Prevalence of Male and Female Pattern Hair Loss in Maryborough

Prevalence of Male and Female Pattern Hair Loss in Maryborough DesmondC.C.GanandRodneyD.Sinclair Department of Dermatology, University of Melbourne, St Vincent s Hospital, Melbourne, Victoria, Australia Maryborough, in central Victoria has an approximate population of 8000 and census data is well matched for Australia overall. Australia has compulsory voting and registration on the electoral roll. To determine the agerelated prevalence of balding among men and women in Maryborough we conducted a postal survey of 5000 men and women aged 20 or older, and 427 were invited to attend for examination. Additional data was collected on dandruff, presence of gray hair. Supplementary questionnaires were sent to 340 children aged 5 9, attending a coeducational primary school. 1456 adults (34.1%) responded to the questionnaire. 396 attended for examination. The prevalence of androgenetic alopecia (AGA) increased with advancing age. 98.6% of men had bitemporal recession and severity was significantly associated with vertex and mid-frontal hair loss (po0.01) but not age (p ¼ 0.06). In all, 64.4% of women had bitemporal hair loss, and similar to men there was a significant association with mid-frontal hair loss (p ¼ 0.042) but not age (p ¼ 0.467). One hundred and forty children with completed questionnaires were examined. All 72 females and 68 males were assessed as stage 1 on the mid-line part and with no bitemporal recession (frequency stage 1 ¼ 100%, 95% CI (confidence interval) 97.4% 100%). A significant but weak positive association existed between presence of gray hair and history of dandruff (po0.01). The prevalence of mid-frontal hair loss increases with age and affects 57% of women and 73.5% of men aged 80 and over. Key words: alopecia/baldness/common/pattern/prevalence J Investig Dermatol Symp Proc 10:184 189, 2005 Hamilton first classified male pattern hair loss (MPHL) on a scale from I to VIII. (Hamilton, 1951). Norwood modified Hamilton s scale eliminating Type III, and including Type A Variant, which represents a more severe frontal recession form of hair loss, and the term vertex, which refers to cases with isolated balding patch on the crown were incorporated to the scales. (Norwood, 1975) Simplified versions of this scale have been used for population surveys (Severi et al, 2003). Ludwig noted that the clinical features of androgenetic alopecia in females differed from common baldness in men (Ludwig, 1977), and presents with diffuse hair loss, over the mid-frontal scalp that he classified into three stages. A modified five-point visual analog grading scale has also been used to score female pattern hair loss (FPHL) (Sinclair et al, 2002; Biondo et al, 2004) FPHL is now the preferred term for androgenetic alopecia in females. (Olsen, 2001). Crossover in the pattern of hair loss is not uncommon and the Ludwig pattern is observed in a proportion of men (Paik et al, 2001) and the Hamilton Norwood patterns with bitemporal recession is seen in some women. (Venning and Dawber, 1988). The prevalence of male and female pattern hair loss has been reported in UK, US, Norway, and Asia. (Norwood, Abbreviations: CI, confidence interval; CTE, chronic telogen effluvium; DHT, dihydrotestosterone; FPHL, female pattern hair loss; M, mean; MPHL, male pattern hair loss; N, total number of subjects; OR, odds ratio; PHL, pattern hair loss; SD, standard deviation; T:V, terminal hair to vellus-like hair ratio 1975; Venning and Dawber, 1988; Demuro-Mercon et al, 2000; Tang et al, 2000; Birch et al, 2001; Norwood, 2001; Paik et al, 2001; Pathomvanich et al, 2002) Apart from the MPHL prevalence studies in Norway (Demuro-Mercon et al, 2000) and Singapore (Tang et al, 2000), these studies recruited in the main from dermatology clinics and health care centres and it is unknown to what extent this sample group is representative of the general community. Further, aside from the Korean study (Paik et al, 2001), the pattern and severity of FPHL was not reported. The purpose of this study was to more precisely estimate the prevalence and pattern of hair loss in the Australian community using a different methodology. Additional information about gray hair and dandruff was recorded. Results Study population Seven hundred and thirty-one (14.6%) envelopes were sent back unopened or marked Return to Sender. These responses are regarded as either deceased, moved away from the address, or unable to be identified as having lived in the address and are excluded from further analysis. The response rate after exclusions was 34.1%. One thousand four hundred and fifty-six completed questionaires were received, 655 from men (mean age 57.6) and 752 from women (mean age 55.6). Forty nine did not report gender. Comparing the age distribution between respondents and the Maryborough population, there Copyright r 2005 by The Society for Investigative Dermatology, Inc. 184

10 : 3 DECEMBER 2005 PREVALENCE OF PATTERN HAIR LOSS 185 Table I. Age distribution of the respondents Maryborough population (2001) (%) Australian population (2001) (%) Respondents Age (%) 20 29 6.5 13.1 18.9 30 39 11.9 15.4 21.0 40 49 15.4 17.7 20.4 50 59 20.1 18.2 16.4 60 69 19.5 14.6 10.7 70 79 18.8 13.9 8.3 80 and above 7.7 7.1 4.3 was an over-representation in the older age groups among the respondents (Table I). All participants were invited to attend an examination. Of the 1456 subjects who completed the questionnaires, 575 (293 males and 282 females) expressed interest in attending hair and scalp examinations. The first 427 were given appointments and 396 participants (203 males and 193 females) turned up for the examinations. Secondary study of pre-pubertal children Two hundred and ten subjects (107 males and 103 females) did not respond. The response rate was 43.77%. All 140 children examined had hair pattern stage 1 (95% CI (confidence interval), 97.4% 100%). No child had temporal recession. One child was excluded from the study because of pubertal signs. In this cohort the normal baseline hair density is stage 1. Any adult or adolescent with a hair pattern of stage 2 or greater is likely to have experienced hair loss. Agreement between clinicians and subjects For both male and female subjects, 60% of the subjects self-ratings of hair thickness were identical to those of the clinicians. Where disconcordance was seen, only 15 male subjects (7.4%) and nine female subjects (4.6%) deviated by more than one grading. Therefore, subjects had difficulty obtaining precise grading but were capable of giving an approximate grading. Kappa statistic revealed that clinicians had a fair agreement with male subjects and a slight agreement with female subjects. Combination of stages 3 5 and also stages 1 and 2 to transform the five-point hair thickness scale into a binary tool improved the agreement considerably (Table II). This showed that the grading scale was an effective tool for self-reporting severe or cosmetically significant AGA. No significant differences were found between the mean (1.60) of females subjects self-ratings and the mean (1.62) of clinicians ratings (t (193) ¼ 0.487, p ¼ 0.627, two-tailed). Figure 1 Age-specific frequency of male pattern hair loss. A comparison of the age-specific frequency of male pattern hair loss in our cohort to that found in similar studies in different patient cohorts. Hair patterns and MPHL in males The age-adjusted prevalence of mid-frontal and vertex scalp hair loss (stages 3 5) was 44.9% (95% CI, 41.1% 48.8%). The prevalence increased with age (Fig 1). Stage 1 was universal among prepubertal boys but was reported by only 4.1% of men aged 80 or older. The age-specific frequency is shown in Table III. Hair patterns and FPHL in females Stage 1 is the normal female hair pattern and found universally among pre-pubertal girls. Most female subjects (66.1%) were stage 1. The frequency, however, of stage 1 decreased from 87.7% of subjects at third decade to 42.6% of subjects over age of 80. Stage 2 hair density increased from 8.77% of subjects at third decade to around 28% of subjects over 60 y of age. Stages 3 5 hair density were uncommon until the sixth decade, and affected over 25% of women aged 70 y and above (Table IV). If stage 2 is defined as mild hair loss and stages 3 5 as moderate to severe hair loss, the age-adjusted prevalence of hair loss in the community is 32.2% (95% CI, 28.8% 35.6%), of whom 10.5% (95% CI, 8.2% 12.7%) have moderate-to-severe hair loss (Fig 2). Only 55.0% of women with stage 3 5 reported history of hair loss. Only four female subjects (2.8% of those who perceived hair loss) had treatment for hair loss. Table II. Agreement between clinicians and subjects Identical grading (%) j statistic Interpretation of j values Original male scale 121/203 ¼ 59.6 0.494 Fair agreement Male two-point scale: stages 1, 2 and 3 5 177/203 ¼ 87.2 0.720 Good agreement Original female scale 115/193 ¼ 59.6 0.291 Slight agreement Female two-point scale: stages 1, 2 and 3 5 170/193 ¼ 88.1 0.540 Fair agreement

186 GAN AND SINCLAIR JID SYMPOSIUM PROCEEDINGS Table III. Hair patterns in male subjects Hair thickness Age Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Total Stages 3 þ 4 þ 5 5 9 68 (100%) 20 29 24 (69%) 5 (14%) 3 (9%) 3 (9%) 35 6 (17.14%) 30 39 34 (46%) 26 (35%) 7 (10%) 4 (5%) 3 (4%) 74 14 (18.9%) 40 49 28 (31%) 21 (24%) 13 (15%) 16 (18%) 11 (12%) 89 40 (44.9.%) 50 59 30 (24%) 33 (26%) 25 (20%) 25 (20%) 14 (11%) 127 64 (50.4%) 60 69 25 (18%) 26 (19%) 20 (14%) 39 (28%) 29 (21%) 139 88 (63.3%) 70 79 20 (16%) 24 (19%) 19 (15%) 33 (26%) 32 (25%) 128 84 (65.6%) 80 2 (4%) 11 (22.4%) 4 (8%) 16 (33%) 16 (33%) 49 36 (73.5%) Total 163 (25%) 146 (23%) 91 (14%) 136 (21%) 105 (16%) 641 332 (44.9% a ) a Adjusted to age. Graying of hair The age- and sex-adjusted prevalence of gray hair was 75.0% (95% CI, 72.7% 77.3%). The agespecific prevalence in males was 75.6%, (95% CI, 72.3% 78.9%) and in females was 74.8% (95% CI, 71.7%, 77.9%). There was a clear increase in prevalence of gray hair with advancing age (Fig 3). Dandruff After adjusting for age and sex, we estimate 42.2% (95% CI, 39.6% 44.8%) of the community are affected by dandruff at some stage of their life. A higher frequency was found in males 48.9% (95% CI, 45.0% 52.7%), compared to females 36.0% (95% CI, 32.6% 39.4%). No clear trend was observed between age and dandruff (Fig 4). Patterns of fronto-temporal hairline Additional information about the frontal hairline in the temporal regions was recorded during the examination. We categorized three types of temporal hairline: absence of bitemporal recession, slight bitemporal recession, and marked bitemporal recession (Fig 5). No pre-pubertal boy had temporal recession. Among adult males, marked bitemporal recession was the commonest temporal hairline (age-adjusted frequency ¼ 79.7%), followed by moderate (18.9%) and absence of recession (1.4%). An ordinal logistic regression analysis of bitemporal recession in men and age and severity of AGA found a significant association between bitmeporal recession and hair thickness (po0.001), but no association with age once hair thickness had been taken into account (p ¼ 0.0648). No pre-pubertal girl had temporal recession. In women, moderate bitemporal recession was the most frequent frontal hairline (age-adjusted frequency ¼ 55.2%), followed by absence of recession (35.6%) and marked recession (9.2%). As with men, ordinal logistic regression analysis found that age was not significantly associated with bitemporal recession (p ¼ 0.467), but that hair thickness was (p ¼ 0.042). Association analysis Logistic regression analysis of selfreported gray hair by self-reported history of dandruff and age showed a significant association between the two in both males and females (po0.001 for both dandruff and age, in both sexes). When this same association was Table IV. Hair patterns in female subjects Hair thickness Age Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Total 1 stage 1 5 9 72 (100%) 20 29 50 (88%) 5 (9%) 1 (2%) 1 (2%) 57 7 (12.3%) 30 39 73 (83%) 14 (16%) 1 (1%) 88 15 (17.0%) 40 49 91 (75%) 28 (23%) 3 (2%) 122 31 (25.4%) 50 59 106 (72%) 29 (20%) 11 (7%) 1 (1%) 147 41 (27.9%) 60 69 73 (59%) 37 (30%) 11 (9%) 2 (2%) 1 (1%) 124 51 (41.1%) 70 79 58 (46%) 35 (28%) 24 (19%) 6 (5%) 2 (2%) 125 67 (53.6%) 80 23 (43%) 15 (28%) 8 (15%) 8 (15%) 54 35 (57.4%) Total 474 (66%) 163 (23%) 55 (8%) 22 (3%) 3 (0.4%) 717 247 (32.2% a ) a Adjusted to age.

10 : 3 DECEMBER 2005 PREVALENCE OF PATTERN HAIR LOSS 187 Figure 4 Age-specific frequency of subjects with history of dandruff. Agespecific frequency of subjects with history of dandruff in our patient population. Figure 2 Hair patterns in the female subjects. The age-specific frequency of hair patterns and hair loss severity in the female subjects in our cohort. examined using only the clinicians examination data, a significant association was present only in females (p ¼ 0.046) but not in males (p ¼ 0.369). Age was again significant in both males and females (po0.001). Discussion This is an epidemiological study exploring hair patterns and FPHL prevalence in the Australian community. We previously reported the prevalence of vertex and full androgenetic hair loss in a different cohort of Australian Men aged 40 69, using a more limited grading scale. Following that study we recognized that a significant proportion of Australian men developed frontal recession as their predominant pattern of hair loss and modified our grading scale to include stage 4. In comparison to previous prevalence studies from the UK, US, Korea, and Thailand (Norwood, 1975; Venning and Dawber, 1988; Birch et al, 2001; Norwood, 2001; Paik et al, 2001; Pathomvanich et al, 2002), our study recruited subjects from a standard study population in Maryborough. The recruitment of our study was similar to the studies done in Norway (Demuro-Mercon et al, 2000) and Singapore (Tang et al, 2000), but the two studies reported only the epidemiology of MPHL. Using census data of Maryborough, we are able to calculate an estimated weighted prevalence, which is a more representative estimate of pattern hair loss prevalence in the general community. The limitations of this research include a response rate of 34%, a potential positive response bias, the limitations of self-reporting hair stages and an over-representation of older age groups in the respondents. Our study showed that MPHL is common in the Australian community. The prevalence correlates well with that reported by Norwood (1975), and supports the contention that MPHL in the western community had an earlier onset and a higher prevalence than that of their Asian counterparts. (Paik et al, 2001; Pathomvanich et al, 2002). There is no gold standard for diagnosis of FPHL particularly in the early stages. Scalp biopsy had been proposed as the gold standard for diagnosis (Whiting, 1990); however, hair follicle miniaturization cannot be demonstrated in all women with visible hair loss, and so these criteria may lack Figure 3 Age-specific frequency of gray hair. Age-specific frequency of any gray hair in our patient population. Figure 5 Classification of temporal hairline. Absence, moderate, and marked.

188 GAN AND SINCLAIR JID SYMPOSIUM PROCEEDINGS sensitivity (Sinclair et al, 2002). Furthermore, follicular miniaturization may be advanced before there is a visible decrease in hair density, especially among women with high baseline scalp hair follicle densities. As reported by Birch et al (2001) hair densities in normal women follow a normal distribution. Although women noted to have FPHL have a lower mean hair density, there is an extensive overlap of hair densities between women with and without FPHL. The finding that all pre-pubertal females have stage 1 hair density allows us to conclude that stage 2 is a deviation from the baseline pre-pubertal hair density and represents hair loss. Women with stage 2 hair loss should be regarded as having mild FPHL, even though this cannot always be demonstrated histologically. Extrapolation of our data indicates that a significant proportion of women with stage 2 hair loss would progress to more severe FPHL. Nevertheless, this may only be proved with a longitudinal study. Application of this yields a much higher estimation of female AGA prevalence. As 55% of women with marked FPHL (stages 3, 4, or 5) did not report hair loss on the questionnaires, it is likely that many women with FPHL will not deem it cosmetically significant or seek medical intervention. In contrast to what was reported by Birch and Messenger (2003), we found that bitemporal recessive frontal hairline is significantly associated with advancing stages of mid-scalp hair loss in both males and females, independent of age. This finding suggests that the pathogenesis of hair loss in both midscalp and bitemporal frontal hairline is probably related, if not the same. Of interest, a history of dandruff has a consistent but weak association with the presence of gray hair in females. The association in males is only significant in the full-cohort self-report data. There are no substantial differences in the numbers of females as compared to males in the full-cohort, or having received clinical examinations, and so this does not explain the differences seen. The explanation may be because of the difference in reporting dandruff. In the questionnaire, subjects are asked about any history of dandruff or dandruff, whereas clinicians are only able to assess any presence of active dandruff. This association has not been previously reported. As Malassezia organisms are involved in both dandruff and pityriasis versicolor which produces depigmentation of the skin (Ljubojevic et al, 2002), we postulate that the Malassezia organisms may act as the mechanistic link between graying of hair and dandruff. As the association is only weakly predictive, however, many other variables are likely to affect presence of gray hair in an individual. Further research is required to explore this association. Sampling frame The sampling frame for this study was adults aged 20 y and above registered on the 2000 Electoral Roll in the city of Maryborough (6400 adults). Maryborough is a rural city in Central Victoria, 100 km north of Melbourne, Australia. Although slightly over-represented in the older age groups, the distribution of age and gender in the population of Maryborough is similar to that of Australia. Previous epidemiological studies have been conducted in this town and have received great support from the local community (Foley et al, 1993; Plunkett et al, 1999). Five thousand adults were randomly selected from the sampling frame. The survey period was February 2003 to May 2003. A pilot study comprising 300 subjects randomly selected from outside the City of Maryborough was conducted prior to the commencement of the survey. Each subject in pilot study was posted questionnaire similar to that used in the main survey. Responses were collected to evaluate questionnaire design and indicate response rate for estimation of sample size in the main study. Recruitment Permission to conduct the survey was obtained from the Shire Council responsible for the City of Maryborough. A notice regarding the survey was published in the local newspaper. Local GP were contacted to prepare them for queries from their patients regarding the survey. A questionnaire with an explanatory letter was posted to each of the 5000 subjects. To maximize response rate, reminders were posted to non-respondents 2 wk after the initial posting. Questionnaire All subjects were asked to complete the questionnaire containing information about history of hair loss, specific forms of hair diseases and treatment, information of presence of gray hair and history of dandruff. More importantly, they were asked to self-rate hair-thickness using the sex-specific grading scales provided (Figs 6 and 7). Examination In order to measure agreement between subject and clinician reporting and to record additional information on frontal hairline, a random sample of 427 subjects who expressed interest on the survey were examined by a consultant dermatologist or dermatology registrar familiar with the hair thickness scales used and blinded to the results of the questionnaires. Hair thickness, hair pattern, dandruff, gray hair, solar keratoses, and skin cancers were recorded. Secondary study of pre-pubertal children To assess mid-frontal scalp hair thickness in pre pubertal children, to determine baseline hair density prior to the possible onset of pattern hair loss, questionnaires and consent forms were sent to the parents of the 350 children at Carey Baptist Grammar School in Kew Victoria. Children aged from 5 10 y were invited to undergo visual examination of their scalp to score the presence and severity of AGA. Examinations were conducted by a consultant dermatologist, a pediatrician, and a research student. Hair wave and color, which might affect the appearance of the mid-frontal hair density, were also scored. Data analysis Questionnaire results were compared to examination results and the k statistic for observed agreement between clinicians ratings and subjects self-ratings was calculated. Descriptive statistics on hair thickness, hair pattern, gray hair and dandruff were analyzed from the questionnaires. Subjects reporting hair loss related to other diagnoses such as alopecia areata, thyroid disease, chemo/radiotherapy etc were excluded from analysis. Age-specific weighting of prevalence estimates was undertaken using the 2001 census data for Maryborough. The 95% CI were calculated for the weighted prevalence estimates with the weighting variable applied. Methods Figure 6 Male hair pattern grading scale. Stages 1 5.

10 : 3 DECEMBER 2005 PREVALENCE OF PATTERN HAIR LOSS 189 Ordinal regression analysis was performed to explore association of frontal hairline with hair thickness and age using clinicians examination data (dependent variable ¼ frontal hairline). Logistic regressions were performed to explore association of grayness and dandruff firstly using self-reported data in the questionnaires, and then clinicians examination data. (dependent variable ¼ grayness; dandruff and grayness both dichotomized) The following people provided assistance throughout the study: Mr Bernie Waixel and members of the Maryborough District Health Service, members of the Maryborough local publishers, Dr Shannon Harrison, Dr Alex Chamberlain, Dr Julia Hyun, Dr Michael Starr, Uma Ramadass, Sem Liew, Adam Chapman and staff members of the Department of Dermatology of St Vincent s Hospital Melbourne. Approved by St Vincent s Hospital Research and Ethics Committee (Reference number 123/02). DOI: 10.1111/j.1087-0024.2005.10102.x Manuscript received September 20, 2004; revised November 18, 2004; accepted for publication December 2, 2004 Address correspondence to: MD, Professor Rodney D. Sinclair, FACD, St Vincent s Hospital, 41 Victoria Parade, Fitzroy, 3065, Melbourne, Victoria, Australia. Email: sinclair@svhm.org.au References Figure 7 Female hair pattern grading scale. Stages 1 5. Biondo S, Goble D, Sinclair R: Women who present with female pattern hair loss tend to underestimate the severity of their hair loss. Br J Dermatol 150:750 752, 2004 Birch MP, Messenger AG: Frontal recession in women: No correlation with hair density. Exp Dermatol 12:231, 2003 Birch MP, Messenger JF, Messenger AG: Hair density, hair diameter and the prevalence of female pattern hair loss [comment]. Br J Dermatol 144:297 304, 2001 DeMuro-Mercon C, Rhodes T, Girman CJ, Vatten L: Male-pattern hair loss in Norwegian men: A community-based study. Dermatology 200:219 222, 2000 Foley P, Nixon R, Marks R, Frowen K, Thompson S: The frequency of reactions to sunscreens: Results of a longitudinal population-based study on the regular use of sunscreens in Australia. Br J Dermatol 128:512 518, 1993 Hamilton JB: Patterned loss of hair in man: Types and incidence. Ann NY Acad Sci 53:708 728, 1951 Ljubojevic S, Skerlev M, Lipozencic J, Basta-Juzbasic A: The role of Malassezia furfur in dermatology. Clin Dermatol 20:179 182, 2002 Ludwig E: Classification of the types of androgenetic alopecia (common baldness) occurring in the female sex. Br J Dermatol 97:247 254, 1977 Norwood O: Incidence of female androgenetic alopecia (female pattern alopecia). Dermatol Surg 27:53 54, 2001 Norwood OT: Male pattern baldness: Classification and incidence. South Med J 68:1359 1365, 1975 Olsen EA: Female pattern hair loss. J Am Acad Dermatol 45:S70 S80, 2001 Paik JH, Yoon JB, Sim WY, Kim BS, Kim NI: The prevalence and types of androgenetic alopecia in Korean men and women. Br J Dermatol 145:95 99, 2001 Pathomvanich D, Pongratananukul S, Thienthaworn P, Manoshai S: A random study of Asian male androgenetic alopecia in Bangkok, Thailand. Dermatol Surg 28:804 807, 2002 Plunkett A, Merlin K, Gill D, Zuo Y, Jolley D, Marks R: The frequency of common nonmalignant skin conditions in adults in central Victoria, Australia. Int J Dermatol 38:901 908, 1999 Severi G, Sinclair R, Hopper JL, et al: Androgenetic alopecia in men aged 40-69 years: Prevalence and risk factors. Br J Dermatol 149:1207 1213, 2003 Sinclair R, Jolley D, Mallari R, Magee J: The reliability of horizontally sectioned scalp biopsies in the diagnosis of chronic diffuse telogen hair loss in women. J Am Acad Dermatol 51:189 199, 2002 Tang PH, Chia HP, Cheong LL, Koh D: A community study of male androgenetic alopecia in Bishan, Singapore. Sing Med J 41:202 205, 2000 Venning VA, Dawber RP: Patterned androgenic alopecia in women. J Am Acad Dermatol 18:1073 1077, 1988 Whiting DA: The value of horizontal sections of scalp biopsies. J Cutan Aging Cosmet Dermatol 1:165 173, 1990