| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Obstetrics and Gynecology (C.M.D., R.A.), Cedars-Sinai Medical Center, Los Angeles, California 90048; Department of Obstetrics and Gynecology (C.M.D., R.A.), The David Geffen School of Medicine at University of California, Los Angeles, Los Angeles, California 90095; and Departments of Biostatistics and Biomathematics (K.S.W., A.A.B.) and Obstetrics and Gynecology (R.A.), University of Alabama at Birmingham, Birmingham, Alabama 35294
Address all correspondence and requests for reprints to: Ricardo Azziz, M.D., M.P.H., M.B.A., Cedars-Sinai Medical Center, Department of Obstetrics and Gynecology, 8635 West Third Street, Suite 160 W, Los Angeles, California 90048. E-mail: azzizr{at}cshs.org.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The purposes of the present study were to define 1) the degree of facial and body terminal hair, as assessed by the modified Ferriman-Gallwey (mFG) score, in unselected women from the general population; 2) the effect of race (Black and White) on the same; and 3) the normative cutoff values.
Design and Setting: We conducted a prospective observational study at a tertiary academic medical center.
Patients/Participants: Participants included 633 unselected White (n = 283) and Black (n = 350) women presenting for a preemployment physical exam.
Interventions: Interventions included history and physical examination.
Main Outcome Measures: Terminal body hair growth was assessed using the mFG scoring system; nine body areas were scored from 0–4 for terminal hair growth distribution.
Results: The mFG scores were not normally distributed; although cluster analysis failed to identify a natural cutoff value or clustering of the population, principal component and univariate analyses denoted two nearly distinct clusters that occurred above and below an mFG value of 2, with the bulk of the scores below. Overall, an mFG score of at least 3 was observed in 22.1% of all subjects (i.e. the upper quartile); of these subjects, 69.3% complained of being hirsute, compared with 15.8% of women with an mFG score below this value, and similar to the proportion of women with an mFG score of at least 8 who considered themselves to be hirsute (70.0%). Overall, there were no significant differences between Black and White women.
Conclusions: Our data indicate that the prevalence and degree of facial and body terminal hair growth, as assessed by the mFG score, is similar in Black and White women and that an mFG of at least 3 signals the population of women whose hair growth falls out of the norm.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The modified (i.e. only nine body areas considered) Ferriman-Gallwey scoring system is the method in general use for visually scoring excess terminal body or facial hair growth for the clinical or investigational assessment of hirsutism (3). Using data from the 161 women 18–38 yr of age seen in a North London general medical clinic, and considering only the nine hormonal body areas, Ferriman and Gallwey (3) observed that 9.9% had scores of at least 6, 4.3% had scores of at least 8, and only 1.2% of women had combined scores of at least 10 for the nine body areas. Although Ferriman himself later chose a cutoff score of five or more when defining hirsutism (4), other investigators indicated that a cutoff value of 8 or greater represented abnormal hair growth, basing this definition on the 95th percentile of the data originally collected by Ferriman and Gallwey (5).
Overall, it remains unclear whether a cutoff of 8 or less should be used to define normality. Although the 95th percentile has often been used to define the upper limit of normal, this cutoff limit may, in fact, not represent a physiological delimitation. In addition, race appears to affect the characteristics of terminal hair growth. For example, in studies of scalp hair follicles, White individuals had a higher density of hair follicles than Blacks, who in turn had more follicles than Asians (6, 7, 8). Consequently, it is important to determine whether the cutoff value for defining abnormal terminal body and facial hair growth is affected by race. The purposes of the present study were to define, in a population of unselected women, 1) the degree of facial and body terminal hair, as assessed the mFG score, in unselected women from the general population; 2) the effect of race (Black and White) on the same; and 3) the normative cutoff values in these women. To determine these parameters, we analyzed data obtained prospectively in 633 unselected White and Black women presenting for a preemployment physical exam.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The methods for recruiting and evaluating the study subjects have been previously described (9, 10). In brief, all prospective employees and staff of the University of Alabama at Birmingham (UAB) undergo an entrance medical evaluation that includes a brief history and physical, and blood sampling. It should be noted that UAB is the single largest employer in the city of Birmingham and the third largest employer in the state of Alabama, and its employees represent a cross-section of the population of the surrounding county. Between March 1996 and July 1997, and between July 1998 and October 1999, all consecutive premenopausal females who were to undergo a preemployment physical exam were asked to participate. The Institutional Review Board of UAB approved the study, and all subjects entered gave written consent. The majority of the subjects in the present study have been previously included in our studies determining the prevalence of PCOS (9, 10). All women included in this study were of either Black or White race; women of other races, because of their low numbers, were excluded from the study.
Study protocol
Subjects completed a standardized history form, including questions regarding frequency of treatment for unwanted hair growth, menstrual cycle length and regularity, medication use, menopause or previous hysterectomy, and whether they felt they were hirsute. Oligomenorrhea was defined as an average cycle length of greater than 35 d in length and included women with frank amenorrhea. All also underwent an assessment of excess terminal hair growth by the same examiner using the previously described modification of the method originally described by Ferriman-Gallwey (i.e. mFG), scoring the presence of terminal hairs over nine body areas (upper lip, chin, chest, upper and lower abdomen, thighs, upper and lower back, and upper arms) from 0–4 (5). Terminal hairs are defined as coarse pigmented medullated hairs, generally growing greater than 1 cm in length if uncut. The research nurses made the initial assessment, after being trained on the use of the scoring system by the principal investigator (R.A.) and noted their findings on a preprinted graphical scoring sheet. To maximize the accuracy of the hirsutism scoring, all subjects with an initial mFG score greater than 3 per the study nurse were reexamined by a single physician and rescored using the same preprinted scoring sheet. In the vast majority (85%) of women reexamined by this investigator, the mFG scores did not change by more than two points from that obtained by the research nurse (9). To minimize treatment bias, we included all women regardless of concomitant hormonal therapy use, including oral contraceptive pills, continuous progestins, glucocorticoids, or insulin sensitizer therapy.
Statistical analysis
All parameters are given as median and range for skewed data. Generalized Wilcoxon and 
2 test were used when appropriate for group comparisons. Logistic regression analysis was used to assess the independent effects of race and age on mFG score. Principal component, cluster, and univariate analysis of the mFG scores were also performed to identify a natural cutoff value or clustering in the population. We also calculated the propensity scores for differentiating groups with different mFG scores. Propensity scores have been used widely in the medical literature for a variety of purposes including bias reduction, multivariable matching, and covariate adjustment (11) and are a scalar measure for each observation summarizing the effect of multiple covariates on different exposure or treatment groups. A P value of <0.05 was considered statistically significant. All data were analyzed using the SAS software, version 9.0 (SAS USA, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Comparing Black and White women
Black subjects were younger and less obese and had fewer pregnancies and gestational deliveries than their White counterparts (Table 1
). Of the subjects studied, 13.2% had a history of oligomenorrhea, higher among Black than White women; 3.3% were menopausal, higher among Whites than Blacks. Overall, 201 subjects (31.8%) were receiving hormonal therapy at the time of their evaluation. Among the 149 Black women using hormones, 126 (84% of those using hormones) used combined oral contraceptive pills, 14 (9.4%) depo-medroxyprogesterone acetate, three (2.0%) levonorgestrel implants, three (2.0%) cyclic progestins, two (1.3%) progesterone-only contraceptive pills, and one (0.7%) menopausal-type hormone replacement therapy. Among White subjects, 52 used hormones, 35 combined oral contraceptive pills (67%), seven (13.5%) menopausal-type hormone replacement therapy, five (9.6%) depo-medroxyprogesterone acetate, three (5.8%) estrogen only, and two (3.8%) cyclic progestins. Women taking hormones did not differ in age from those not receiving such medications but were more likely to be Black, were of slightly greater body mass, and had a slightly greater mFG hair growth score (Table 2).
|
|
). The mFG scores were not normally distributed and were skewed toward the lower end of the range (Fig. 1). However, there was no significant difference in the distribution of scores between Black and White women by nonparametric analysis.
|
= 0.05, we estimated that the study had a power of 0.83 to detect a medium effect size difference in the mean mFG scores of the two race groups. The medium effect size difference was set at 0.25 SD separations of the means. Defining the abnormal mFG score in the population
The 95th percentile mFG value of the combined population was 7.7. Consequently, using an mFG score of at least 8 to define the upper normal limit of the population, we observed that 5.4% (15 of 278) of White and 4.3% (15 of 349) of Black women evaluated would be considered abnormal (i.e. hirsute), a difference that was not statistically different.
However, we should note that the 95th percentile value might not necessarily indicate the populational or natural cutoff value. Consequently, a number of statistical methods (principal component, univariate, and cluster analyses) were used to identify structure or pattern in the distribution of the mFG scores in an attempt to identify potentially distinct subpopulations. Combining all subjects, cluster analysis of the mFG scores failed to identify a natural cutoff value or clustering of the population. However, principal component analysis indicated that three components explained 97% of the variation in the data. The first component was weighed heavily by age, the second by BMI, and the third by the mFG score. When the first component was plotted vs. the other components, two nearly distinct clusters occurred above and below the value of 2, with the bulk of the scores greater than 2. Likewise, in a univariate distribution of the mFG scores, the skewed bar graph indicated that the bulk of the data were at least 2 vs. the remainder of the data, which were at least 3. Finally, the value of 2 conformed to 2 SE above the mean of the raw distribution of FG scores.
Overall, it appeared that an mFG score of at least 3 was out of the norm, or abnormal, in our population. Taken together, 21.9% (61 of 278) of White and 22.6% (79 of 349) of Black women had mFG scores above this cutoff value, a nonsignificant difference; 18.3% of Black and 15.8% of White women had mFG scores between 3 and 7.
Estimating the effect of race on the normative values or prevalence of abnormal hair growth in the population
Multiple logistic regression analysis was performed to determine the impact of race on the mFG score cutoff values after adjusting for age, use of hormonal medications, and menopausal status. Nonetheless, the prevalence of mFG scores of at least 3 and at least 8, or the cutoff value, remained not different by race. In addition, we calculated the propensity score for race, composed of age, BMI, hormonal treatment, menstrual status, and menopausal status, to determine whether the groups of mFG scores were affected by the possible imbalance of prognostic variables. That is to say, we constructed predicted probability scores or propensities of belonging to either race, setting race as the dependent variable in a logistic model and as a function with other important confounding variables. This new score or probability becomes the new single confounding variable in the model for mFG and is a composite of the variables from which it was made and has the remarkable property of adjusting for all the covariates that went into its estimation (12). When balancing race for all other relevant covariates using the propensity technique, race remained a nonsignificant factor for determining the mFG score; i.e. the propensity for either race was not significant at an mFG cutoff of at least 3 (P = 0.0.089) or at least 8 (P = 0.084). Although these P values would appear to approach significance, the absence of a significant relationship with race is valid considering the large sample size. Overall, the power of our study to detect a difference of 12% in the frequency of scores above or below the cutoff of 2 was greater than 0.80.
Self-reported cosmetic treatment for excess hair growth: relationship to mFG scores, cutoff values, and prevalence of abnormal hair growth
One hundred forty-seven women (23.2% of the total) used some sort of cosmetic treatment for excess hair growth (i.e. bleach, pluck, shave, wax, or electrology). There was no racial difference in the proportion of women using treatment for unwanted hair growth (23.3 and 23.1% of White or Black women, respectively). Of women using hair treatments 36.7% used it approximately once monthly, 29.2% twice monthly, 19.0% once weekly, 3.4% twice weekly, 6.8% every other day, and 4.8% daily. The frequency of treatment correlated positively with the mFG (r = 0.27; P < 0.0001). Sixty-two of 147 (42.2%) women using treatment for unwanted hair had an mFG score less than 3.
Overall, 175 subjects (27.6%) stated that they felt they were hirsute, and 144 of these (82.3%) were using some method of hair removal. In turn, fully 143 of 147 (97.3%) of subjects treating unwanted hair also stated they felt they were hirsute. Women who self-reported being hirsute were more obese with a BMI median (range) of 25.1 (15.7–58.4) vs. 26.8 (17.7–53.2) kg/m2, respectively (P < 0.007) and had a higher median mFG score of 0 (0–20) vs. 3 (0–17), respectively (P < 0.03) but did not differ in median age of 27 (18–66) vs. 28 (19–64) yr, respectively, or race (proportion of Black subjectsm 53.1 vs. 56.1%, respectively) compared with subjects who self-reported that they were not hirsute.
Ninety-seven of 140 subjects (69.3%) with an mFG score of at least 3 complained of being hirsute, compared with 78 of 493 (15.8%) with an mFG score below this value, a significant difference (2 = 155.8; P < 0.0001). Of women with an mFG score of at least 3, 60.7% used some form of treatment for unwanted hair. This was not statistically different from the proportion of women with an mFG score of at least 8 who considered themselves to be hirsute or used some form of hair treatment (70.0 and 36.7%, respectively).
We then estimated the proportion of women that could be hirsute but were missed because of the use of hair removal techniques by analyzing the subpopulation of women who used treatment for unwanted hair at least twice weekly. Our results indicated that of women using treatment for unwanted hair at least twice weekly, six had an mFG less than 3. This suggests that the proportion of women with abnormal hair growth according to a cutoff value of at least 3 and potentially not detected because of the use of treatments for unwanted hair removal was 4.2% (6 of 140 + 6), with the number of affected women increasing from 140 to 146 affected women, or from 22.1 to 23.1% of the total. Alternatively, 18 women with an mFG score less than 8 used treatment for unwanted hair at least twice weekly, suggesting that the possible number of women with an mFG score of at least 8 could increase by approximately 50% (from 30 to 48) women, or from 4.7 to 7.6% of the total.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Defining what is normal in a population is not always obvious. If the 95th percentile of the distribution of mFG scores in our population is used to define normalcy, we note that abnormal would be a score of 8 or greater, similar to that observed by Ferriman and Gallwey (3). However, although the 95th percentile has often been used to define the upper range of normal, this cutoff limit may in fact not represent a physiological delimitation. For example, one of the differences between what defines a genetic variant as a mutation vs. a normal polymorphism is its relative frequency in the population; DNA sequence polymorphisms are usually defined when a genetic variation is present at greater than 1% frequency in the population (16). Alternatively, when the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus of the American Diabetes Association revised their cutoff value for fasting glucose levels used to diagnose type 2 diabetes mellitus, they did not use percentile values (17) but rather the risk of microvascular complications. Subsequent studies using these criteria in the United States noted that the prevalence of individuals with an abnormal glucose cutoff vale ranged from 7.8% of adults at least 20 yr old to 18.8% of individuals at least 60 yr old (18).
We attempted to define normalcy using statistical methods (principal component, univariate, and cluster analyses) intended to identify structure or pattern in the distribution of the mFG scores. Although traditional cluster analysis failed to identify specific distinct subpopulations, principal component analysis denoted two nearly distinct clusters occurred above and below the value of 2 with the bulk of the scores less than 2. Likewise, in a univariate distribution of the mFG scores, the bulk of the data were no more than 2 vs. the remainder of the data, which were at least 3. In addition, the value of 2 conformed to 2 SE above the mean of the raw distribution of FG scores. Our data then suggest that an mFG score of 3 or greater in our population could be considered above the norm, i.e. abnormal. Consequently, approximately 22% of unselected women of reproductive age, whether White or Black, had excess body or facial terminal hair growth, defined by an mFG score of 3 or greater.
A cutoff value of at least 3 as indicative of abnormality is consistent with the findings of other investigators. A study of 81 healthy female volunteers and 71 hirsute patients of childbearing age and Dutch ancestry noted that a score of more than 1 for the chin, upper back, upper abdomen, and upper arm, or more than 2 for upper lip, sideburns, chest, lower back, lower abdomen, thighs, or forearm, was abnormal (19). In addition, various other data support an mFG score of at least 3 as abnormal. For example, in our population, approximately 70% of subjects with an mFG score of at least 3 considered themselves hirsute, compared with only 16% of those with an mFG score below this value. In addition, over 60% of women with an mFG score of at least 3 used some form of cosmetic treatment for unwanted hair. Finally, recent data suggest that even minimal unwanted hair growth is an important indicator of underlying pathology, primarily an androgen excess disorder. In a study of 228 consecutive patients complaining of minimal unwanted hair growth (mFG scores between 1 and 5), we observed that 55% had androgen excess, primarily PCOS (20).
The mFG score in our population was found not to correlate, in addition to race, with body mass or age. This differs somewhat from the study by Sagsoz and colleagues (21) who observed a negative correlation between age, but not BMI, and total mFG score among 204 Turkish women. Other investigators also observed a decrease in facial hair growth with age (22, 23). Muti et al. (23), studying 314 healthy women, also observed a weakly positive correlation between BMI and facial hirsutism. An explanation for the difference between these results and ours is not readily apparent, although we should note that these previous studies tended to include greater numbers of older and menopausal women, which may have accentuated the changes in hair growth occurring with age.
This study has a number of strengths, including the inclusion of large numbers of women who were examined for reasons independent of medical necessity or disease, the inclusion of roughly equivalent numbers of Black and White study subjects, and the fact that all women with a screening mFG score of more than 3 were reexamined by one investigator. This latter investigational strategy is important considering the known between-observer variations in mFG scores (24). A major limitation to the study lies in the principal study endpoint, the mFG score. Nonetheless, the visual hair growth scoring method described and used in this study is the principal system currently used for clinically assessing the extent of body or facial terminal hair growth and consequently for diagnosing the presence of hirsutism.
Second, excess hair growth and/or hirsutism is likely to be readily treated cosmetically by most women, such that our study may have underestimated the true prevalence of the disorder. We estimated the proportion of women that could be hirsute but were missed because of the use of hair removal techniques, by analyzing the subpopulation of women who used cosmetic treatment for unwanted hair at least twice weekly. Our results indicated that of these women, six had an mFG less than 3, suggesting that the number of women with abnormal hair growth (according to a cutoff value of 
3) and not detected because of the use of treatments for unwanted hair removal is minimal (
4%).
Finally, hair growth will ameliorate with the use of sex steroid therapies, particularly oral contraceptives (25), and approximately 30% of subjects studied were on some form of hormonal therapy. Notably, subjects on hormonal therapy were slightly more obese and more hirsute than those not receiving these medications, potentially because these women are more likely to also have menstrual abnormalities that would be treated with hormonal therapy. We should stress, however, that these differences were small in degree. Less clear is the underlying reasons for the difference in the prevalence of hormonal therapy usage between Black and White women, although the younger age of the former group may, at least in part, explain their greater use of hormonal contraception. Nonetheless, although the use of hormones could have led to an underestimation of the prevalence and/or severity of hair growth, the impact of hormonal therapy on the mFG score was not obvious in our analysis.
In conclusion, our data indicate that the prevalence and degree of facial and body terminal hair growth is similar in Black and White unselected women and that an mFG of at least 3 signals the population of women whose hair growth falls out of the norm. Our data also indicate that a racially specific cutoff value for diagnosing hirsutism in not necessary, at least between Black and White women. Approximately 20% of our unselected women of reproductive age, whether White or Black, had excess body or facial terminal hair growth, defined by an mFG score of 3 or greater. Overall, it would appear that a stricter definition of what should be considered normal body or facial terminal hair growth in the general population of women is required. These data are consistent with our recent data indicating that up to 50% of patients who present with a complaint of unwanted hair growth and visually have minimal excess hairs suffer from an androgen excess disorder, notably PCOS (20). Finally, additional studies in other ethnic or racial groups are needed.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
C.M.D., K.S.W., and A.A.B. have nothing to declare; R.A. is a consultant for Procter and Gamble Pharmaceuticals and Merck and Co.
First Published Online January 31, 2006
Abbreviations: BMI, Body mass index; mFG, modified Ferriman-Gallwey; PCOS, polycystic ovary syndrome.
Received November 24, 2004.
Accepted January 20, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  W. A. March, V. M. Moore, K. J. Willson, D. I.W. Phillips, R. J. Norman, and M. J. Davies The prevalence of polycystic ovary syndrome in a community sample assessed under contrasting diagnostic criteria Hum. Reprod., November 12, 2009; (2009) dep399v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. O. Yildiz, S. Bolour, K. Woods, A. Moore, and R. Azziz Visually scoring hirsutism Hum. Reprod. Update, June 30, 2009; (2009) dmp024v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Koulouri and G. S Conway Management of hirsutism BMJ, March 27, 2009; 338(mar27_1): b847 - b847. [Full Text]
|
|
|
|
|
|
S. J. Chae, J. J. Kim, Y. M. Choi, K. R. Hwang, B. C. Jee, S. Y. Ku, C. S. Suh, S. H. Kim, J. G. Kim, and S. Y. Moon Clinical and biochemical characteristics of polycystic ovary syndrome in Korean women Hum. Reprod., August 1, 2008; 23(8): 1924 - 1931. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Does {blacktriangledown}eflornithine help women face hirsutism? DTB, August 1, 2007; 45(8): 62 - 64. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Azziz, E. Carmina, D. Dewailly, E. Diamanti-Kandarakis, H. F. Escobar-Morreale, W. Futterweit, O. E. Janssen, R. S. Legro, R. J. Norman, A. E. Taylor, et al. Criteria for Defining Polycystic Ovary Syndrome as a Predominantly Hyperandrogenic Syndrome: An Androgen Excess Society Guideline J. Clin. Endocrinol. Metab., November 1, 2006; 91(11): 4237 - 4245. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
