DEPRESSION AND ANXIETY 32:728–736 (2015)
Research Article FAMILIAL PATTERNS OF HOARDING SYMPTOMS Gail Steketee, Ph.D.,1 ∗ Andrea A. Kelley, M.S.W., M.P.H.,1 Jeremy A. Wernick, B.A.,1 Jordana Muroff, Ph.D.,1 Randy O. Frost, Ph.D.,2 and David F. Tolin, Ph.D.3
Background: Previous research suggests that hoarding aggregates in families and is associated with health and safety risks and family problems. The present study examined gender- and diagnosis-related differences in reports of hoarding symptoms among first-degree relatives of people who hoard, and of clinical and community samples. Methods: The present study included 443 participants in a study of hoarding behavior: 217 with hoarding disorder (HD), 96 with obsessivecompulsive disorder (OCD), and 130 nonclinical community controls (CC). Assessment included a detailed interview of familial patterns of hoarding behaviors among parents and siblings and measures of hoarding severity. Results: In the combined sample, participants reported more hoarding among female (mothers, sisters) than male (fathers, brothers) relatives. Significantly more female than male participants indicated they had a parent or any first-degree relative with hoarding behaviors. However, within the HD sample no significant gender effects were found for household, safety, and functioning variables, or for hoarding symptom severity. In an age- and gender-matched subsample (total n = 150), HD participants reported more hallmark hoarding symptoms (difficulty discarding and saving/clutter), and acquiring among their relatives compared to OCD and CC samples, and parents had higher rates than siblings. Conclusions: Hoarding symptoms appear to be common among first-degree relatives of people who hoard and are also found among relatives of control samples. The predominance of hoarding symptoms among female relatives may indicate genetic or modeling transmission but this requires further study using large twin samples. Clinicians should consider that family members may also have signifi C 2015 cant hoarding symptoms. Depression and Anxiety 32:728–736, 2015. Wiley Periodicals, Inc.
Key words: hoarding, family/marital, OCD; gender, gene/environment
FAMILIAL PATTERNS IN HOARDING DISORDER
1 Boston
University, School of Social Work, Boston, MA Department of Psychology, Northampton, MA 3 Anxiety Disorders Center, Institute of Living, Hartford, CT and Department of Psychiatry, Yale University School of Medicine, New Haven, CT 2 Smith College,
Contract grant sponsor: NIMH; Contract grant numbers: R01 MH068008 and R01 MH068007. ∗ Correspondence
to: Gail Steketee, School of Social Work, Boston University, 264 Bay State Rd., Boston, MA 02215. E-mail: [email protected] Received for publication 20 October 2014; Revised 6 June 2015; Accepted 9 June 2015 DOI 10.1002/da.22393
C 2015 Wiley Periodicals, Inc.
Hoarding disorder (HD) is a debilitating psychiatric
problem that can lead to considerable health risks, functional impairment, family conflict, and substantial financial burden for sufferers, family members, and the community.[1–3] Recently included in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5),[4] HD is characterized by excessive saving and difficulty discarding possessions, resulting in substantial clutter in active living areas; additional features include excessive acquisition and varying levels of insight. Prevalence estimates for HD range from 2 to 5%
Published online 30 June 2015 in Wiley Online Library (wileyonlinelibrary.com).
Research Article: Familial Patterns of Hoarding
among adults in the United States and other western countries.[5, 6] The few studies that have examined gender differences in hoarding prevalence and associated features have reported mixed results. Although women predominate in clinical HD samples, it is unclear whether hoarding affects women more than men.[6, 7] Samuels et al.[6] found that men and women with hoarding differed in obsessivecompulsive disorder (OCD) subtypes and comorbid disorders, but because they recruited from an OCD sample, it is difficult to know whether findings are representative of HD. Frost, Steketee, and Tolin[1] examined gender in relation to comorbid disorders for 217 people recruited for clinical hoarding1 and reported that significantly more men than women had nonhoarding OCD symptoms. Interestingly, the most frequent comorbid condition, MDD, did not differ by gender, despite its higher prevalence among women in the population; nor did the frequency of social phobia differ by gender. In two controlled studies that recruited OCD patients, Samuels et al.[8, 9] reported that first-degree relatives of OCD participants with hoarding symptoms (ns = 24 and 85, respectively) were more likely to have hoarding symptoms than were relatives of nonhoarding participants. In an early study of participants recruited specifically for hoarding behavior, Frost and Gross[10] reported high rates of hoarding among first-degree relatives in two small samples: (1) 78% of 32 participants in an uncontrolled study, and (2) 85% of 20 people with hoarding compared to 54% of 50 community controls. Pertusa et al.[11] reported higher familial rates of hoarding among HD participants compared to OCD, other anxiety disorders, and community controls, although the rate for HD participants was somewhat lower than earlier studies (52%). Thus, family studies have supported the notion that hoarding is more common among first-degree relatives compared to control samples, but samples were small to moderate in size, and the percentage of affected relatives has ranged widely from 12 in an OCD sample[8] to 85% in a hoarding sample.[10] Findings from genetic studies are consistent with family studies in suggesting familial inheritance of hoarding behavior[12] but so far, nearly all of these studies recruited patients with OCD or Tourette’s syndrome and did not employ standardized measures of hoarding symptoms. Interestingly, a UK study that determined hoarding cases using the Hoarding Rating Scale which closely represents HD criteria[5] found more similarity in hoarding behaviors among monozygotic than dizygotic twins when controlling for shared environment and a higher rate of hoarding among males compared to females. They reported that in female twins, genetic factors accounted for about half of the variance in hoarding symptoms, with environmental effects accounting for the remainder. 1 Frost et al. (2011) studied the same sample of HD participants as for the present paper.
729
Overall, these findings suggest that hoarding aggregates in families and generates discord and risk, especially for family members living in a hoarded environment. However, research on the frequency of hoarding disorder in men and women is limited, and gender patterns as well as the frequency of hoarding symptoms among male and female relatives have received little study. The present study compared a large sample of people recruited for clinical hoarding symptoms to clinical OCD and nonclinical control groups on patterns of hoarding symptoms among their parents and siblings, as well as on possible gender-related aspects of hoarding. Note that this study did not examine the rate of hoarding disorder (HD) among relatives, as measures of impairment or distress were not included.
METHODS PARTICIPANTS The present study included 217 adult participants with clinical levels of hoarding (HD), 96 with obsessive–compulsive disorder without hoarding (OCD), and 130 community controls (CC), for a total sample of 443 who participated in a parent study of the psychopathology of hoarding behavior. Participants were recruited through advertisements for a study about saving, news media, clinics and mental health settings (for OCD and to a lesser extent HD), and via word of mouth. Figure 1 provides a recruitment flow chart. Trained interviewers used the Anxiety Disorders Interview Schedule (ADIS-IV-Lifetime)[13] to determine diagnostic criteria for inclusion. Consistent with current DSM-5 criteria for hoarding disorder[4] inclusion in the HD group required interviewer ratings of moderate or greater clutter, difficulty discarding, and distress/impairment (either or both) from hoarding according to the Hoarding Rating ScaleInterview (HRS-I, see below).[14] In addition, the clutter and difficulty discarding could not be attributed to another OCD symptom (e.g., contamination, checking). To ensure inclusion of a broad spectrum of people with clinical hoarding in the parent study, HD was not required to be the primary diagnosis, and a diagnosis of nonhoarding OCD was permitted. According to ADIS severity ratings, HD was the principal or co-principal diagnosis for 92% of the HD sample, and was secondary to another disorder in 8% of the sample; 18% had comorbid OCD.[1] OCD participants met criteria for a DSM-IV diagnosis of OCD (nonhoarding) as their primary (most severe) problem according to the ADIS-IV-L; hoarding symptoms could be present but were required to be below moderate levels (not qualify for HD group) based on the HRS-I. CC participants were not permitted to meet criteria for any mental health disorder, except specific phobia. Subthreshold hoarding symptoms could be present if they did not meet diagnostic criteria. Criteria for exclusion were suicidal ideation or other risk factors requiring immediate attention, current psychotic symptoms, substance abuse or dependence within the past 3 months, and significant cognitive impairment such as mental retardation or dementia that could compromise informed consent or assessments.
MEASURES The Anxiety Disorders Inventory Schedule for DSM-IV Lifetime version (ADIS-IV-L)[13] guided diagnosis of OCD, anxiety, mood, somatoform, and substance use disorders. Clinical interviews were conducted by master’s level clinical psychologists or postdoctoral fellows trained to criteria using the ADIS-IV-L and supervised by licensed Depression and Anxiety
730
Steketee et al.
Figure 1. Flowchart of participants.
psychologists. This measure was given to participants but not to family members. The Hoarding Rating Scale-Interview (HRS-I)[14] is a semistructured interview containing five questions about the extent of clutter, difficulty discarding, acquisition, distress, and impairment; items relate closely to DSM-5 diagnostic criteria for HD.[4] This measure has demonstrated excellent reliability (test-retest, interrater, home vs. office, internal consistency) and validity (concurrent, discriminant) validity. The HRS was used in conjunction with the ADIS to determine diagnosis of HD for participants (but not for family members). The internal consistency (Cronbach’s α) of the HRS-I for the HD sample was .915. The Saving Inventory-Revised (SI-R)[15] is a widely used 23-item selfreport inventory with established reliability and validity; it contains three subscales that measure clutter, difficulty discarding, and excessive acquisition. Participants (not family members) completed this measure. The internal consistency of the SI-R for the HD sample in this study was .79. The SI-R and the HRS were used to examine the relationship of family variables to participant hoarding severity. The Hoarding Interview (HI) was developed to assess various aspects and consequences of hoarding behavior. Questions addressed who lived in the home (parent, child, sibling, grandparent, partner, roommate, other); current hoarding behaviors; intervention from outside agencies (housing/landlords, fire department, police, child/adult/animal protection); whether clutter presented a problem for health or safety (“yes/no” for falling, risk of fire, personal hygiene, nutrition, medical, home infestation); and problems with daily living activities (rated 0 = no problem to 8 = extreme for job/school, social activities, daily routine, family activities, financial burden, family conflict). The interviewer asked whether participants had grown up in home with “a lot of clutter”, and inquired about family history of hoarding symptoms among relatives (only first-degree relatives— mother, father, sister, brother—are reported in this study) with the following “yes/no” questions: “Did any of your biological relatives: (1) save many things or have a lot of clutter in their living space?, (2) have difficulty discarding things?, (3) buy things excessively?, and (4) pick up free things excessively?” To be consistent with DSM-5 criteria for HD, saving/clutter plus difficulty discarding were combined to measure the hallmark symptoms of HD, and buying or acquiring free items were combined to measure excessive acquiring, a specifier for HD. As the degree of impairment in relatives was not assessed, Depression and Anxiety
a likely diagnosis of HD among participants’ relatives could not be determined. Although all participants completed the interview, items rated “not applicable” were excluded from these analyses, and some data were missing as noted in sample sizes reported for analyses.
PROCEDURE This study was approved by the Institutional Review Boards at Smith College, Boston University, and Hartford Hospital. All participants signed an informed consent form. Graduate students or postdoctoral fellows who were supervised by experienced mental health clinicians completed the ADIS, HRS-I, and Hoarding Interview at baseline, and participants completed self-report measures.
ANALYSES All analyses were conducted in SPSS version 20. Groups were compared using general linear modeling analyses for continuous variables and Pearson chi square analyses for categorical data; significant findings were followed up with post hoc comparisons. Independent samples t-tests were used for comparisons of two groups on continuous variables. To reduce type I error, P values were set to .01 whenever multiple comparisons were conducted. Sample sizes vary in analyses reported below due to missing data for some variables. Because demographic findings for the full sample (see Results below) indicated significant differences on age and gender across groups that might affect the findings for family variables in this study, subsamples matched on age and gender were constituted. For this purpose, participants were matched on gender within age cohorts by decade (i.e., age 18–30, 31–40, 41–50, etc.). A maximum of 150 participants could be matched, including 50 people with HD (50% men; mean age = 44.1, range = 19–74), 50 with OCD (50% men; mean age = 43.5, range = 19–74), and 50 CC (50% men; mean age = 44.2; range = 21–74). As expected, matched samples did not differ on age or gender (Ps > .814); the two clinical groups (HD, OCD) also did not differ on severity with regard to ADIS rating of primary diagnosis (P = .190) or number of comorbid diagnoses (P > .317). Because the full and matched samples were generally similar in rates of relatives’ hoarding symptoms and on statistical comparisons, between-group comparisons are presented below only for the matched sample. This sample likely provided a better estimate of hoarding among siblings (siblings of younger OCD
Research Article: Familial Patterns of Hoarding
participants may not yet show hoarding symptoms which typically develop later in life), and also increased the representation of men across groups. Analyses that did not involve between-group comparisons used the full sample, either all three groups combined or the HD sample only, depending on the goal of the analysis. Accordingly, analyses of participant gender effects used the combined sample, as all groups reported some degree of hoarding among relatives. Only the HD participants were used to examine the impact of participant gender on the consequences of hoarding and the relationship of participant hoarding severity to hoarding symptoms among relatives.
RESULTS DEMOGRAPHIC INFORMATION
Demographic information for all participants and statistical comparisons among groups are given in Table 1. Although the age range was similar across all groups, the OCD sample had a significantly lower mean age (F (2, 438) = 84.7, P < .001) than did the HD and CC groups, which did not differ. Gender distribution also varied significantly across groups (χ 2 (2, n = 443) = 24.9, P < .001), with HD and CC participants including more women, whereas the OCD sample had nearly equal numbers of men and women. The three groups did not differ in education (mainly college), employment (about 50% employed), marital status (35–35% married), or race/ethnicity (about 85% White). Over half of HD participants lived alone, a significantly higher percentage than for the OCD and CC groups, about a third of whom lived alone (χ 2 (2, n = 427) = 14.62, P < .001). The most frequent joint living arrangement was with a partner or roommate, followed by living with children and living with parents. Not surprisingly, the younger OCD sample was significantly more likely to live with their parents (χ 2 (2, n = 427) = 39.64, P < .001) than HD and CC, which did not differ from each other. Comparisons for living with children were not significant (P > .052). The two clinical samples did not differ in clinical severity according to ADIS severity ratings for their primary diagnosis (P = .247); OCD participants had more comorbid current Axis I disorders than did HD participants (1.93 vs. 1.44, t(311) = 4.109, P < .001). As noted above under analyses, following the matching procedure, groups no longer differed significantly on age, gender or clinical severity. They also no longer differed on living situation (Ps > .268). PARTICIPANT GENDER EFFECTS
To explore gender effects for hoarding symptoms that might derive from modelling of behaviors or sex-linked heritability, chi square analyses tested whether women participants would report a higher proportion of female relatives (mothers, sisters) with hoarding and acquiring symptoms, whereas men would report having more male relatives (fathers, brothers) with these symptoms. This analysis used all HD, OCD, and CC participants (ns = 346–408) to examine gender patterns across all groups combined. Women participants reported more
731
hallmark hoarding symptoms (saving/clutter plus discarding) in their mothers (P = .024–.048) and more acquiring among their sisters (P = .024–.048), but as P values exceeded .01, these findings were not considered statistically significant given multiple comparisons. We also examined whether more women than men participants would report having relatives with hoarding symptoms. In the combined sample, significantly more female than male participants indicated they had a first-degree relative with hoarding behaviors (48.4% women vs. 30.7% men, χ 2 = 11.25, P = .001, n = 412) and that they had a parent with hoarding behaviors (40.6% of women vs. 27.6% of men, χ 2 = 6.46, P = .011, n = 410). Potential effects of gender on the severity and consequences of HD were examined within the HD sample with regard to impairment in functioning (employment, personal and family, financial), health and safety (falling, fire, hygiene, medical), and service interventions (housing, fire, child/adult/animal protection, police). Compared to men, scores for women suggested more impairment in social activities (P = .039) and in daily routine (P = .057), but these findings were not considered statistically significant given the multiple comparisons; no other comparisons approached significance, Ps > .129, ns = 145–210. MATCHED SAMPLE COMPARISONS OF HOARDING SYMPTOMS IN RELATIVES
In matched sample comparisons, significantly more HD participants (36.7%) reported growing up in a cluttered home compared to OCD (16%) and CC (10.0%) participants, χ 2 (2, n = 149) = 11.790, P < .003. OCD and CC participants did not differ from each other, P = .372. These rates did not differ significantly for men and women for the three groups combined (P = .052) or for the HD sample alone (P = .483). Figures 2–4 present findings comparing groups on rates of hoarding among first-degree relatives for matched samples. In these figures, circles and squares indicate female and male relatives, respectively, and the size of the object represents the proportion of relatives with hoarding symptoms. P values for sample differences (HD, OCD, CC) are reported on the left or right side of the colored objects, and P values for differences within relative type (mothers vs. fathers; sisters vs. brothers) are given in the center. The overall chi square analyses indicated significant differences among groups on saving/clutter plus difficulty discarding (Fig. 2) for mothers (χ 2 (2, n = 149) = 8.26, P = .016) and for fathers (χ 2 (2, n = 147) = 9.92, P = .007). Similar findings occurred on acquiring (Fig. 3) for mothers (χ 2 (2, n = 147) = 10.26, P = .006) and for fathers (χ 2 (2, n = 144) = 7.00, P = .030). Post hoc pairwise comparisons showed that HD parents had higher rates of hoarding symptoms than did OCD and CC groups which did not differ from each other (HD > [OCD = CC]). The exception was fathers’ acquiring where the HD sample had higher rates than community controls and the two clinical and two control Depression and Anxiety
Steketee et al.
732
TABLE 1. Demographic information for the full sample Descriptive statistics Age Mean (SD) Range Percentage of women Percentage of race/ethnicity White African American Asian American Native American Other Hispanic Years education (SD) Percentage of employed Percentage of married/partnered Percentage of living situation Live alone With partner/roommate With child(ren) With parent(s)
HD (N = 217)
OCD (N = 96)
CC (N = 130)
N
P value
52.6a (10.3) 19–78 77.0a
34.5b (13.7) 18–74 47.9b
52.6a (13.5) 21–83 70.0a
438
.345).
DISCUSSION To our knowledge, this is the first study to examine symptoms of hoarding disorder among parents and siblings in a large clinical sample diagnosed with hoarding disorder independent of OCD symptoms, and including both clinical (OCD) and community control samples. It is noteworthy that while recruitment for the OCD sample occurred mainly through mental health settings, this was less common for the HD sample, many of whom were not seeking treatment. As study recruitment materials referred to saving behavior, the CC sample may have had higher rates of hoarding symptoms than in the population as a whole. Unfortunately, we did not collect data on the referral source for each participant. While the three groups were generally similar on many demographic measures (mainly college educated, about 50% employed, about 40% married or partnered,
Figure 4. First-degree relatives’ with hoarding symptoms of saving/clutter and difficulty discarding in matched samples. Depression and Anxiety
734
Steketee et al.
85% White), the OCD sample was significantly younger and contained more men than the HD and CC samples. This is typical for OCD clinic samples, but their younger age undoubtedly affected findings for hoarding among relatives, especially as about one-quarter (vs. less than 5% of the hoarding and CC samples) lived with their parents rather than living alone as was more common for the HD sample. Our effort to control for this through age and gender matching was effective in yielding groups of about 50% men with an average age in the mid-40s, a wide age range, and clinical groups that were comparable in diagnostic severity and in the general patterns of relatives’ hoarding symptoms. Although the matching process reduced the overall sample size considerably, this study is one of the largest to examine hoarding among relatives and the only one to report on hallmark hoarding symptoms and on acquiring. Of course, in matching the two clinical groups, it is important to note that neither sample can be considered typical of the clinic population, as mean ages for these two clinical groups are often nearly a generation apart and more women are common in HD research samples. At the same time, perusal of the rates of hoarding symptoms among relatives for the full and matched samples indicated that the overall magnitude and general patterns of data were very similar, suggesting that matching provided a reasonable comparison among groups. We examined gender effects in possible sex-linked transmission patterns, the patterns of hoarding symptoms among relatives, and the relationship of participant gender to baseline family variables and hoarding severity. Overall, significantly more mothers and sisters than fathers and brothers were reported to have hoarding symptoms as evident from Figs. 2 and 3. However, sex-linked patterns whereby women participants report more symptoms for female relatives and men for male relatives were not significant. Perhaps both men and women participants were more aware of the patterns of hoarding behavior among their female relatives. It is also possible that men may hide their symptoms more than do women, and that they are less likely to invite people into their home. Both possibilities may be indicative of more modeling of hoarding by mothers than fathers for both girls and boys. Further research is needed to clarify the source of these findings. Acquiring behavior is a specifier rather than a diagnostic requirement for hoarding disorder. With regard to gender differences for acquiring in the combined matched samples (Fig. 3), both mothers and sisters had higher rates than fathers and brothers, but the higher rate for sisters appeared to be due to gender differences in the community sample rather than in the hoarding or OCD samples. Higher rates of impulsive buying behavior, which is not necessarily associated with hoarding have been observed among younger women,[16] and this may account for the findings. Why women should have a higher frequency of hoarding symptoms compared to men is not clear. Efforts to find other gender differences within the HD samDepression and Anxiety
ple yielded no significant findings for family and functioning variables or service interventions, shedding little light on why hoarding appears more often among female than male relatives, even among nonhoarding samples. Unfortunately, our data do not clarify whether hoarding behaviors may be inherited, learned through familial modeling, or both. While it is interesting that over one-third of the HD sample reported growing up in a cluttered home, more than twice the rate for OCD and CC participants, this too does not clarify how hoarding is transmitted. As noted earlier, few studies have identified gender differences associated with hoarding. Samuels et al.’s[6] findings of different OCD subtypes and other psychiatric diagnoses for men versus women with hoarding symptoms do not appear to explain the greater frequency of hoarding among female relatives. In their study of twins in the UK, Iervolino et al.[5] found a higher prevalence of hoarding among men than women (4.1 vs. 2.1%), but they could only confirm heritability among women as the sample size of male twin pairs was too small. Both of our original HD and community samples contained many more women than men, perhaps because women volunteer more often for research on mental health problems, although this does not appear to be true for OCD. Findings from this study echo Iervolino et al.’s[5] call for better ways to sample men with hoarding behaviors. The HD sample generally showed higher rates of symptoms in their first-degree relatives compared to the OCD and community controls when matched groups were compared on hoarding symptoms reported for parents and for siblings. Parents were reported to have higher rates than siblings, perhaps due to their older age as hoarding symptom severity has been shown to increase with age.[17] Interestingly, excessive acquiring (Fig. 3), which is a specifier for HD does not appear to occur at a lower rate than the hallmark symptoms of saving/clutter and discarding (Fig. 2). It is important to note that the relatively high rates of symptoms across all three samples, especially for the OCD and CC samples, reflects the assessment of hoarding symptoms rather than an HD diagnosis that requires additional assessment of impairment and distress. As apparent from Fig. 4, within the HD sample, the overall rate of the major hoarding symptoms reported for any first-degree relative was quite high at 57%, with parents (53%) accounting for the largest portion. Difficulty discarding and saving/clutter also occurred with surprising frequency among relatives of nonhoarding samples, with over one-quarter of firstdegree relatives and one-fifth of parents showing these symptoms among the OCD and community groups (Fig. 4). That is, while hoarding behavior does “run in families” of those with HD, a substantial proportion of those without clinical hoarding may also have relatives with hoarding problems. The overall rate of hoarding saving/clutter and difficulty discarding among first-degree relatives in our study (57%) is slightly higher than reports using moderately large sample sizes by Pertusa et al.[11] (52%, n = 71)
Research Article: Familial Patterns of Hoarding
who employed similar diagnostic criteria for HD and by Samuels et al.[9] (49%, n = 142) using an OCD recruited sample. An early case study by Frost and Gross[10] reported unusually high rates of hoarding among firstdegree relatives of 78–85% (n = 52). The differences in rates are likely accounted for by methodological differences in recruitment and assessment strategies for the hoarding and relative samples. The 30% rate of hoarding symptoms among relatives in our OCD sample is higher than that reported by Pertusa et al.[11] (10%) but very similar to the 33% reported by Samuels et al.,[9] again, likely due to methodological differences. A substantial portion (26%) of our community sample also reported hoarding among their relatives. This figure lies midway between Frost and Gross’s[10] much higher rate of 54% in their early study and the 3–5% rates reported by Samuels et al.[8] and Pertusa et al.[11] which is similar to prevalence estimates for HD in the population. As our study assessed only individual hoarding behaviors/symptoms rather than HD diagnosis, it is not surprising that the rates among relatives exceed HD prevalence estimates. Hirschtritt and Mathews’[12] review of the genetic evidence for hoarding disorder indicates that most family studies support the notion that hoarding is more common among first-degree relatives of people with HD compared to control samples, and genetic studies suggest a strong genetic component in a complex pattern of inheritance, perhaps accounting for half the phenotypic variance. They point to limitations in the literature that relies on OCD samples, rather than those diagnosed with HD, and lacks replication studies. The present study found that hoarding symptoms are fairly common among first-degree relatives of people with HD, although the proportion with a diagnosis of HD among family members cannot be estimated from our data, which lacked an assessment of impairment and direct assessment of relatives’ hoarding symptoms. On the other hand, the severity of HD participants’ hoarding was not associated with the presence of relatives’ hoarding behavior, and thus, it does not appear that those with more severe clinical hoarding are more likely to have first-degree relatives with this problem as might be expected with a strong genetic loading. Again, however, limitations in assessment limit conclusions. There are several additional limitations of the current study. Our samples were recruited through a variety of methods but may not be representative of the populations under study. As noted earlier, the reliance on participant reports of family hoarding behavior without corroboration by other direct assessment of relatives’ behaviors is an important limitation. The rates of reported hoarding symptoms may be elevated for this reason, and as no formal diagnoses could be made, the actual prevalence of hoarding disorder among relatives is likely much lower than the reported frequency of individual symptoms. Because the preponderance of women in most clinic HD samples can bias findings with regard
735
to gender, future studies should utilize epidemiological recruiting strategies and, if necessary, oversample men in order to guard against such bias. Finally, the Hoarding Interview used here was developed to assess a variety of variables associated with hoarding, but its psychometric properties have not been established. Overall, the current study supports the notion that hoarding aggregates in families, appearing among a large percentage of first-degree relatives of people who hoard and also among a surprisingly large number of relatives of those who do not have clinical hoarding. The apparent predominance of hoarding symptoms among female relatives is noteworthy, evident not only for hoarding participants but also for the nonhoarding samples. Determination of hoarding transmission through genetic or modelling methods will require careful twin studies on large samples, which may be more easily accomplished in countries where twin registries are in place. These findings do suggest that, given the relatively high frequency of hoarding symptoms within families, clinicians working with family members of people who hoard must keep in mind the likelihood that relatives, especially mothers and sisters, will also have significant hoarding symptoms. As the role of genetic transmission and modeling are better understood, strategies for prevention can be developed to take this information into account.
REFERENCES 1. Frost R, Steketee G, Tolin D. Comorbidity in hoarding disorder. Depress Anxiety 2011;28:1–9. 2. Tolin DF, Frost RO, Steketee G, Fitch KE. Family burden of compulsive hoarding: results of an internet survey. Behav Res Ther 2008;46(3):334–344. 3. Tolin DF, Frost RO, Steketee G, Gray KD, Fitch KE. The economic and social burden of compulsive hoarding. Psychiatry Res 2008;160(2):200–211. 4. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Washington, DC: American Psychiatric Association; 2013. 5. Iervolino AC, Perroud N, Fullana MA, Guipponi M, Cherkas L, Collier DA, Mataix-Cols D. Prevalence and heritability of compulsive hoarding: a twin study. Am J Psychiatry 2009;166(10):1156–1161. 6. Samuels JF, Bienvenu OJ, Grados MA, et al. Prevalence and correlates of hoarding behavior in a community-based sample. Behav Res Ther 2008;46(7):836–844. 7. Steketee G, Frost RO. Phenomenology of hoarding. In: Frost RO, Steketee G, editors. The Oxford Handbook of Hoarding and Acquiring. New York: Oxford University Press; 2014:19–32. 8. Samuels J, Bienvenu OJ, 3rd, Riddle MA, et al. Hoarding in obsessive compulsive disorder: results from a case-control study. Behav Res Ther 2002;40(5):517–528. 9. Samuels JF, Bienvenu OJ, 3rd, Pinto A, et al. Hoarding in obsessive-compulsive disorder: results from the OCD Collaborative Genetics Study. Behav Res Ther 2007;45(4):673– 686. 10. Frost RO, Gross RC. The hoarding of possessions. Behav Res Ther 1993;31(4):367–381.
Depression and Anxiety
736
Steketee et al.
11. Pertusa A, Fullana M, Singh S, Alonso P, Menchon ´ J, Mataix-Cols D. Compulsive hoarding: a symptom of OCD, a distinct clinical syndrome, or both? Am J Psychiatry 2008;165:1289–1298. 12. Hirschtritt ME, Mathews CA. Genetics and family models of hoarding disorder. In: Frost RO, Steketee G, editors. Oxford Handbook of Hoarding and Acquiring. New York, NY: Oxford University Press; 2014:159–176. 13. Brown TA, Di Nardo PA, Barlow DH. Anxiety Disorders Interview Schedule for DSM-IV: Lifetime Version (ADIS-IV-L). San Antonio, TX: Psychological Corporation; 1994.
Depression and Anxiety
14. Tolin D, Frost, RO, Steketee, G. A brief interview for assessing compulsive hoarding: The Hoarding Rating Scale-Interview. Psychiatry Res 2010;178:147–152. 15. Frost RO, Steketee G, Grisham J. Measurement of compulsive hoarding: saving inventory-revised. Behav Res Ther 2004;42(10):1163–1182. 16. Black DW. Compulsive buying disorder: a review of the evidence. CNS Spectr 2007;12(2):124–132. 17. Grisham J, Frost RO, Steketee G, Kim H-J, Hood, S. Age of onset of compulsive hoarding. J Anxiety Disorders 2006;20:675–686.
Research Article FAMILIAL PATTERNS OF HOARDING SYMPTOMS Gail Steketee, Ph.D.,1 ∗ Andrea A. Kelley, M.S.W., M.P.H.,1 Jeremy A. Wernick, B.A.,1 Jordana Muroff, Ph.D.,1 Randy O. Frost, Ph.D.,2 and David F. Tolin, Ph.D.3
Background: Previous research suggests that hoarding aggregates in families and is associated with health and safety risks and family problems. The present study examined gender- and diagnosis-related differences in reports of hoarding symptoms among first-degree relatives of people who hoard, and of clinical and community samples. Methods: The present study included 443 participants in a study of hoarding behavior: 217 with hoarding disorder (HD), 96 with obsessivecompulsive disorder (OCD), and 130 nonclinical community controls (CC). Assessment included a detailed interview of familial patterns of hoarding behaviors among parents and siblings and measures of hoarding severity. Results: In the combined sample, participants reported more hoarding among female (mothers, sisters) than male (fathers, brothers) relatives. Significantly more female than male participants indicated they had a parent or any first-degree relative with hoarding behaviors. However, within the HD sample no significant gender effects were found for household, safety, and functioning variables, or for hoarding symptom severity. In an age- and gender-matched subsample (total n = 150), HD participants reported more hallmark hoarding symptoms (difficulty discarding and saving/clutter), and acquiring among their relatives compared to OCD and CC samples, and parents had higher rates than siblings. Conclusions: Hoarding symptoms appear to be common among first-degree relatives of people who hoard and are also found among relatives of control samples. The predominance of hoarding symptoms among female relatives may indicate genetic or modeling transmission but this requires further study using large twin samples. Clinicians should consider that family members may also have signifi C 2015 cant hoarding symptoms. Depression and Anxiety 32:728–736, 2015. Wiley Periodicals, Inc.
Key words: hoarding, family/marital, OCD; gender, gene/environment
FAMILIAL PATTERNS IN HOARDING DISORDER
1 Boston
University, School of Social Work, Boston, MA Department of Psychology, Northampton, MA 3 Anxiety Disorders Center, Institute of Living, Hartford, CT and Department of Psychiatry, Yale University School of Medicine, New Haven, CT 2 Smith College,
Contract grant sponsor: NIMH; Contract grant numbers: R01 MH068008 and R01 MH068007. ∗ Correspondence
to: Gail Steketee, School of Social Work, Boston University, 264 Bay State Rd., Boston, MA 02215. E-mail: [email protected] Received for publication 20 October 2014; Revised 6 June 2015; Accepted 9 June 2015 DOI 10.1002/da.22393
C 2015 Wiley Periodicals, Inc.
Hoarding disorder (HD) is a debilitating psychiatric
problem that can lead to considerable health risks, functional impairment, family conflict, and substantial financial burden for sufferers, family members, and the community.[1–3] Recently included in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5),[4] HD is characterized by excessive saving and difficulty discarding possessions, resulting in substantial clutter in active living areas; additional features include excessive acquisition and varying levels of insight. Prevalence estimates for HD range from 2 to 5%
Published online 30 June 2015 in Wiley Online Library (wileyonlinelibrary.com).
Research Article: Familial Patterns of Hoarding
among adults in the United States and other western countries.[5, 6] The few studies that have examined gender differences in hoarding prevalence and associated features have reported mixed results. Although women predominate in clinical HD samples, it is unclear whether hoarding affects women more than men.[6, 7] Samuels et al.[6] found that men and women with hoarding differed in obsessivecompulsive disorder (OCD) subtypes and comorbid disorders, but because they recruited from an OCD sample, it is difficult to know whether findings are representative of HD. Frost, Steketee, and Tolin[1] examined gender in relation to comorbid disorders for 217 people recruited for clinical hoarding1 and reported that significantly more men than women had nonhoarding OCD symptoms. Interestingly, the most frequent comorbid condition, MDD, did not differ by gender, despite its higher prevalence among women in the population; nor did the frequency of social phobia differ by gender. In two controlled studies that recruited OCD patients, Samuels et al.[8, 9] reported that first-degree relatives of OCD participants with hoarding symptoms (ns = 24 and 85, respectively) were more likely to have hoarding symptoms than were relatives of nonhoarding participants. In an early study of participants recruited specifically for hoarding behavior, Frost and Gross[10] reported high rates of hoarding among first-degree relatives in two small samples: (1) 78% of 32 participants in an uncontrolled study, and (2) 85% of 20 people with hoarding compared to 54% of 50 community controls. Pertusa et al.[11] reported higher familial rates of hoarding among HD participants compared to OCD, other anxiety disorders, and community controls, although the rate for HD participants was somewhat lower than earlier studies (52%). Thus, family studies have supported the notion that hoarding is more common among first-degree relatives compared to control samples, but samples were small to moderate in size, and the percentage of affected relatives has ranged widely from 12 in an OCD sample[8] to 85% in a hoarding sample.[10] Findings from genetic studies are consistent with family studies in suggesting familial inheritance of hoarding behavior[12] but so far, nearly all of these studies recruited patients with OCD or Tourette’s syndrome and did not employ standardized measures of hoarding symptoms. Interestingly, a UK study that determined hoarding cases using the Hoarding Rating Scale which closely represents HD criteria[5] found more similarity in hoarding behaviors among monozygotic than dizygotic twins when controlling for shared environment and a higher rate of hoarding among males compared to females. They reported that in female twins, genetic factors accounted for about half of the variance in hoarding symptoms, with environmental effects accounting for the remainder. 1 Frost et al. (2011) studied the same sample of HD participants as for the present paper.
729
Overall, these findings suggest that hoarding aggregates in families and generates discord and risk, especially for family members living in a hoarded environment. However, research on the frequency of hoarding disorder in men and women is limited, and gender patterns as well as the frequency of hoarding symptoms among male and female relatives have received little study. The present study compared a large sample of people recruited for clinical hoarding symptoms to clinical OCD and nonclinical control groups on patterns of hoarding symptoms among their parents and siblings, as well as on possible gender-related aspects of hoarding. Note that this study did not examine the rate of hoarding disorder (HD) among relatives, as measures of impairment or distress were not included.
METHODS PARTICIPANTS The present study included 217 adult participants with clinical levels of hoarding (HD), 96 with obsessive–compulsive disorder without hoarding (OCD), and 130 community controls (CC), for a total sample of 443 who participated in a parent study of the psychopathology of hoarding behavior. Participants were recruited through advertisements for a study about saving, news media, clinics and mental health settings (for OCD and to a lesser extent HD), and via word of mouth. Figure 1 provides a recruitment flow chart. Trained interviewers used the Anxiety Disorders Interview Schedule (ADIS-IV-Lifetime)[13] to determine diagnostic criteria for inclusion. Consistent with current DSM-5 criteria for hoarding disorder[4] inclusion in the HD group required interviewer ratings of moderate or greater clutter, difficulty discarding, and distress/impairment (either or both) from hoarding according to the Hoarding Rating ScaleInterview (HRS-I, see below).[14] In addition, the clutter and difficulty discarding could not be attributed to another OCD symptom (e.g., contamination, checking). To ensure inclusion of a broad spectrum of people with clinical hoarding in the parent study, HD was not required to be the primary diagnosis, and a diagnosis of nonhoarding OCD was permitted. According to ADIS severity ratings, HD was the principal or co-principal diagnosis for 92% of the HD sample, and was secondary to another disorder in 8% of the sample; 18% had comorbid OCD.[1] OCD participants met criteria for a DSM-IV diagnosis of OCD (nonhoarding) as their primary (most severe) problem according to the ADIS-IV-L; hoarding symptoms could be present but were required to be below moderate levels (not qualify for HD group) based on the HRS-I. CC participants were not permitted to meet criteria for any mental health disorder, except specific phobia. Subthreshold hoarding symptoms could be present if they did not meet diagnostic criteria. Criteria for exclusion were suicidal ideation or other risk factors requiring immediate attention, current psychotic symptoms, substance abuse or dependence within the past 3 months, and significant cognitive impairment such as mental retardation or dementia that could compromise informed consent or assessments.
MEASURES The Anxiety Disorders Inventory Schedule for DSM-IV Lifetime version (ADIS-IV-L)[13] guided diagnosis of OCD, anxiety, mood, somatoform, and substance use disorders. Clinical interviews were conducted by master’s level clinical psychologists or postdoctoral fellows trained to criteria using the ADIS-IV-L and supervised by licensed Depression and Anxiety
730
Steketee et al.
Figure 1. Flowchart of participants.
psychologists. This measure was given to participants but not to family members. The Hoarding Rating Scale-Interview (HRS-I)[14] is a semistructured interview containing five questions about the extent of clutter, difficulty discarding, acquisition, distress, and impairment; items relate closely to DSM-5 diagnostic criteria for HD.[4] This measure has demonstrated excellent reliability (test-retest, interrater, home vs. office, internal consistency) and validity (concurrent, discriminant) validity. The HRS was used in conjunction with the ADIS to determine diagnosis of HD for participants (but not for family members). The internal consistency (Cronbach’s α) of the HRS-I for the HD sample was .915. The Saving Inventory-Revised (SI-R)[15] is a widely used 23-item selfreport inventory with established reliability and validity; it contains three subscales that measure clutter, difficulty discarding, and excessive acquisition. Participants (not family members) completed this measure. The internal consistency of the SI-R for the HD sample in this study was .79. The SI-R and the HRS were used to examine the relationship of family variables to participant hoarding severity. The Hoarding Interview (HI) was developed to assess various aspects and consequences of hoarding behavior. Questions addressed who lived in the home (parent, child, sibling, grandparent, partner, roommate, other); current hoarding behaviors; intervention from outside agencies (housing/landlords, fire department, police, child/adult/animal protection); whether clutter presented a problem for health or safety (“yes/no” for falling, risk of fire, personal hygiene, nutrition, medical, home infestation); and problems with daily living activities (rated 0 = no problem to 8 = extreme for job/school, social activities, daily routine, family activities, financial burden, family conflict). The interviewer asked whether participants had grown up in home with “a lot of clutter”, and inquired about family history of hoarding symptoms among relatives (only first-degree relatives— mother, father, sister, brother—are reported in this study) with the following “yes/no” questions: “Did any of your biological relatives: (1) save many things or have a lot of clutter in their living space?, (2) have difficulty discarding things?, (3) buy things excessively?, and (4) pick up free things excessively?” To be consistent with DSM-5 criteria for HD, saving/clutter plus difficulty discarding were combined to measure the hallmark symptoms of HD, and buying or acquiring free items were combined to measure excessive acquiring, a specifier for HD. As the degree of impairment in relatives was not assessed, Depression and Anxiety
a likely diagnosis of HD among participants’ relatives could not be determined. Although all participants completed the interview, items rated “not applicable” were excluded from these analyses, and some data were missing as noted in sample sizes reported for analyses.
PROCEDURE This study was approved by the Institutional Review Boards at Smith College, Boston University, and Hartford Hospital. All participants signed an informed consent form. Graduate students or postdoctoral fellows who were supervised by experienced mental health clinicians completed the ADIS, HRS-I, and Hoarding Interview at baseline, and participants completed self-report measures.
ANALYSES All analyses were conducted in SPSS version 20. Groups were compared using general linear modeling analyses for continuous variables and Pearson chi square analyses for categorical data; significant findings were followed up with post hoc comparisons. Independent samples t-tests were used for comparisons of two groups on continuous variables. To reduce type I error, P values were set to .01 whenever multiple comparisons were conducted. Sample sizes vary in analyses reported below due to missing data for some variables. Because demographic findings for the full sample (see Results below) indicated significant differences on age and gender across groups that might affect the findings for family variables in this study, subsamples matched on age and gender were constituted. For this purpose, participants were matched on gender within age cohorts by decade (i.e., age 18–30, 31–40, 41–50, etc.). A maximum of 150 participants could be matched, including 50 people with HD (50% men; mean age = 44.1, range = 19–74), 50 with OCD (50% men; mean age = 43.5, range = 19–74), and 50 CC (50% men; mean age = 44.2; range = 21–74). As expected, matched samples did not differ on age or gender (Ps > .814); the two clinical groups (HD, OCD) also did not differ on severity with regard to ADIS rating of primary diagnosis (P = .190) or number of comorbid diagnoses (P > .317). Because the full and matched samples were generally similar in rates of relatives’ hoarding symptoms and on statistical comparisons, between-group comparisons are presented below only for the matched sample. This sample likely provided a better estimate of hoarding among siblings (siblings of younger OCD
Research Article: Familial Patterns of Hoarding
participants may not yet show hoarding symptoms which typically develop later in life), and also increased the representation of men across groups. Analyses that did not involve between-group comparisons used the full sample, either all three groups combined or the HD sample only, depending on the goal of the analysis. Accordingly, analyses of participant gender effects used the combined sample, as all groups reported some degree of hoarding among relatives. Only the HD participants were used to examine the impact of participant gender on the consequences of hoarding and the relationship of participant hoarding severity to hoarding symptoms among relatives.
RESULTS DEMOGRAPHIC INFORMATION
Demographic information for all participants and statistical comparisons among groups are given in Table 1. Although the age range was similar across all groups, the OCD sample had a significantly lower mean age (F (2, 438) = 84.7, P < .001) than did the HD and CC groups, which did not differ. Gender distribution also varied significantly across groups (χ 2 (2, n = 443) = 24.9, P < .001), with HD and CC participants including more women, whereas the OCD sample had nearly equal numbers of men and women. The three groups did not differ in education (mainly college), employment (about 50% employed), marital status (35–35% married), or race/ethnicity (about 85% White). Over half of HD participants lived alone, a significantly higher percentage than for the OCD and CC groups, about a third of whom lived alone (χ 2 (2, n = 427) = 14.62, P < .001). The most frequent joint living arrangement was with a partner or roommate, followed by living with children and living with parents. Not surprisingly, the younger OCD sample was significantly more likely to live with their parents (χ 2 (2, n = 427) = 39.64, P < .001) than HD and CC, which did not differ from each other. Comparisons for living with children were not significant (P > .052). The two clinical samples did not differ in clinical severity according to ADIS severity ratings for their primary diagnosis (P = .247); OCD participants had more comorbid current Axis I disorders than did HD participants (1.93 vs. 1.44, t(311) = 4.109, P < .001). As noted above under analyses, following the matching procedure, groups no longer differed significantly on age, gender or clinical severity. They also no longer differed on living situation (Ps > .268). PARTICIPANT GENDER EFFECTS
To explore gender effects for hoarding symptoms that might derive from modelling of behaviors or sex-linked heritability, chi square analyses tested whether women participants would report a higher proportion of female relatives (mothers, sisters) with hoarding and acquiring symptoms, whereas men would report having more male relatives (fathers, brothers) with these symptoms. This analysis used all HD, OCD, and CC participants (ns = 346–408) to examine gender patterns across all groups combined. Women participants reported more
731
hallmark hoarding symptoms (saving/clutter plus discarding) in their mothers (P = .024–.048) and more acquiring among their sisters (P = .024–.048), but as P values exceeded .01, these findings were not considered statistically significant given multiple comparisons. We also examined whether more women than men participants would report having relatives with hoarding symptoms. In the combined sample, significantly more female than male participants indicated they had a first-degree relative with hoarding behaviors (48.4% women vs. 30.7% men, χ 2 = 11.25, P = .001, n = 412) and that they had a parent with hoarding behaviors (40.6% of women vs. 27.6% of men, χ 2 = 6.46, P = .011, n = 410). Potential effects of gender on the severity and consequences of HD were examined within the HD sample with regard to impairment in functioning (employment, personal and family, financial), health and safety (falling, fire, hygiene, medical), and service interventions (housing, fire, child/adult/animal protection, police). Compared to men, scores for women suggested more impairment in social activities (P = .039) and in daily routine (P = .057), but these findings were not considered statistically significant given the multiple comparisons; no other comparisons approached significance, Ps > .129, ns = 145–210. MATCHED SAMPLE COMPARISONS OF HOARDING SYMPTOMS IN RELATIVES
In matched sample comparisons, significantly more HD participants (36.7%) reported growing up in a cluttered home compared to OCD (16%) and CC (10.0%) participants, χ 2 (2, n = 149) = 11.790, P < .003. OCD and CC participants did not differ from each other, P = .372. These rates did not differ significantly for men and women for the three groups combined (P = .052) or for the HD sample alone (P = .483). Figures 2–4 present findings comparing groups on rates of hoarding among first-degree relatives for matched samples. In these figures, circles and squares indicate female and male relatives, respectively, and the size of the object represents the proportion of relatives with hoarding symptoms. P values for sample differences (HD, OCD, CC) are reported on the left or right side of the colored objects, and P values for differences within relative type (mothers vs. fathers; sisters vs. brothers) are given in the center. The overall chi square analyses indicated significant differences among groups on saving/clutter plus difficulty discarding (Fig. 2) for mothers (χ 2 (2, n = 149) = 8.26, P = .016) and for fathers (χ 2 (2, n = 147) = 9.92, P = .007). Similar findings occurred on acquiring (Fig. 3) for mothers (χ 2 (2, n = 147) = 10.26, P = .006) and for fathers (χ 2 (2, n = 144) = 7.00, P = .030). Post hoc pairwise comparisons showed that HD parents had higher rates of hoarding symptoms than did OCD and CC groups which did not differ from each other (HD > [OCD = CC]). The exception was fathers’ acquiring where the HD sample had higher rates than community controls and the two clinical and two control Depression and Anxiety
Steketee et al.
732
TABLE 1. Demographic information for the full sample Descriptive statistics Age Mean (SD) Range Percentage of women Percentage of race/ethnicity White African American Asian American Native American Other Hispanic Years education (SD) Percentage of employed Percentage of married/partnered Percentage of living situation Live alone With partner/roommate With child(ren) With parent(s)
HD (N = 217)
OCD (N = 96)
CC (N = 130)
N
P value
52.6a (10.3) 19–78 77.0a
34.5b (13.7) 18–74 47.9b
52.6a (13.5) 21–83 70.0a
438
.345).
DISCUSSION To our knowledge, this is the first study to examine symptoms of hoarding disorder among parents and siblings in a large clinical sample diagnosed with hoarding disorder independent of OCD symptoms, and including both clinical (OCD) and community control samples. It is noteworthy that while recruitment for the OCD sample occurred mainly through mental health settings, this was less common for the HD sample, many of whom were not seeking treatment. As study recruitment materials referred to saving behavior, the CC sample may have had higher rates of hoarding symptoms than in the population as a whole. Unfortunately, we did not collect data on the referral source for each participant. While the three groups were generally similar on many demographic measures (mainly college educated, about 50% employed, about 40% married or partnered,
Figure 4. First-degree relatives’ with hoarding symptoms of saving/clutter and difficulty discarding in matched samples. Depression and Anxiety
734
Steketee et al.
85% White), the OCD sample was significantly younger and contained more men than the HD and CC samples. This is typical for OCD clinic samples, but their younger age undoubtedly affected findings for hoarding among relatives, especially as about one-quarter (vs. less than 5% of the hoarding and CC samples) lived with their parents rather than living alone as was more common for the HD sample. Our effort to control for this through age and gender matching was effective in yielding groups of about 50% men with an average age in the mid-40s, a wide age range, and clinical groups that were comparable in diagnostic severity and in the general patterns of relatives’ hoarding symptoms. Although the matching process reduced the overall sample size considerably, this study is one of the largest to examine hoarding among relatives and the only one to report on hallmark hoarding symptoms and on acquiring. Of course, in matching the two clinical groups, it is important to note that neither sample can be considered typical of the clinic population, as mean ages for these two clinical groups are often nearly a generation apart and more women are common in HD research samples. At the same time, perusal of the rates of hoarding symptoms among relatives for the full and matched samples indicated that the overall magnitude and general patterns of data were very similar, suggesting that matching provided a reasonable comparison among groups. We examined gender effects in possible sex-linked transmission patterns, the patterns of hoarding symptoms among relatives, and the relationship of participant gender to baseline family variables and hoarding severity. Overall, significantly more mothers and sisters than fathers and brothers were reported to have hoarding symptoms as evident from Figs. 2 and 3. However, sex-linked patterns whereby women participants report more symptoms for female relatives and men for male relatives were not significant. Perhaps both men and women participants were more aware of the patterns of hoarding behavior among their female relatives. It is also possible that men may hide their symptoms more than do women, and that they are less likely to invite people into their home. Both possibilities may be indicative of more modeling of hoarding by mothers than fathers for both girls and boys. Further research is needed to clarify the source of these findings. Acquiring behavior is a specifier rather than a diagnostic requirement for hoarding disorder. With regard to gender differences for acquiring in the combined matched samples (Fig. 3), both mothers and sisters had higher rates than fathers and brothers, but the higher rate for sisters appeared to be due to gender differences in the community sample rather than in the hoarding or OCD samples. Higher rates of impulsive buying behavior, which is not necessarily associated with hoarding have been observed among younger women,[16] and this may account for the findings. Why women should have a higher frequency of hoarding symptoms compared to men is not clear. Efforts to find other gender differences within the HD samDepression and Anxiety
ple yielded no significant findings for family and functioning variables or service interventions, shedding little light on why hoarding appears more often among female than male relatives, even among nonhoarding samples. Unfortunately, our data do not clarify whether hoarding behaviors may be inherited, learned through familial modeling, or both. While it is interesting that over one-third of the HD sample reported growing up in a cluttered home, more than twice the rate for OCD and CC participants, this too does not clarify how hoarding is transmitted. As noted earlier, few studies have identified gender differences associated with hoarding. Samuels et al.’s[6] findings of different OCD subtypes and other psychiatric diagnoses for men versus women with hoarding symptoms do not appear to explain the greater frequency of hoarding among female relatives. In their study of twins in the UK, Iervolino et al.[5] found a higher prevalence of hoarding among men than women (4.1 vs. 2.1%), but they could only confirm heritability among women as the sample size of male twin pairs was too small. Both of our original HD and community samples contained many more women than men, perhaps because women volunteer more often for research on mental health problems, although this does not appear to be true for OCD. Findings from this study echo Iervolino et al.’s[5] call for better ways to sample men with hoarding behaviors. The HD sample generally showed higher rates of symptoms in their first-degree relatives compared to the OCD and community controls when matched groups were compared on hoarding symptoms reported for parents and for siblings. Parents were reported to have higher rates than siblings, perhaps due to their older age as hoarding symptom severity has been shown to increase with age.[17] Interestingly, excessive acquiring (Fig. 3), which is a specifier for HD does not appear to occur at a lower rate than the hallmark symptoms of saving/clutter and discarding (Fig. 2). It is important to note that the relatively high rates of symptoms across all three samples, especially for the OCD and CC samples, reflects the assessment of hoarding symptoms rather than an HD diagnosis that requires additional assessment of impairment and distress. As apparent from Fig. 4, within the HD sample, the overall rate of the major hoarding symptoms reported for any first-degree relative was quite high at 57%, with parents (53%) accounting for the largest portion. Difficulty discarding and saving/clutter also occurred with surprising frequency among relatives of nonhoarding samples, with over one-quarter of firstdegree relatives and one-fifth of parents showing these symptoms among the OCD and community groups (Fig. 4). That is, while hoarding behavior does “run in families” of those with HD, a substantial proportion of those without clinical hoarding may also have relatives with hoarding problems. The overall rate of hoarding saving/clutter and difficulty discarding among first-degree relatives in our study (57%) is slightly higher than reports using moderately large sample sizes by Pertusa et al.[11] (52%, n = 71)
Research Article: Familial Patterns of Hoarding
who employed similar diagnostic criteria for HD and by Samuels et al.[9] (49%, n = 142) using an OCD recruited sample. An early case study by Frost and Gross[10] reported unusually high rates of hoarding among firstdegree relatives of 78–85% (n = 52). The differences in rates are likely accounted for by methodological differences in recruitment and assessment strategies for the hoarding and relative samples. The 30% rate of hoarding symptoms among relatives in our OCD sample is higher than that reported by Pertusa et al.[11] (10%) but very similar to the 33% reported by Samuels et al.,[9] again, likely due to methodological differences. A substantial portion (26%) of our community sample also reported hoarding among their relatives. This figure lies midway between Frost and Gross’s[10] much higher rate of 54% in their early study and the 3–5% rates reported by Samuels et al.[8] and Pertusa et al.[11] which is similar to prevalence estimates for HD in the population. As our study assessed only individual hoarding behaviors/symptoms rather than HD diagnosis, it is not surprising that the rates among relatives exceed HD prevalence estimates. Hirschtritt and Mathews’[12] review of the genetic evidence for hoarding disorder indicates that most family studies support the notion that hoarding is more common among first-degree relatives of people with HD compared to control samples, and genetic studies suggest a strong genetic component in a complex pattern of inheritance, perhaps accounting for half the phenotypic variance. They point to limitations in the literature that relies on OCD samples, rather than those diagnosed with HD, and lacks replication studies. The present study found that hoarding symptoms are fairly common among first-degree relatives of people with HD, although the proportion with a diagnosis of HD among family members cannot be estimated from our data, which lacked an assessment of impairment and direct assessment of relatives’ hoarding symptoms. On the other hand, the severity of HD participants’ hoarding was not associated with the presence of relatives’ hoarding behavior, and thus, it does not appear that those with more severe clinical hoarding are more likely to have first-degree relatives with this problem as might be expected with a strong genetic loading. Again, however, limitations in assessment limit conclusions. There are several additional limitations of the current study. Our samples were recruited through a variety of methods but may not be representative of the populations under study. As noted earlier, the reliance on participant reports of family hoarding behavior without corroboration by other direct assessment of relatives’ behaviors is an important limitation. The rates of reported hoarding symptoms may be elevated for this reason, and as no formal diagnoses could be made, the actual prevalence of hoarding disorder among relatives is likely much lower than the reported frequency of individual symptoms. Because the preponderance of women in most clinic HD samples can bias findings with regard
735
to gender, future studies should utilize epidemiological recruiting strategies and, if necessary, oversample men in order to guard against such bias. Finally, the Hoarding Interview used here was developed to assess a variety of variables associated with hoarding, but its psychometric properties have not been established. Overall, the current study supports the notion that hoarding aggregates in families, appearing among a large percentage of first-degree relatives of people who hoard and also among a surprisingly large number of relatives of those who do not have clinical hoarding. The apparent predominance of hoarding symptoms among female relatives is noteworthy, evident not only for hoarding participants but also for the nonhoarding samples. Determination of hoarding transmission through genetic or modelling methods will require careful twin studies on large samples, which may be more easily accomplished in countries where twin registries are in place. These findings do suggest that, given the relatively high frequency of hoarding symptoms within families, clinicians working with family members of people who hoard must keep in mind the likelihood that relatives, especially mothers and sisters, will also have significant hoarding symptoms. As the role of genetic transmission and modeling are better understood, strategies for prevention can be developed to take this information into account.
REFERENCES 1. Frost R, Steketee G, Tolin D. Comorbidity in hoarding disorder. Depress Anxiety 2011;28:1–9. 2. Tolin DF, Frost RO, Steketee G, Fitch KE. Family burden of compulsive hoarding: results of an internet survey. Behav Res Ther 2008;46(3):334–344. 3. Tolin DF, Frost RO, Steketee G, Gray KD, Fitch KE. The economic and social burden of compulsive hoarding. Psychiatry Res 2008;160(2):200–211. 4. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Washington, DC: American Psychiatric Association; 2013. 5. Iervolino AC, Perroud N, Fullana MA, Guipponi M, Cherkas L, Collier DA, Mataix-Cols D. Prevalence and heritability of compulsive hoarding: a twin study. Am J Psychiatry 2009;166(10):1156–1161. 6. Samuels JF, Bienvenu OJ, Grados MA, et al. Prevalence and correlates of hoarding behavior in a community-based sample. Behav Res Ther 2008;46(7):836–844. 7. Steketee G, Frost RO. Phenomenology of hoarding. In: Frost RO, Steketee G, editors. The Oxford Handbook of Hoarding and Acquiring. New York: Oxford University Press; 2014:19–32. 8. Samuels J, Bienvenu OJ, 3rd, Riddle MA, et al. Hoarding in obsessive compulsive disorder: results from a case-control study. Behav Res Ther 2002;40(5):517–528. 9. Samuels JF, Bienvenu OJ, 3rd, Pinto A, et al. Hoarding in obsessive-compulsive disorder: results from the OCD Collaborative Genetics Study. Behav Res Ther 2007;45(4):673– 686. 10. Frost RO, Gross RC. The hoarding of possessions. Behav Res Ther 1993;31(4):367–381.
Depression and Anxiety
736
Steketee et al.
11. Pertusa A, Fullana M, Singh S, Alonso P, Menchon ´ J, Mataix-Cols D. Compulsive hoarding: a symptom of OCD, a distinct clinical syndrome, or both? Am J Psychiatry 2008;165:1289–1298. 12. Hirschtritt ME, Mathews CA. Genetics and family models of hoarding disorder. In: Frost RO, Steketee G, editors. Oxford Handbook of Hoarding and Acquiring. New York, NY: Oxford University Press; 2014:159–176. 13. Brown TA, Di Nardo PA, Barlow DH. Anxiety Disorders Interview Schedule for DSM-IV: Lifetime Version (ADIS-IV-L). San Antonio, TX: Psychological Corporation; 1994.
Depression and Anxiety
14. Tolin D, Frost, RO, Steketee, G. A brief interview for assessing compulsive hoarding: The Hoarding Rating Scale-Interview. Psychiatry Res 2010;178:147–152. 15. Frost RO, Steketee G, Grisham J. Measurement of compulsive hoarding: saving inventory-revised. Behav Res Ther 2004;42(10):1163–1182. 16. Black DW. Compulsive buying disorder: a review of the evidence. CNS Spectr 2007;12(2):124–132. 17. Grisham J, Frost RO, Steketee G, Kim H-J, Hood, S. Age of onset of compulsive hoarding. J Anxiety Disorders 2006;20:675–686.
