mass spectrometry, following either gas or, more commonly, liquid chromatography (LC/MS) (103). These assays have a high degree of sensitivity and specificity, and they generally provide accurate results. LC/MS is used as the reference method at centers such as the US National Institute of Standards and Technology and is the “gold standard” for measurement of serum T (103). A few commercial radioimmunoassays (104) and chemiluminescent immunoassay (105) that do not include prior extraction and purification of the sample may be as accurate as LC/MS assays in patients with PCOS, and even at very low T concentrations such as those from prepubertal and postpubertal children (104). However, more studies are needed to confirm these observations.
Circulating free T concentrations are more sensitive than the measurement of total T for establishing the presence of androgen excess (102). Accurate techniques for the estimation of free T include equilibrium dialysis and ultracentrifugation assays, which determine the percentage of free T, and ammonium sulfate precipitation techniques, which determine bioavailable T by precipitation of the SHBG-bound T fraction (106). In any of these free T assays, a high-quality total T measurement is needed (102, 106). However, these assays are cumbersome, and most commercial laboratories use direct radioimmunoassay or chemiluminescent immunoassay assays that rely on the competitive binding of a nontestosterone analog for measuring free T, which are notoriously inaccurate (107, 108).
Alternatively, it may also be useful to perform concomitant measurement of SHBG along with the total T measurement, permitting the calculation of free T (106). This estimate of free T has good concordance and correlation with free T measured by equilibrium dialysis, provided that an accurate method is used for the total T measurement (106).
Testing for androgens or androgen metabolites, other than total and free T, such as androstenedione, DHEA, and DHEAS, does not appear to add much to the detection of hyperandrogenism in women with PCOS and FPHL (72, 74). The potential value of 11-oxo androgens, as described above (44), in the evaluation of androgen excess in women with FPHL remains to be determined.
Other hormone analyses
The frequency of thyroid disorders in patients with FPHL does not appear to be significantly higher than in the general population. It was 0% in the prospective series of 109 women with alopecia reported by Futterweit et al. (9). In a survey study of 64 patients (ages 20 to 88) with FPHL who were being seen at a hair clinic, 31.25% indicated they were hypothyroid; however, there was no significant association between FPHL severity and thyroid disease (68). In a retrospective study of 210 patients (ages 68 to 86 years) with FPHL, 12% (n = 26) had a preexisting history of thyroid disease (and presumably were being treated), and thyroid screening identified only one more affected individual (109). However, subclinical hypothyroidism is common in older women, and prevalence up to 18% in women in their 80s has been reported (110). Hypothyroidism, similar to nutritional deficiencies, should be sought in FPHL, although not necessarily as a causative factor, but as a condition that itself can cause a telogen effluvium and make hair regrowth difficult in FPHL.
The utility of measuring serum prolactin in patients with FPHL is unclear. In the prospective series of 109 women with FPHL examined by Futterweit et al. (9), 7.2% of patients had elevated prolactin and two (1.8%) had a prolactinoma. The possible link between high prolactin and pattern hair loss is unclear and further studies are needed. Increased prolactin levels may be secondary to hypothyroidism or to hyperestrogenism (102).
Mineral and vitamin deficiency assessment
Recently, there has been much interest on the possible role of nutrient deficiency in FPHL (111, 112). Many studies have been dedicated to the possibility that iron deficiency contributes to hair loss in FPHL (113). In a retrospective study of 100 multiethnic patients with FPHL in Vancouver (BC, Canada), deficiency in ferritin was observed in 29% at the first consultation (113). In a prospective, case-controlled study of 38 women (18 to 45 years of age), mean ferritin levels were lower in patients with FPHL than in controls; serum ferritin levels were lowest in those patients with the severest degrees of hair loss (114). However, in another case-controlled study that included 170 women with FPHL, 58 with CTE, and 47 control subjects, there was no statistically significant increase in the incidence of iron deficiency in premenopausal or postmenopausal women with FPHL or CTE vs controls, regardless of whether a definition of ferritin <15 μg/L or <40 μg/L was used as the definition of iron deficiency (115). Other investigators have reported that serum ferritin is lower in premenopausal, but not in postmenopausal, patients, compared with controls (116). In the same study supplementation with iron in patients with low-normal serum ferritin (<70 μg/L) did not improve the hair loss (116). Overall, there remains an uncertain relationship of iron deficiency and FPHL, given the frequent occurrence in the general premenopausal population of iron deficiency and a lack of well controlled studies on whether correction of iron deficiency affects FPHL.
Mean vitamin D levels were reported to be reduced in many patients with hair loss (109, 116, 117). In the Vancouver study, 62% of patients with FPHL presented low serum levels of 25-hydroxyvitamin D (109). In a prospective study, mean zinc levels, but not mean copper levels, were lower in 77 women with FPHL vs 32 controls (118). In the Vancouver study, 15% of patients with FPHL presented with low zinc levels (109). Well-controlled studies are needed to establish the relatedness of these deficiencies to FPHL and whether correction of the deficiency specifically improves hair growth.
The diagnosis of FPHL is mainly clinical and requires a careful medical history and physical examination, including evaluation of the scalp, and assessment for the presence of other immune or endocrine disorders (level of evidence B).
The differentiation of FPHL from other causes of hair loss in women may require a dermatoscopic examination and/or a scalp biopsy (level of evidence C).
Clinicians taking care of patients with FPHL should be able to recognize the different clinical alterations that may be present in female hair loss.
In women with FPHL circulating androgens should be evaluated (level of evidence C).
Androgen measures should include measuring total T using a high-quality assay, preferably by LC/MS or by an immunoassay including sample extraction and purification, and free T, either using equilibrium dialysis, ultracentrifugation, or ammonium sulfate precipitation, or estimated using SHBG (level of evidence C).
Measurements of blood iron studies, vitamin D, zinc, thyroid profile, and prolactin may be useful to rule out and treat other conditions that may affect hair regrowth in FPHL (level of evidence C).
Treatment of FPHL in hyperandrogenic women
A recent Cochrane review noted that most studies of therapy in women with FPHL were of low quality, mainly due to bias in design (e.g., no blinding) or a small sample size and were only able to detect enough evidence to support the use of topical minoxidil in women with FPHL, although even then the effect size was small (119).
Topical minoxidil is considered the first-line treatment of FPHL. Minoxidil is a potassium channel opener hypothesized to enhance angiogenesis and vasodilation, and to have anti-inflammatory and antiandrogenic effects (14, 120). In animal studies, minoxidil shortens the telogen phase, resulting in premature entry of resting hair follicles into the anagen phase (120). The anagen phase is also prolonged and miniaturized follicles are enlarged with minoxidil, resulting in increased hair counts, weight, and density (120).
Topical minoxidil is available in 2% and 5% solution (TMS) or in 5% foam (TMF). In both MPHL and FPHL there is evidence of greater efficacy of the 5% vs 2% TMS (121, 122). The 5% TMF applied once a day has been shown to be equivalent to 2% TMS applied twice a day in FPHL (123), and both 5% preparations are used most frequently in clinical practice (14). During the initial treatment period of 2 to 8 weeks, there can be a transient increase in scalp hair shedding, which results from the shedding of telogen hairs in preparation for new anagen growth (14, 120). Scalp pruritus and flaking may occur with therapy, more common with TMS, which contains propylene glycol, vs TMF (121, 123). Facial hypertrichosis may also occur and is more common with 5% than 2% formulations (14, 121). Topical minoxidil should be used for at least 12 months for optimal results (14).
When the patient has severe hair loss or hyperandrogenism, a combination of skin-directed and systemic therapy with an antiandrogen or 5α-reductase inhibitor may enhance results (3). Although one meta-analysis suggested no significant improvement with either a 5α-reductase inhibitor or antiandrogen (119), multiple small studies, case reports, and general prescribing practices suggest otherwise. More data with outcomes stratified by age, hair loss severity, and hyperandrogenism are needed to better understand the efficacy of these agents.
When adding an antiandrogen or 5α-reductase inhibitor, it must be remembered that these products have teratogenic potential for feminization of a male fetus and patients should be counseled accordingly, and strict contraception should be used during and for at least 30 days after discontinuation of these drugs.
Finasteride is often used in women with FPHL, especially when hyperandrogenism is present. However, clear-cut evidence of its effectiveness is less clear, and a meta-analysis evaluation failed to demonstrate a significant effect of finasteride in FPHL (119). However, as previously observed, interpretation of data by meta-analysis is difficult because of inclusion of patients with FPHL with different age and inflammatory and androgen status.
Differences in age and in finasteride doses may be important. Differently from what has been reported in men (124, 125), in postmenopausal women with FPHL 1 mg daily of finasteride did not appear to be any more effective than placebo (85, 124, 126). Higher doses of finasteride (2.5 or 5 mg) have been associated with improvement of FPHL, but it has been shown mainly in postmenopausal nonhyperandrogenic patients. In fact, a retrospective study of premenopausal and postmenopausal patients with FPHL prescribed finasteride 2.5 mg/d found that of 112 patients studied 30% demonstrated slight and 65% significant improvement using global photographs, whereas 5% had no change. In this study, finasteride had a better effect on hair growth when patients had a lower Ludwig score and an older age at onset (126). Similar results have been reported in other studies (127–130) that found improvement of FPHL in postmenopausal normoandrogenic women with both 2.5- and 5-mg doses of finasteride.
Few data are available in young women with FPHL. In a retrospective study including 30 women <50 years of age and 30 women >50 years of age, 1.25 mg/d finasteride administered for 3 years was found to be effective in both age groups (131). No data on androgenic status of these patients are available. Instead, in a group of hyperandrogenic young women with FPHL, finasteride treatment did not result in significant changes in the Ludwig score (132). Alternatively, several case studies have shown that low doses (1 or 1.25 mg) may be also effective in treatment of FPHL (133, 134).
These data suggest that finasteride may be useful for the treatment of FPHL in some patients in doses of 2.5 or 5 mg daily. However, the task force recognizes that most finasteride studies have been conducted in postmenopausal women because of the higher prevalence of FPHL in this age range and to avoid the potential for women of childbearing age becoming pregnant while on the drug. Consequently, although current data support the use of finasteride for FPHL primarily in postmenopausal women, various clinical reports suggest that it may also be effective in selected premenopausal women with FPHL. Large unbiased prospective controlled studies are needed to determine the efficacy and best approach for the use of finasteride in FPHL.
Dutasteride, an inhibitor of both type I and type II 5α-reductase, has been found to be more effective than finasteride in men with MPHL (135, 136). In patients with FPHL, data on dutasteride are almost absent. In the retrospective study of 30 women <50 years of age and 30 women >50 years of age, 3 years of therapy with 0.15 mg of dutasteride was similarly effective to 1.25 mg of finasteride in improving hair density and thickness but more effective in the central and vertex sites in women <50 years old (131). Dutasteride has a much longer half-life than finasteride, and women of child-bearing potential should avoid its use without effective contraception for at least 6 months after discontinuation owing to its potential teratogenic effect.
Spironolactone, a potassium-sparing diuretic that is commonly used in hirsutism treatment because of its antiandrogen activity, represents a possible alternative to finasteride in patients with FPHL (68, 137, 138). Doses of 100 to 200 mg/d and prolonged treatments (at least 12 months) are needed. Hair regrowth or stabilization of hair loss is observed in a variable number of patients ranging from 44% (136) to 74% (68). Patients with FPHL and hirsutism or acne have a significantly better response to spironolactone (68) than do women with only FPHL.
Flutamide is a pure antiandrogen that has been successfully used in FPHL, but that bears the risk of important, albeit rare, hepatotoxicity. Because of this, few studies have evaluated the efficacy of flutamide in women with FPHL. In a randomized open-label trial in 36 hyperandrogenic young women with FPHL, 250 mg/d flutamide resulted in a significant reduction (−21%) in Ludwig scores (132). In a prospective nonblinded cohort study, 101 women diagnosed with FPHL received yearly reducing doses of flutamide (250, 125, and 62.5 mg/d) for 4 years (139). The cohort included 33 patients treated with flutamide alone and 68 treated with flutamide and an oral contraceptive. Both groups demonstrated marked decrease in alopecia scores after 12 months of flutamide therapy, compared with baseline values; the maximum drug effect occurred after 2 years and was maintained during the subsequent 2 years of therapy (139). In conclusion, although flutamide may be more effective than spironolactone or finasteride in treatment of hyperandrogenic FPHL, its use is not recommended because of the risk of hepatotoxicity.
Cyproterone acetate is an antiandrogen product that is not available in United States. This product has been used at high doses (50 to 100 mg/d) for hyperandrogenic women with FPHL with mixed results (132, 140).
Alternative treatments for FPHL
Low-level laser light therapy
Red or near-infrared laser light has been reported to promote tissue repair and regeneration and, more recently, low-intensity light laser (LLL) therapy has been demonstrated to stimulate hair growth, probably by increasing the mitotic rate of the hair follicle stem cells or follicular keratinocytes (141). Additionally, LLL may modulate inflammation by increasing TGF-β1 and IL-10 while reducing proinflammatory cytokines and prostaglandin E2 (141). Usually wavelengths between 600 and 1550 nm (in red/infrared spectrum), emitted by helium-neon and fractional erbium glass lasers, are used. Several studies have shown that LLL may improve FPHL (141–144) but the effect is generally modest and, in a male study, although hair count increased, no differences with global photographs were observed (145). Adverse effects are few but may include headache, erythema, pruritus, burning, pain, and mild paresthesia (141).
Platelet-rich plasma (PRP) is produced by centrifuging a patient’s own blood, concentrating the platelets, and injecting the concentrate into the patient’s scalp in affected areas. In addition to platelets, PRP contains numerous growth factors, chemokines, and cytokines (146). Results are promising; however, evidence is very limited and studies that have been performed have used different protocols without sufficient follow-up (147).
In women with FPHL, hair transplantation may also be useful (148) although the hair loss process, as in men, will continue outside of the transplant areas, so continued use of medical treatment after surgery is necessary.
In men with MPHL, ketoconazole shampoo has been shown to increase hair density, hair size, and percentage of hairs in the anagen phase (149), and topical ketoconazole has been found to increase hair growth (150). However, data in women with FPHL are lacking.
Topical minoxidil is an effective treatment of FPHL and should be used as the first-line treatment (level of evidence B).
Therapeutic results of minoxidil may be enhanced with the addition of 5α-reductase inhibitors or antiandrogen therapy. The 5α-reductase inhibitor finasteride, in doses of 2.5 or 5 mg/d, and the AR-blocker spironolactone, in doses of 100 to 200 mg/d, may be used with the lowest risk (level of evidence C). Dutasteride may be useful, but clinical evidence is limited (level of evidence D). Flutamide is potentially effective, although its use is associated with significant, albeit rare, hepatotoxicity (level of evidence C), and consequently use in patients with FPHL is not encouraged. Cyproterone acetate appears to have unclear effectiveness (level of evidence C).
Note that most of the data generated with the use of 5α-reductase inhibitors and antiandrogen for FPHL therapy are in postmenopausal women. In women of reproductive age, 5α-reductase inhibitor and antiandrogen treatment should be used concomitant with secure contraception to minimize the teratogenic risk of these medications (level of evidence B).
LLL therapy may be of modest benefit for FPHL (evidence level C). Scalp injections of PRP are a promising emerging therapy, although evidence is limited (level of evidence C). Hair transplantation is a possible therapy to assist in hair recuperation (level of evidence B), although it should be used in conjunction with continued medical therapy (level of evidence D).
The term “female pattern hair loss” should be used, avoiding the previous terms of alopecia or androgenetic alopecia.
The two typical patterns of hair loss in FPHL are centrifugal expansion in mid scalp or a frontal accentuation or Christmas tree pattern.
Isolated FPHL should not be considered a sign of hyperandrogenism when androgen levels are normal and no other clinical signs of hyperandrogenism are present.
The assessment of patients with FPHL is primarily clinical and requires differentiation from other forms of hair loss in women.
In all patients with FPHL, assessment of a possible androgen excess is mandatory. Measurement of vitamin D, iron, zinc, thyroid hormones, and prolactin are optional but recommended.
Treatment of FPHL should begin with minoxidil (5%), adding 5α-reductase inhibitors or antiandrogens when there is severe hair loss or associated hyperandrogenism.