A twin concordance study of trichotillomania†
How to Cite this Article: Novak CE, Keuthen NJ, Stewart SE, Pauls DL. 2009. A Twin Concordance Study of Trichotillomania. Am J Med Genet Part B 150B:944–949.
Abstract
Trichotillomania (TTM) is a disorder with putative genetic underpinnings. Family studies report higher than expected rates of TTM among relatives of affected individuals, but no twin concordance studies have been completed to estimate heritability rates. Same-sex twin pairs with hair pulling in at least one co-twin were included. Subjects were recruited following phone screens and questionnaire completion for zygosity and hair pulling variables. Three sets of criteria were used to define hair pulling and TTM. Two other sets of criteria were widened to include skin picking and bothersome hair manipulation. Fisher exact tests assessed pairwise concordance rates for monozygotic and dizygotic twin pairs and heritability estimates were calculated where significant differences existed. Among 34 identified twin pairs, 24 were monozygotic (MZ) and 10 were dizygotic (DZ). Respective concordance rates for MZ and DZ twin pairs were significantly different at 38.1% and 0% for DSM-IV TTM criteria, 39.1% and 0% using modified DSM criteria, and 58.3% and 20% for noticeable non-cosmetic hair pulling (heritability estimates 76.2%). MZ and DZ concordance rates were not significantly different when broadening hair pulling criteria to include skin picking or when including bothersome hair manipulation. Concordance rates from this study suggest that genetic factors play a significant role in the etiology of TTM. Given the reported discordance rates among the MZ twins, further research is required to fully understand contributory non-genetic factors. © 2009 Wiley-Liss, Inc.
INTRODUCTION
Trichotillomania (TTM) can affect between 1% and 3% of the population depending on the stringency of diagnostic criteria used [Christenson et al., 1991a; Rothbaum et al., 1993]. It is characterized by the presence of chronic, repetitive or compulsive hair pulling resulting in noticeable hair loss that is problematic and distressing to the sufferer. TTM is a psychologically devastating disease with frequent physical complications. It has characteristics that overlap with neuropsychiatric illnesses such as Tourette disorder (TD) and obsessive–compulsive disorder (OCD). The extent to which TD and OCD are genetically inherited has been extensively studied [Pauls and Leckman, 1986; Nestadt et al., 2000; Rosario-Campos et al., 2005] although similar work has yet to be done with TTM.
Preliminary evidence suggests that TTM is a familial and potentially genetically mediated illness. Multiple case reports of familial hair pulling have been published [Sanderson and Hall-Smith, 1970; Galski, 1983; Kerbeshian and Burd, 1991]. Increased rates of hair pulling have been reported among relatives of probands with TTM [Swedo and Rapoport, 1991; Christenson et al., 1992]. Two studies also reported higher than expected rates of TTM (5%) among first-degree relatives of TTM probands [Swedo and Rapoport, 1991; Schlosser et al., 1994]. Furthermore, indirect evidence for a genetic association with OCD, TD and/or other “OCD spectrum” disorders has been identified in family studies [Lenane et al., 1992; King et al., 1995; Bienvenu et al., 2000].
Twin concordance studies are frequently used to determine whether observed familiality of a disorder is attributable to influences of genetics, common environmental factors or both. In twin studies, an initial suggestion of the relative etiologic importance of genetic, shared and non-shared environmental contributors is obtained by examining MZ and DZ twin resemblance. MZ twins share 100% of their genetic material while DZ twins share on average 50%. Given this difference, it is expected that if genetic influences on twin resemblance are important, MZ concordance should be significantly higher than DZ concordance. Alternatively, in diseases for which shared environmental factors (e.g., a parent modeling hair pulling behavior to a child) but no genetic factors impact on phenotypic variance, MZ concordance rates would be expected to equal DZ concordance rates. Lastly, non-shared environmental influences may act together with both genetic and shared environmental influences to impact on phenotype.
Twin concordance studies of TTM have not been conducted to date. In the only study reporting on TTM twins, SPECT scans were conducted on an MZ twin pair concordant for TTM [Vythilingum et al., 2002]; perfusion defects were larger in the more severely affected twin. Among related disorders, Ooki et al. studied genetic effects on nail-biting and thumb-sucking. They reported that the percentage of phenotypic variance attributable to genetic influences was 50% for nail-biting regardless of gender, and was 66% in males and 50% in females for finger-sucking [Ooki, 2005].
Finally, molecular genetic research on both animals and humans has begun to examine candidate genes for TTM and related disorders. Mice with a homozygous loss of function mutation in the Hoxb8 gene demonstrated excessive pathological grooming [Greer and Capecchi, 2002]. More recently, SAPAP3 knock-out mice demonstrated a 100% penetrant phenotype of pathologic overgrooming and anxiety, which was rescued by restoring gene function [Sestan and State, 2005; Welch et al., 2007]. In human studies, genotype distributions of the 5-HT2A receptor T102C variant differed significantly between TTM-affected individuals and controls [Hemmings et al., 2006]. Finally, a frameshift mutation and rare occurrence of Slit and Trk-like 1 (SLITRK1) was found in 3 of 174 unrelated subjects with TD [Abelson et al., 2005]. One subject's mother with the mutation had TTM but not TD. This is of particular interest as TD has been described as most phenotypically and biologically related to TTM [Price et al., 1985].
Thus, current evidence suggests that TTM and related disorders have familial and potential genetic underpinnings. This study examines estimated heritability of TTM and less strictly defined hair pulling. It was hypothesized that the TTM concordance rate among MZ twins would significantly exceed that among DZ twins.
Given the overlapping phenomenology and high comorbidity between TTM and pathological skin picking [Christenson et al., 1991b; Lochner et al., 2002; Odlaug and Grant, 2008a,b] it was hypothesized that they represent an alternate phenotype of the same heritable condition. Thus, it was expected that adding pathological skin picking to the phenotype analysis would result in similar concordance rate differences between MZ and DZ twins. On the other hand, hair twisting, twirling and other manipulation is common in the general population and is usually non-pathological even if annoying or bothersome to an individual or observer. In contrast to skin picking, it has not been thought to be part of the broader TTM phenotype, thus it was hypothesized that such hair manipulation would show no concordance differences when the TTM phenotype was broadened to include them.
MATERIALS AND METHODS
Participants
Study volunteers were recruited by a direct mailing to members of the Trichotillomania Learning Center (TLC), a study posting on the TLC website (www.trich.org), and an advertisement in the Mid-Atlantic Twin Registry newsletter. To minimize the ascertainment bias towards recruiting twins who are monozygotic and/or concordant, care was taken in wording the study advertisements to state, “It is important for us to learn from both identical and non-identical twins so we need both types of twin pairs to volunteer. It is also important that twins who don't both have trichotillomania volunteer for the study.” Study approval was obtained from the HealthPartners (Minnesota) Institutional Review Board prior to initiation of recruitment efforts.
Hair pullers with same-sex twins volunteered or were identified by their twin as possible participants. Study inclusion criteria included: same-sex twin pairs with self-report of non-cosmetic hair pulling in one or both co-twins, hair pulling not attributable to dermatological etiologies, and participant age ≥18 years old. Both co-twins had to express a willingness both to (1) complete a questionnaire and several self-report instruments and (2) to contribute a cheek swab for DNA if there was uncertainty regarding zygosity. Given the known female preponderance in this disorder [Christenson et al., 1991b], only same-sex twin pairs were recruited to eliminate the potential for artificial lowering of concordance rates in opposite-sex twin pairs. Similarly, given typical onset during early adolescence [Christenson et al., 1991b], pairs less than 18 years of age were excluded to minimize false negative results attributable to variability in hair pulling symptom onset.
Study Procedures
After an initial phone screen to ascertain satisfaction of study inclusion criteria, twin pairs were mailed a study packet including the Zygosity Questionnaire [ZQ; Torgersen, 1979], the Massachusetts General Hospital Hair Pulling Scale [MGH-HPS; Keuthen et al., 1995], the Trichotillomania Impact Scale [TIS; O'Sullivan et al., 1994] and a questionnaire developed for the purposes of this study (described below). Participants were instructed to complete study instruments independently and anonymously from their co-twin to avoid bias in self-report. Completion of all study measures took approximately 60–90 min. Written informed consent was obtained from all participants. The consent form was mailed with the questionnaire packet.
Study Instruments
The study questionnaire developed specifically for this study was designed to assess demographics, medical history, diagnostic criteria for TTM, hair pulling history, past treatment, family history and presence of other body-focused repetitive behaviors (e.g., skin picking).
The ZQ is a 5-item instrument designed for zygosity ascertainment in twin pairs with questions assessing eye and hair color and perceived similarity in appearance. It has been demonstrated to be 96–98% accurate in zygosity diagnosis [Torgersen, 1979; Magnus et al., 1983; Reed et al., 2005].
The MGH-HPS is a self-report instrument for the assessment of hair pulling severity in adults. It consists of seven items including frequency and intensity of urges, ability to control urges, frequency of hair pulling, resistance to and control over hair pulling, and associated distress. Items are rated on a severity scale ranging from 0 to 4. It is a scale with good internal consistency (89%) [Keuthen et al., 1995]. Scale evaluation with an independent sample documents its test–retest reliability (97%), convergent and divergent validity, and sensitivity to change in symptoms [O'Sullivan et al., 1995].
The TIS is a 28-item self-report scale that assesses emotional and psychosocial consequences to pulling. Item severity ratings range from 0 (“none”) to 5 (“severe”).
Classification Criteria
Five different classification criteria sets were utilized to assess twin pair concordance for hair pulling and skin picking. The most conservative classification definition, Criteria Set 1, assessed for a diagnosis of TTM and was identical to that of DSM-IV. This required a response of “yes” to the question “Do you pull out your hair repeatedly to the point of having noticeable hair loss?” Or if they responded “no” to the above question, they needed to endorse a rating of “1” (mild hair loss) or higher in one or more of ten listed body areas. It also required a response of “yes” to “Do you experience an increasing sense of tension immediately before pulling out the hair or when you attempt to resist the behavior?” In addition, participants were required to endorse a score of “2” or higher (range = 0–5) on at least one of the three TIS items which assess completeness, satisfaction, or pleasure upon pulling (e.g., “I get a sense of completeness from hair pulling.”). Thirdly, a negative response was required to the question “Can your hair pulling be accounted for by another existing mental disorder (besides trichotillomania) or by a general medical condition, for example, skin inflammation or itching?” Where a positive response was given, it was revised to a negative one if the subject clarified in writing or by phone their answer to mean anxiety or other common mental condition. Finally, participants had to endorse distress or impairment associated with pulling by endorsing one of the following three criteria: (1) “yes” to the question “Does your hair pulling cause you significant distress or impairment in social, occupational, or other important areas of functioning?” or (2) a score of “3” (moderate) or higher on the questionnaire item “Please rate the extent to which hair pulling has interfered with each of the following areas of your life” for any one of several domains including family relationships, social relationships outside of your family, job, career or school work, or daily life (dressing, eating, driving, or bathing); or (3) a score of “3” (moderately) or higher on either question “How much have you suffered emotionally because of your hair pulling?” or “Considering all aspects of your life, how much do you consider your hair pulling to be a problem?”
The next classification scheme (Criteria Set 2) identified participants with clinically significant hair pulling using modified DSM-IV criteria that retained the essential components of Criteria Set 1 in that the hair pulling is noticeable and distressing (first and last criteria of Set 1), yet endorsed only tension prior to pulling or when attempting to resist OR completeness, satisfaction or pleasure upon pulling (second or third criterion of Set 1, not both). This relaxation of the DSM-IV definition was made due to observations that a significant proportion of individuals with the clinical characteristics of trichotillomania do not have both tension before or relief/gratification after pulling hair [e.g., 17% in Christenson et al., 1991b].
Criteria Set 3 assessed non-cosmetic hair pulling. Participants had only to endorse hair pulling currently or in the past and not for the purposes of grooming (i.e., eyebrow hairs).
Criteria Set 4 assessed the presence of skin picking, hair pulling or both (one of each twin pair met criteria for non-cosmetic hair pulling). Participants had to endorse either Criteria Set 3 for non-cosmetic hair pulling OR endorsed Skin/Acne/Scab Picking as a “nervous” habit or “compulsive” behavior and a score of “3” (moderately) or higher in response to the question “To what degree has skin picking bothered you overall?”
Lastly, Criteria Set 5 assessed bothersome hair manipulation. To satisfy these criteria, participants had to endorse “yes” to ever having engaged in one or more of several identified behaviors (including rubbing hair, twirling bunches of hair around your finger, twisting hair strands, picking or scratching hair areas, tugging at strands or tufts of hair, feeling/searching for certain hairs, cutting hair excessively such as to even it out or get rid of unwanted hairs, chewing on hair or playing with hair in any other way) “to the point where it bothered or annoyed you or people around you.”
Data Analysis
Descriptive analyses were completed with respect to demographic variables and measurement scale responses. Fisher exact tests were used to assess pairwise concordance rates for monozygotic and dizygotic twin pairs for all five classification criteria sets defined above. In contrast to a proband-wise method of calculating concordance, each twin pair was counted only once. Comparing the concordance rates between MZ and DZ twins allowed a determination of genetic versus environmental influences on the development of TTM. A significance threshold of P < 0.05 was used. Heritability estimates were calculated for each diagnostic threshold for which MZ–DZ rates were statistically distinct, by doubling the difference between concordance rates for MZ and DZ twin pairs.
RESULTS
One hundred forty-one individuals (70 hair pullers with same-sex twins and 1 hair puller whose twin was deceased) were identified where at least one of the twins initially volunteered to participate in the study. Subsequently, 24 volunteers and/or their co-twins declined to participate; 6 volunteers' co-twins did not return their study questionnaires and in 4 cases the co-twin returned the questionnaire whereas the volunteering twin did not. Thus, we initially had 70 twin pairs (we did not include the one with a deceased co-twin) of whom 24 declined to participate and 10 did not complete all of the questionnaires. Thus, 36 twin pairs were subsequently enrolled in the study. However, data from 1 twin pair was excluded given disagreement in co-twin responses on the ZQ. Follow-up phone calls to each twin and their parents failed to resolve response differences and one co-twin subsequently refused genetic testing to ascertain zygosity. Another twin pair was excluded as neither co-twin endorsed noticeable hair loss, distress due to pulling, or associated functional impairment.
The final sample of study completers consisted of 33 female twin pairs and 1 male twin pair. Twenty-four twin pairs were monozygotic and 10 twin pairs were dizygotic as identified by ZQ responses. The final study sample consisted of 66 (97.1%) females and 2 (2.9%) males. Mean participant age was 35.28 years (SD = 11.89; range = 19–71). Mean age of hair pulling onset (n = 43) was 12.49 years (SD = 3.00; range = 6–19). Lifetime endorsement of other “nervous” habits or compulsive behaviors included the following: nail-biting (n = 47), cuticle picking/biting (n = 39), skin/acne/scab picking (n = 41), lip or cheek picking/biting (n = 35), nose picking (n = 29), cracking knuckles (n = 31) and thumb sucking (n = 19).
The mean MGH-HPS total score indicated moderate current hair pulling severity in our study sample (n = 55, mean = 12.8, SD = 8.1, range = 0–24). Responses to each item of the MGH-HPS are reported in Table I. The mean TIS total score indicated overall mild psychosocial impact secondary to hair pulling (mean = 44.4, SD = 35.2, range = 0–114). Location of hair pulling is summarized in Table II. The most commonly pulled sites by report included the scalp (N = 37), pubic area (N = 25), eyebrows (N = 21) and eyelashes (N = 17) for which 73.8%, 10.2%, 25.1%, and 28.4% of pullers extracted hairs in the previous 3 months. Least commonly reported sites of pulling included underarms (N = 2) and chest (N = 2), for which 8% and 3.5% of pullers extracted hairs in the previous 3 months.
| % (N) | |||||
|---|---|---|---|---|---|
| Rating = 0 | Rating = 1 | Rating = 2 | Rating = 3 | Rating = 4 | |
| Urge frequency | 20.7% (12) | 19.0% (11) | 19.0% (11) | 34.5% (20) | 6.9% (4) |
| Urge intensity | 20.7% (12) | 17.2% (10) | 24.1% (14) | 32.8% (19) | 5.2% (3) |
| Urge control | 22.4% (13) | 15.5% (9) | 24.1% (14) | 27.6% (16) | 10.3% (6) |
| Hair pulling frequency | 24.1% (13) | 27.6% (9) | 17.2% (14) | 27.6% (16) | 3.4% (6) |
| Hair pulling resistance | 22.4% (13) | 25.9% (15) | 27.6% (16) | 19.0% (11) | 5.2% (3) |
| Hair pulling control | 23.2% (13) | 12.5% (7) | 12.5% (7) | 28.6% (16) | 23.2% (13) |
| Associated distress | 29.8% (17) | 21.1% (12) | 17.5% (10) | 22.8% (13) | 8.8% (5) |
- Scale items are rated from 0 to 4 with higher scores reflecting greater severity.
| Site of hair pulling | Subjects pulling from this site | Total site-specific hairs pulled/Total body hairs pulled in the past 3 months | |
|---|---|---|---|
| N | Mean % | SD (range) | |
| Scalp | 37 | 73.8% | 30.6 (1–100) |
| Brows | 21 | 25.1% | 25.5 (0–80) |
| Lashes | 17 | 28.4% | 27.1 (1–99) |
| Pubic | 25 | 10.2% | 10.3 (1–50) |
| Arms | 3 | 10.7% | 12.5 (2–25) |
| Underarms | 2 | 8.0% | 9.9 (1–15) |
| Legs | 6 | 4.8% | 5.1 (1–15) |
| Beard | 5 | 38.6% | 49.3 (1–95) |
| Mustache | 3 | 6.3% | 4.7 (1–10) |
| Chest | 2 | 3.5% | 2.1 (2–5) |
| Other | 5 | 10.0% | 3.5 (5–15) |
Significant differences in MZ and DZ concordance rates were observed for three of the five diagnostic schemes (DSM-IV (0.381 vs. 0.00, P = 0.047), Modified DSM (0.391 vs. 0.21, P = 0.021) and non-cosmetic hair pulling (0.583 vs. 0.200, P = 0.046)) yielding heritability estimates ranging from 0.76 to 0.78 (see Table III).
| TTM and Related Syndrome Definition | Monozygotic (MZ) pairs | Dizygotic (DZ) pairs | MZ versus DZ | Heritability estimate | ||
|---|---|---|---|---|---|---|
| N | Concordance | N | Concordance | P-value | ||
| 1: DSM-IV defined TTM | 21 | 0.381 | 8 | 0 | 0.047 | 76% |
| 2: Modified DSM defined TTM | 23 | 0.391 | 10 | 0 | 0.021 | 78% |
| 3: Non-cosmetic hair pulling | 24 | 0.583 | 10 | 0.20 | 0.046 | 76.6% |
| 4: HP and/or skin picking | 24 | 0.750 | 10 | 0.40 | 0.062 | n/a |
| 5: Bothersome hair manipulation | 24 | 0.708 | 10 | 0.50 | 0.221 | n/a |
- DZ, dizygotic; MZ, monozygotic; TTM, trichotillomania; n/a, not-applicable.
DISCUSSION
These results suggest that a significant proportion of TTM phenotypic variation is accounted for by genetic factors. Of note is that the concordance rates for the phenotypes that included skin picking and hair manipulation were not significantly different, suggesting that they may not be influenced by the same etiologic factors responsible for hair pulling. However, the fact that concordance among MZ pairs in this study was not equal to 100% indicates that non-genetic variables also significantly influence the onset of TTM and hair pulling.
There has been some debate regarding the validity of current diagnostic criteria for DSM-IV defined TTM. It has been suggested by clinicians that not all TTM-affected individuals pull to the point of noticeable hair loss or have an increasing sense of tension prior to pulling. Yet these individuals appear similar to classic TTM patients from all other perspectives. For this reason, the present study examined several diagnostic thresholds to more fully examine the hypothesis that TTM has genetic underpinnings. Findings from this study support the supposition that broader diagnostic criteria than those described by the DSM-IV may be used in future diagnostic studies. The DSM-IV threshold, described by Criteria Set 1 in this study, resulted in very similar heritability estimates to those for Criteria Sets 2 and 3. Criteria 2 loosens the requirement for both tension prior to pulling or when attempting to resist OR completeness, satisfaction or pleasure upon pulling. Further, Criteria Set 3 does not require the presence of associated distress or impairment.
The lack of significant differences between MZ and DZ concordance rates when replacing hair pulling criteria with hair manipulation or skin picking suggests that these are not alternate phenotypes reflecting common TTM genetic influence. For hair pulling Criteria Sets 1–3, the MZ concordance rates were greater than double the DZ concordance rates, implicating the presence of genetic and environmental effects on the phenotype. For Criteria Sets 4 and 5 for skinpicking and hair manipulation, as there was no significant difference between MZ and DZ concordance rates the potential effect of genetic influence appears to be much less appreciable, if present at all.
Several efforts were made in this study to ensure conservative heritability estimates and a rigorous methodology. Broad-sense heritability estimates were reported, which reflect the portion accounted for by all potential genetic influences. These influences may include factors such as gene dominance and epistasis. In comparison, narrow-sense heritability would have only estimated the genetic influence that is passed from parents to offspring. Further, this study controlled for potential divergent sex effects on environment by requiring that all twin pairs be same-sex. This may not have been necessary since empirical evidence supports the assumption of an equal environment for both male [Xian et al., 2000] and female [Kendler, 1993] twin pairs. However, there may be different underlying genetic contributions so that one might actually expect sex differences because of different genetic influence between males and females (e.g., hormones).
Limitations of this study must be acknowledged. There is a possibility of ascertainment bias, given that the study sample was recruited via one twin with hair pulling. These twin pairs may have distinct qualities related to severity or familiality that differentiates them from general population samples. In addition, assessment by questionnaire rather than clinical interview is not ideal, though hair pulling and skin picking are relatively objective behaviors and so lend themselves to this study method. Another limitation is the small sample size. The limited sample size did not permit the use of structural equation modeling to calculate more precise heritability rates rather than the estimates provided herein. However, despite this limited power to detect differences, MZ and DZ concordance rates differed significantly when using DSM-IV, modified-DSM-IV TTM and non-cosmetic hair pulling. Finally, comparing concordance rates between MZ and DZ twins to determine heritability is based on the assumption that MZ twins do not share more environmental similarity than DZ twins. This assumption is difficult to assess based on potential biases related to recall.
In conclusion, concordance rates in this twin study indicate significant genetic influences on the development of TTM and non-cosmetic hair pulling. Family and molecular genetic studies are required to clarify inheritance patterns and neuropathologic mechanisms for TTM. Potential interactions between genetic, shared and non-shared environmental influences on this disorder also require further elucidation. This may guide earlier detection and intervention and novel treatment development for this debilitating illness.
Acknowledgements
This work was supported by grants from the HealthPartners Research Foundation, Minneapolis, MN (CEN) and the Harvard Scholar in Medicine award (SES). Authors would like to thank the Trichotillomania Learning Center for their assistance in recruitment, and Casey Walsh, M.S.W. and Dianne Hezel, B.A. for their assistance in manuscript preparation.
