Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239

Contents lists available at SciVerse ScienceDirect

Progress in Neuro-Psychopharmacology & Biological

Psychiatry
journal homepage: www.elsevier.com/locate/pnp

Symptom dimensions are associated with age of onset and clinical course of
obsessive–compulsive disorder
Stephen A. Kichuk a, Albina R. Torres b, Leonardo F. Fontenelle c, Maria Conceição Rosário d,
Roseli G. Shavitt e, Eurípedes C. Miguel e, Christopher Pittenger a, f, g, Michael H. Bloch f, g,⁎
a
Department of Psychiatry, Yale University, United States
b
Department of Neurology, Psychology and Psychiatry, Botucatu Medical School, Univ Estadual Paulista, Botucatu (SP), Brazil
c
Anxiety and Depression Research Program — Institute of Psychiatry, Federal University of Rio de Janeiro, Rio de Janeiro (RJ), Brazil
d
Department of Psychiatry, Federal University of São Paulo Medical School, São Paulo (SP), Brazil
e
Department of Psychiatry, University of São Paulo Medical School, São Paulo (SP), Brazil
f
Department of Psychology, Yale University, United States
g
Child Study Center, Yale University, United States

a r t i c l e i n f o a b s t r a c t

Article history: Meta-analysis of the heterogeneous symptoms of obsessive–compulsive disorder (OCD) has found a four-factor
Received 24 November 2012 structure of symptom dimensions consisting of cleaning, forbidden thoughts, symmetry, and hoarding. Research
Received in revised form 2 February 2013 into age of onset of symptom dimensions has yielded inconsistent results, and it is unknown whether symptoms
Accepted 5 February 2013
along these dimensions differ in their clinical course. We assessed age of onset and clinical course of different
Available online 12 February 2013
OCD symptom dimensions in a large cohort of adult patients. Nine-hundred fifty-five subjects were assessed
using the Dimensional Yale–Brown Obsessive–Compulsive Scale. For age of onset analysis, we tested across
three methods of classification: (1) primary (more severe) symptom dimension (2) clinically significant symp-
toms within a dimension or (3) any symptoms within a dimension. Age of onset was defined as the earliest age of
onset reported for any individual item within a symptom dimension. For analysis of different types of clinical
course, we used chi-square tests to assess for differences between primary symptom dimensions. OCD symptoms
in the symmetry dimension had an earlier age of onset than other OCD symptom dimensions. These findings
remained significant across all three methods of classification and controlling for gender and comorbid tics. No
significant differences were found between the other dimensions. Subjects with primary OCD symptoms in
the forbidden thoughts dimension were more likely to report a waxing-and-waning course, whereas symmetry
symptoms were less likely to be associated with a waxing-and-waning course.
© 2013 Published by Elsevier Inc.

1. Introduction of cleaning, forbidden (aggressive/sexual/religious) thoughts, symme-

try (ordering/counting/repeating) and hoarding (Bloch et al., 2008). Re-
Obsessive–compulsive disorder (OCD) is a highly heterogeneous cent studies suggest that these symptom dimensions differ in genetic
condition (Miguel et al., 2005). While current nosology (DSM-IV and association (Hasler et al., 2007; Samuels et al., 2007), neurocircuitry
ICD-10) considers OCD to be a unitary entity, there is great variability (Mataix-Cols et al., 2004; Rauch et al., 1998), and response to pharma-
in symptomatic presentation (Lochner and Stein, 2003). Indeed, symp- cological (Landeros-Weisenberger et al., 2010; Mataix-Cols et al.,
toms can present in such dramatically different ways that several indi- 1999) and behavioral (Mataix-Cols et al., 2002) treatments.
viduals with equally severe OCD may have no specific symptoms in It remains unclear whether symptoms along the different dimensions
common (Bloch et al., 2008; Ferrao et al., 2006). have a different natural history. Several studies have yielded conflicting
OCD symptom dimensions are a tool for capturing this heterogene- results, with some indicating differences in age of onset (Honjo et al.,
ity. Factor analytic studies have yielded a fairly consistent four-factor 1989; Mataix-Cols et al., 1999; Minichiello et al., 1990) and others not
structure of OCD symptom dimensions across the lifespan, consisting finding any age-related patterns (Rettew et al., 1992). Interpretation of
these conflicting results is further burdened by the limited sample sizes
differing methodology; particularly problematic is their different
Abbreviations: OCD, Obsessive-Compulsive Disorder; ANOVA, Analysis of Variation; methods of assessing and classifying symptom dimensions. For instance,
DY-BOCS, Dimensional Yale-Brown Obsessive Compulsive Scale; FT, Forbidden several common OCD symptoms are often evaluated in a manner that
Thoughts; SYM, symmetry; CLEAN, Cleaning; HRD, Hoarding.
⁎ Corresponding author at: Child Study Center, Yale University School of Medicine,
leaves it unclear which dimension certain symptoms belong within
P.O. Box 2070900, New Haven, CT 06520, United States. (e.g. was “checking” due to harm-related obsessions, or due to
E-mail address: michael.bloch@yale.edu (M.H. Bloch). obsessions that something was not done correctly). Additionally,

0278-5846/$ – see front matter © 2013 Published by Elsevier Inc.

http://dx.doi.org/10.1016/j.pnpbp.2013.02.003
234 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239

participants generally only indicated whether or not particular classes of sensitivity analysis we also report results for sexual/religious and aggres-
symptoms were present (and in some cases whether they were a major sive OCD symptoms separately. For analysis involving primary symptom
problem), meaning there was no distinction made between symptoms dimension, subjects that scored highest on their worst-ever ratings in the
at varying levels of severity. This distinction is important, as without it, na- miscellaneous dimension or had a tie between two or more symptom di-
scent and minor symptoms within a dimension might have been coded in mensions were excluded. Clinically significant symptoms within a dimen-
the same manner as well-established and problematic symptoms. Dis- sion was defined by having a worst-ever DY-BOCS score in a dimension
tinct patterns of onset and natural history may be found at different sever- equal to or greater than 6, which roughly corresponds to a score on the
ity thresholds — that is, if one considers all symptoms, or only clinically more widely used Yale–Brown Obsessive Compulsive Scale (Y-BOCS)
significant symptoms, or only primary symptoms. Further, identifying of 16, a standard threshold to identify clinically significant symptoms.
whether different symptoms are associated with distinct clinical course Any symptom within a dimension was defined as having a worst-ever
can potentially help clinicians to provide more accurate prognostic infor- score greater than 0 in that dimension. Thus, a participant could (and
mation and identify symptoms that demand treatment priority. often did) qualify for more than one dimension when symptom dimen-
In an effort to enhance understanding of OCD symptom dimensions, sions were defined by clinical significance or presence of symptoms but
we examined age of onset and clinical course of OCD symptomatology could only qualify for one primary dimension.
according to symptom dimension in a large cohort of OCD patients. Age of onset for each symptom dimension was determined by the
We hypothesized that differences in age of onset would exist between earliest age of onset reported for any individual item within a symptom
dimensions, with the symmetry dimension having the earliest onset. dimension on the DY-BOCS checklist. This information was reported at
Given that converging evidence (i.e. genetics, neuroimaging, clinical the same time to when a clinical diagnosis was established. Clinical
evaluation) suggests that OCD symptom dimensions may have distinct course of OCD was determined using a graphical-based question in
neurobiological underpinnings and differential response to treatment, which individuals were given six possible graphical options to describe
we additionally hypothesized that OCD symptom dimensions would their OCD course. These options are depicted in Fig. 1 and include
display differences in their clinical course. (1) constant, (2) episodic, (3) waxing-and-waning, (4) deteriorating
then constant, (5) progressively deteriorating and (6) other. Subjects
2. Methods who chose “other” were excluded from this analysis.
Analyses were conducted in SPSS 19.0. One-way omnibus analysis of
2.1. Participants variance (ANOVA) was used to test for overall significance in age of
onset of DY-BOCS symptom dimensions. Analyses were conducted testing
A sample of 1001 OCD patients were recruited from 7 sites located primary symptoms, clinically significant symptoms, and any symptoms.
across Brazil as part of the Brazilian Research Consortium on Obsessive– Post-hoc pairwise comparisons were then conducted to identify signifi-
Compulsive Spectrum Disorders from 2005 to 2010. A full description of cant differences between OCD symptom dimensions when the overall
the recruitment and assessment procedure of this sample has been pub- test was significant. For analysis of the association between primary
lished elsewhere (Miguel et al., 2008). Due to the possibility of the clin- symptom dimension and type of clinical course, chi-square tests were uti-
ical course not having fully emerged in adolescents, it was decided prior lized. When the overall chi-square test was significant, individual
to data analysis that those under age 18 would be excluded; fifty-six post-hoc tests were performed to determine which clinical course options
subjects were excluded as a result. Nine-hundred fifty-five adults aged were reported at significantly increased or decreased rates for each
18–65 were ultimately included in our analyses. dimension.
For results related to the primary symptom dimension, we conducted
2.2. Procedure additional analyses to determine the possible confounding effects of gen-
der and comorbid tic disorders on our findings, as male gender and co-
Each site's ethics review boards reviewed and approved all methods. morbid tic disorders have been associated with both an earlier age of
Experts in OCD trained to reliably use protocol instruments interviewed onset (de Mathis et al., 2009) and increased likelihood of symptoms in
participants. Participants with primary OCD (per DSM-IV criteria) were particular OCD dimensions (Labad et al., 2008; Rosario-Campos et al.,
included. An OCD diagnosis was established by administration of the 2006). For the age of onset data, we added gender and tic disorder as ad-
Structured Clinical Interview for DSM-IV (First et al., 1995). OCD was ditional covariates in a one-way ANOVA in SPSS.
considered the primary psychiatric condition when this was the disorder The evidence concerning a relationship between gender, the pres-
for which the patient sought treatment and the diagnosis was verified by ence of tic disorders, and longitudinal course of symptoms is consid-
an experienced clinician. This diagnostic evaluation was conducted at the erably sparser, but there exists some evidence, at least in pediatric
same session or in close temporal proximity to when symptom ratings OCD, that gender and presence of a comorbid tic disorder is associated
were conducted. Exclusions included comorbid schizophrenia or any with adulthood outcome (Bloch et al., 2006, 2009). Therefore, we decid-
condition that interfered with participants' ability to effectively take ed to examine possible confounders in this analysis as well. The Mantel–
part in the protocol. Demographic and clinical data was obtained through Haenszel chi-square test was used to assess confounding across gender
use of a questionnaire produced by the Consortium. and tic disorder strata.
OCD symptom severity within each symptom dimension was
assessed with the Dimensional Yale–Brown Obsessive–Compulsive 3. Results
Scale (DY-BOCS) (Rosario-Campos et al., 2006). The time spent on
symptoms, level of anxiety, and level of interference are each rated on 3.1. Participants
a 0 (least severe) to 5 (most severe) scale, for a total score of 0–15, for
each dimension at the current and worst-ever time period. For the age Demographics characteristics of subjects are depicted in Table 1.
of onset analysis, OCD symptom dimensions of subjects were classified
in three ways: (1) primary symptom dimension (2) clinically significant 3.2. Age of onset
symptoms within a dimension or (3) any symptoms within a dimension.
The DY-BOCS ratings of symptom dimensions were restructured into the We found a significant association between age of onset and primary
four symptom dimensions identified by a large meta-analysis of previous dimension of OCD symptoms F(3, 519) =5.43, pb 0.001. Symptoms in the
studies in the area (Bloch et al., 2008). Severity in the forbidden thoughts symmetry dimension (age of onset=13.6±8.6) were associated with
dimension was defined for a subject as the higher of the severity scores an earlier age of onset than symptoms in the hoarding (age of onset=
in either the sexual/religious or aggressive DY-BOCS dimensions. As a 18.3±11.3, pb .05), cleaning (age of onset=16.9±8.3, pb .01), and
 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239 235

Fig. 1. Graphical question that probed clinical course of OCD symptoms.

forbidden thoughts dimensions (age of onset=16.9±8.9, pb .001). No 14.4± 7.7) had an earlier age of onset than hoarding (age of onset =
other significant differences in age of onset of symptoms were found be- 15.8± 8.8, p b .01), cleaning (age of onset = 16.5± 8.8, p valueb 0.001)
tween the OCD symptom dimensions. Fig. 2A depicts a cumulative histo- and forbidden thoughts (age of onset = 16.1± 8.7, p b .001). No other
gram of the age of onset of symptoms in each subject's primary OCD significant differences in age of onset of symptoms were found between
dimension when defined by the primary dimension of symptomatology the OCD symptom dimensions. Fig. 2B depicts a cumulative histogram
in each subject. of the age of onset of symptoms in each OCD dimension when defined
We additionally conducted a sensitivity analysis where symptoms in by the presence of lifetime clinically significant symptoms. When a sen-
the forbidden thoughts dimensions were separated into either sexual/ sitivity analysis was conducted separating the forbidden thoughts di-
religious or aggressive symptoms. Both sexual/religious (age of onset= mension into two dimensions, both sexual/religious (age of onset =
16.6±7.3, pb .001) and aggressive (age of onset=19.3±10.5, pb .001) 16.5± 8.1, p b .001) and aggressive (age of onset = 17.5 ± 9.3, p b .001)
OCD symptoms were associated with a significantly later age of onset OCD symptoms were associated with a significantly later age of onset
than symptoms in the symmetry dimension. The age of onset of sexual/ than symptoms in the symmetry dimension. The age of onset of sexu-
religious and aggressive OCD symptoms did not significantly differ from al/religious and aggressive OCD symptoms did not significantly differ
other OCD symptom dimensions or from each other. from other OCD symptom dimensions or from each other.
A significant association between OCD symptom dimension and age A significant association between OCD symptom dimension and age
of onset was also demonstrated when symptom dimension was defined of onset was demonstrated when symptom dimension was defined by
by clinically significant symptoms (F(3, 2366) = 8.53, p b .0001). As previ- the presence of any lifetime symptoms (F(3, 2738) =8.67, pb .0001).
ously, OCD symptoms in the symmetry dimension (age of onset= Pairwise comparisons found that symptoms in the symmetry dimension
(age of onset=14.5±7.8) had an earlier onset than symptoms in the
hoarding dimension (age of onset=16.1±8.8, pb .001), the cleaning di-
Table 1 mension (age of onset=16.5±8.8, pb .001), and the forbidden thoughts
Demographics.
dimension (age of onset=16.3±8.8, pb .001). No other significant dif-
Demographic characteristics Mean ± SD ferences in age of onset of symptoms were found between the OCD
N (%) symptom dimensions. Fig. 2C depicts a cumulative histogram of the
Age 35.82 ± 12.49 age of onset of symptoms in each OCD dimension when defined by the
Male gender (%) 400 (42%) presence of any lifetime symptoms. When a sensitivity analysis was
YBOCS total score 25.6 ± 7.46 conducted separating the forbidden thoughts dimension into two di-
YBOCS obsessions subscale 12.73 ± 3.91
mensions, both sexual/religious (age of onset=16.7±8.6, p b .001) and
YBOCS compulsions subscale 12.87 ± 4.11
aggressive (age of onset=17.5±9.3, pb .001) OCD symptoms were as-
Lifetime comorbid conditions sociated with a significantly later age of onset than symptoms in the
Generalized anxiety disorder 321 (33.6%) symmetry dimension. The age of onset of sexual/religious and aggressive
Social anxiety disorder 312 (32.7%)
OCD symptoms did not significantly differ from other OCD symptom di-
Post-traumatic stress disorder 97 (10.2%)
Panic disorder 44 (4.6%)
mensions or from each other.
Specific phobia 296 (31%)
Major depression 549 (57.5%)
Bipolar disorder 51 (5.3%) 3.3. Effects of gender and comorbid tics on age of onset
Body dysmorphic disorder 105 (11%)
Trichotillomania 45 (4.7%) The association between symptom dimension and age of onset
Tic disorder 264 (27.6%) remained significant when we adjusted for gender and tic comorbidity.
Primary OCD dimension
In a model adjusting for the presence of comorbid tics, a significant
Forbidden thoughts 202 (22%) main effect was found for primary symptom dimension (F(3, 511) =
Symmetry 148 (16.1%) 2.84, p = .038) and for comorbid tics (F(1, 511) = 9.69, p = .002); OCD pa-
Cleaning 146 (15.9%) tients with comorbid tics had an earlier age of onset. There was also a
Hoarding 27 (2.8%)
significant interaction between primary symptom dimension and co-
Previous treatments morbid tics F(3, 511) = 2.83, p = .038. Fig. 3A depicts age of onset of
Psychotherapy 614 (61.3%) OCD symptom dimensions stratified by presence of comorbid tics.
Serotonin reuptake inhibitors 679 (71%) OCD patients without comorbid tics reported an earlier age of onset
Antipsychotic augmentation 194 (20.3%) for symptoms in the symmetry dimension whereas OCD patients with
YBOCS = Yale–Brown Obsessive–Compulsive Scale. comorbid tics did not. OCD patients with comorbid tics had earlier age
236 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239

Fig. 2. A–C: Cumulative histograms of age of onset by OCD symptom dimension. Cumulative histograms depict the proportion of subjects experiencing onset of symptoms within
each OCD symptom dimension by a given age. Figures define presence of symptoms within OCD dimensions differently — Fig. 2A primary symptom dimension, Fig. 2B by presence
of clinically significant symptoms and Fig. 2C by presence of any symptoms.
 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239 237

the effect of primary symptom dimension on clinical course was only

at trend level in the patients with tics (χ2(df=12, n=115) = 16.2, p =
0.18) and without comorbid tics (χ2(df=12, n =351) = 20.6, p = 0 .06),
perhaps due to the lower number of subjects in these subgroups and
therefore to reduced statistical power.
In OCD patients without tics, subjects with primary symptoms in the
forbidden thoughts dimension more often reported a waxing-and-
waning course (χ 2(df=1, n=351) = 9.9, p = .002) and less often reported
a deteriorating then constant course at trend levels (χ 2(df=1, n=351) =
3.2, p = .07). Subjects with primary symmetry symptoms less often
reported a waxing-and-waning course (χ 2(df=1, n=351) = 4.8, p = .03).
In OCD patients with comorbid tics, subjects with primary symptoms
in the forbidden thoughts dimension were also more likely to report a
waxing-and-waning course (χ2(df=1, n=115) =3.8, p=.05) and less like-
ly to report a deteriorating then constant course (χ2(df=1, n=115) =5.6,
p=.02).
We also examined the potential confounding effects of gender on
the association between OCD symptom dimension and clinical course
(Fig. 4A, D and E). There was a strong, statistically significant associa-
tion between OCD symptom dimension and clinical course in females
(χ 2(df = 12, n = 262) = 31.6, p = 0.002) but not males (χ 2(df = 12, n = 207) =
9.8, p = 0.63).
Females with primary symptoms in the forbidden thoughts
dimension more often reported a waxing-and-waning course
(χ 2(df = 1, n = 262) = 14.2, p = .0002) and less often reported a deteri-
orating then constant course (χ2(df=1, n=262) =4.7, p =.03). Females
with primary symmetry symptoms were less likely to report a waxing-
and-waning course (χ2(df=1, n=262) =6.1, p=.01). Additionally, females
Fig. 3. A–B: Age of onset in symptom dimensions stratified by potential confounders. with primary cleaning symptoms were more likely to report a deteriorat-
Fig. 3A is stratified by the presence (blue) or absence (green) of a comorbid tic disorder.
Fig. 3B is stratified by gender (males — blue and females — green). (For interpretation of
ing course (χ2(df=1, n=262) =10.2, p=.001) and less likely to report an
the references to color in this figure legend, the reader is referred to the web version of episodic course (χ2(df=1, n=262) =10.2, p=.001).
this article.) Males did not demonstrate any significant association between
OCD symptom dimensions and clinical course. The significant associ-
ations present in the overall sample all occurred in the same direction
of onset of cleaning and hoarding symptoms compared to OCD patients among males only but did not approach statistical significance (forbidden
without comorbid tics. thoughts/waxing-and-waning course: χ2(df=1, n=204) =1.0, p=.31; for-
In a model adjusting for gender, one-way ANOVA found a significant bidden thoughts/deteriorating then constant course: χ2(df=1, n=204) =
main effect for primary symptom dimension (F(3, 511) = 4.6, p = .003) 1.5, p=.22; symmetry/waxing-and-waning course: χ2(df=1, n=204) =0,
but not gender (F(1, 511) = .92, p = .338). There was an interaction of pri- p=.96).
mary dimension and gender trending towards but not reaching signifi-
cance (F(3, 511) = 2.6, p = .051). A greater difference in age of onset by 4. Discussion
symptom dimension tended to be reported by females compared to
males (Fig. 3B). The heterogeneity of OCD symptoms can be described along several
orthogonal dimensions (1). We report that symptoms along these dis-
3.4. Clinical course tinct dimensions are associated with different natural histories. Specifi-
cally, symmetry symptoms have an earlier onset than symptoms in
There was a significant association between primary OCD symp- other dimensions. Longitudinally, patients with primary symptoms
tom dimension and reported clinical course of OCD symptoms in the forbidden thoughts dimension were more likely to report a
(χ 2(df = 12, n = 466) = 25.7, p = .012). Participants with primary OCD waxing-and-waning course whereas patients with primary symptoms
symptoms in the forbidden thoughts dimension were more likely in the symmetry dimension were less likely to report a waxing-and-
to endorse a fluctuating course (χ 2(df = 1, n = 466) = 11.9, p = .0006) waning course. Females with primary OCD symptoms in the cleaning
and less likely to endorse a deteriorating then constant course dimension were more likely to report a progressively deteriorating
(χ 2(df=1, n=466) = 7.1, p = .008). Participants with primary symptoms course.
in the symmetry dimension were less likely to report their clinical This association between age of onset and symptom dimension is in
course as waxing-and-waning (χ2(df= 1, n=466) = 5.8, p = .02). Partici- line with previous research on this topic (Honjo et al., 1989; Minichiello
pants with primary symptoms in the cleaning dimension more often et al., 1990). Direct comparisons are difficult, as previous studies used
reported a deteriorating course at trend levels (χ 2(df=1, n =466) = 3.4, very different methods of assessment and classification. Nevertheless,
p = .07). Fig. 4A depicts the proportion of patients reporting each type the association between symmetry symptoms and earlier age of onset
of clinical course by primary OCD symptom dimension. has been reported previously (Denys et al., 2004; Mataix-Cols et al.,
1999; Matsunaga et al., 2010). We show this association to be robust
3.5. Effects of gender and comorbid tics on clinical course to method of analysis, as our results did not vary on how symptom di-
mension was defined (Fig. 2B–D).
We examined the potential confounding effects of comorbid tics on In contrast to an earlier study, we did not find an association be-
the associations we found between OCD symptom dimensions and clin- tween an earlier onset and symptoms of the forbidden thoughts (FT) di-
ical course (Fig. 4A, B and C). The pattern of association was similar in mension (Mataix-Cols et al., 1999). Differences in methodology may
subjects with and without tic disorders, although the significance of explain the different results between the two studies. Specifically, the
238 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239

Fig. 4. A–E: Clinical course of OCD symptom dimensions. Fig. 4A includes the entire sample, Figs. 4B and C depict OCD patients with and without comorbid tic disorders respectively.
Fig. 4D and E depict women and men with OCD respectively. Turquoise denotes a deteriorating course; purple denotes a worsening with plateau course; green denotes a constant
course; red denotes an episodic course; blue denotes a waxing and waning course. Abbreviations: FT, Forbidden Thoughts; SYM, symmetry; CLEAN, Cleaning; HRD, Hoarding. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

earlier study assessed global age of onset of OCD, while we assessed to other symptom dimensions, and a consequent lack of statistical
dimension-specific age of onset (Mataix-Cols et al., 1999). It may be power may partially explain this lack of association. However, this was
that patients with primary FT symptoms in adulthood had earlier onset a large sample, and included over 300 patients that reported some
of symptoms in another dimension. This would obscure the age of hoarding symptoms (which were included in the analyses presented in
onset of the FT dimension itself. Our methodology avoids this confound. Fig. 2B and C even if they were not the primary OCD symptom). This
The association of symmetry dimension symptoms with an earlier lack of distinct age of onset for hoarding symptoms should not be viewed
age of onset may be understood in a developmental context. A need as evidence against the growing consensus that these symptoms are dis-
for symmetry, ordering and arranging compulsions, and “just right” tinct from the rest of OCD; many other diagnostically discrete conditions
feelings are very common in young children (b4 years of age) (Evans such as depression, schizophrenia, and substance abuse also have similar
et al., 1999). In some cases, these symptoms can persist into adulthood. ages of onset as OCD (Torres et al., 2012).
The greater prevalence of symmetry-related issues in children (as com- Our study represents a significant advance over previous investiga-
pared to other OCD-type issues) could explain our observed difference tions of the relationship between symptom subtypes and the natural his-
in age of onset. tory of OCD. We analyzed a large, clinically referred sample of 955 adults
The association we found between the forbidden thoughts dimension with OCD, which provided increased power relative to previous research
and a more fluctuating course may be due to active coping attempts by with much smaller samples. Our use of the standard DY-BOCS provided a
affected individuals. These symptoms are often triggered by the presence continuous measure of symptom severity, whereas previous research
of other people or objects (e.g. obsessions about accidentally harming only provided categorical data. Additionally, we were able to assess the
others), and a primary coping strategy for individuals with such symp- age of onset of each OCD symptom dimension, as opposed to previous re-
toms is the avoidance of external symptom triggers. However, since search that relied on global age of onset. Nevertheless, there remain sev-
complete avoidance is rarely possible, chance encounters with symp- eral limitations to our study. As our data was derived from retrospective
tomatic triggers may result in symptom spikes, which may at least par- patient reports, its accuracy may be affected by factors such as level of in-
tially account for our finding. sight, comorbid conditions, or the distortions in memory that can always
We found no association between hoarding symptoms and age of complicate retrospective data. Our main question probing retrospective
onset, relative to other OCD symptoms. This result is somewhat surpris- clinical course has never been validated. In addition, treatment factors
ing given the accumulating evidence suggesting that hoarding symp- were not controlled for, and thus it is possible that the treatment history
toms are distinct from the rest of OCD (Saxena, 2007). The relatively of our subjects affected the results, particularly as related to clinical
few number of subjects with primary hoarding symptoms, compared course.
 S.A. Kichuk et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 44 (2013) 233–239 239

In conclusion, we find that different dimensions of OCD symptoms Evans DW, Gray FL, Leckman JF. The rituals, fears and phobias of young children: in-
sights from development, psychopathology and neurobiology. Child Psychiatry
are associated with different patterns of natural history. Specifically, Hum Dev 1999;29:261–76.
symmetry symptoms tend to have the earliest onset and a less fluctuat- Ferrao YA, Shavitt RG, Bedin NR, de Mathis ME, Carlos Lopes A, Fontenelle LF, et al.
ing longitudinal course; forbidden thoughts symptoms more often had a Clinical features associated to refractory obsessive-compulsive disorder. J Affect
Disord 2006;94:199–209.
waxing-and-waning course; and cleaning symptoms were more likely to First M, Spitzer R, Gibbon M, Williams J. Scructured clinical interview for DSM-IV axis I
have a progressively deteriorating course, among females only. Future disorders-Patient edition. New York: Biometrics Reseach Department, New York
studies prospectively evaluating the age of onset and course of symptom Psychiatric Institute; 1995.
Hasler G, Pinto A, Greenberg BD, Samuels J, Fyer AJ, Pauls D, et al. Familiality of factor
dimensions would be beneficial. Age of onset can be evaluated in several analysis-derived YBOCS dimensions in OCD-affected sibling pairs from the OCD
ways (initial subclinical symptoms vs. initial clinical symptoms; global Collaborative Genetics Study. Biol Psychiatry 2007;61:617–25.
symptoms vs. specific subtypes of symptoms). Heterogeneity in the def- Honjo S, Hirano C, Murase S, Kaneko T, Sugiyama T, Ohtaka K, et al. Obsessive-compulsive
symptoms in childhood and adolescence. Acta Psychiatr Scand 1989;80:83–91.
inition of symptom onset may explain the heterogeneity in research on
Labad J, Menchon JM, Alonso P, Segalas C, Jimenez S, Jaurrieta N, et al. Gender differ-
these matters (although we found similar results when onset was de- ences in obsessive-compulsive symptom dimensions. Depress Anxiety 2008;25:
scribed in different ways; Fig. 2A–C). 832–8.
Landeros-Weisenberger A, Bloch MH, Kelmendi B, Wegner R, Nudel J, Dombrowski P, et al.
Dimensional predictors of response to SRI pharmacotherapy in obsessive-compulsive
Acknowledgments disorder. J Affect Disord 2010;121:175–9.
Lochner C, Stein DJ. Heterogeneity of obsessive-compulsive disorder: a literature re-
view. Harv Rev Psychiatry 2003;11:113–32.
This study was supported by the Fundação de Amparo à Pesquisa do
Mataix-Cols D, Marks IM, Greist JH, Kobak KA, Baer L. Obsessive-compulsive symptom
Estado de São Paulo (FAPESP, grant 2005/55628-8 to Dr. Miguel) and dimensions as predictors of compliance with and response to behaviour therapy:
from the Conselho Nacional de Desenvolvimento Científico e Tecnológico results from a controlled trial. Psychother Psychosom 2002;71:255–62.
(CNPQ, grant 573974/2008-0 to ECM). We thank Aristides V. Cordioli, Mataix-Cols D, Rauch SL, Manzo PA, Jenike MA, Baer L. Use of factor-analyzed symptom
dimensions to predict outcome with serotonin reuptake inhibitors and placebo in the
MD, PhD; Katia Petribu, MD, PhD; Christina H. Gonzalez, MD, PhD; treatment of obsessive-compulsive disorder. Am J Psychiatry 1999;156:1409–16.
Maria Alice de Mathis, MsC, and all colleagues from the Brazilian Mataix-Cols D, Wooderson S, Lawrence N, Brammer MJ, Speckens A, Phillips ML. Dis-
Research Consortium of Obsessive–Compulsive Spectrum Disorders tinct neural correlates of washing, checking, and hoarding symptom dimensions
in obsessive-compulsive disorder. Arch Gen Psychiatry 2004;61:564–76.
(C-TOC) for their help with data collection. Matsunaga H, Hayashida K, Kiriike N, Maebayashi K, Stein DJ. The clinical utility of
The authors acknowledge the National Institute of Mental Health symptom dimensions in obsessive-compulsive disorder. Psychiatry Res 2010;180:
support of the Yale Child Study Center Research Training Program 25–9.
Miguel EC, Ferrao YA, Rosario MC, Mathis MA, Torres AR, Fontenelle LF, et al. The
(MHB), National Institutes of Health grants K23MH091240 (MHB) Brazilian Research Consortium on Obsessive-Compulsive Spectrum Disorders:
and K08MH081190 (CP), the APIRE/Eli Lilly Psychiatric Research Fel- recruitment, assessment instruments, methods for the development of multicenter
lowship (MHB), the AACAP/Eli Lilly Junior Investigator Award (MHB), collaborative studies and preliminary results. Rev Bras Psiquiatr 2008;30:185–96.
Miguel EC, Leckman JF, Rauch S, do Rosario-Campos MC, Hounie AG, Mercadante MT,
the Trichotillomania Learning Center (MHB), NARSAD (MHB), the
et al. Obsessive-compulsive disorder phenotypes: implications for genetic studies.
Doris Duke Charitable Foundation (CP), and UL1 RR024139 from the Mol Psychiatry 2005;10:258–75.
National Center for Research Resources, a component of the National Minichiello WE, Baer L, Jenike MA, Holland A. Age of onset of major subtypes of
obsessive-compulsive disorder. J Anxiety Disord 1990;4:147–50.
Institutes of Health, and NIH roadmap for Medical Research (MHB,
Rauch SL, Dougherty DD, Shin LM, Alpert NM, Manzo PA, Leahy L, et al. Neural corre-
CP). lates of factor-analyzed OCD symptom dimensions. CNS Spectr 1998;3:37–43.
Rettew DC, Swedo SE, Leonard HL, Lenane MC, Rapoport JL. Obsessions and compul-
sions across time in 79 children and adolescents with obsessive-compulsive disor-
References der. J Am Acad Child Adolesc Psychiatry 1992;31:1050–6.
Rosario-Campos MC, Miguel EC, Quatrano S, Chacon P, Ferrao Y, Findley D, et al. The
Bloch MH, Craiglow BG, Landeros-Weisenberger A, Dombrowski PA, Panza KE, Peterson
Dimensional Yale-Brown Obsessive-Compulsive Scale (DY-BOCS): an instrument
BS, et al. Predictors of early adult outcomes in pediatric-onset obsessive-compulsive
for assessing obsessive-compulsive symptom dimensions. Mol Psychiatry 2006;11:
disorder. Pediatrics 2009;124:1085–93.
495–504.
Bloch MH, Landeros-Weisenberger A, Rosario MC, Pittenger C, Leckman JF. Meta-
Samuels J, Shugart YY, Grados MA, Willour VL, Bienvenu OJ, Greenberg BD, et al. Sig-
analysis of the symptom structure of obsessive-compulsive disorder. Am J Psychiatry
nificant linkage to compulsive hoarding on chromosome 14 in families with
2008;165:1532–42.
obsessive-compulsive disorder: results from the OCD Collaborative Genetics Study.
Bloch MH, Peterson BS, Scahill L, Otka J, Katsovich L, Zhang H, et al. Adulthood outcome
Am J Psychiatry 2007;164:493–9.
of tic and obsessive-compulsive symptom severity in children with Tourette syn-
Saxena S. Is compulsive hoarding a genetically and neurobiologically discrete syn-
drome. Arch Pediatr Adolesc Med 2006;160:65–9.
drome? Implications for diagnostic classification. Am J Psychiatry 2007;164:380–4.
de Mathis MA, Diniz JB, Shavitt RG, Torres AR, Ferrao YA, Fossaluza V, et al. Early onset
Torres AR, Fontenelle LF, Ferrao YA, do Rosario MC, Torresan RC, Miguel EC, et al. Clinical
obsessive-compulsive disorder with and without tics. CNS Spectr 2009;14:362–70.
features of obsessive-compulsive disorder with hoarding symptoms: a multicenter
Denys D, de Geus F, van Megen HJ, Westenberg HG. Use of factor analysis to detect
study. J Psychiatr Res 2012;46:724–32.
potential phenotypes in obsessive-compulsive disorder. Psychiatry Res 2004;128:
273–80.