A Developmental and Evolutionary Perspective on Obsessive-Compulsive Disorder: Whence and Whither Compulsive Hoarding?

Obsessive-compulsive disorder (OCD) is a prevalent condition that is heterogeneous in its clinical presentation. In addition, there is compelling evidence that supports the value of developmental and dimensional approaches to OCD as well as viewing OCD as part of a larger spectrum of disorders including tic disorders, trichotillomania, body dysmorphic disorder, and possibly compulsive hoarding. Specific subtypes of OCD also need to be considered, especially pediatric-onset and tie-related OCD, as well as pediatric autoimmune neuropsychiatric disorders associated with streptococcal (PANDAS) infections. As a result, there are conflicting views regarding how best to classify OCD in the next edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) (2, 3) . Three of the articles in this issue of the Journal are relevant to this topic and warrant our attention. Before each of these articles and their contributions to the field are considered, a brief overview may be of value.

Beginning around the second year of life most children develop a variety of rituals, habits, routines, and preferences, some of which resemble the behaviors associated with OCD (4, 5) . The idea that compulsive ritualistic behaviors may be normative in young children is not new. Gesell and his colleagues (1, 6) were among the first to recognize that young children—particularly those around the age of two and a half—begin to establish rigid routines that Gesell termed the “ritualisms of the ritualist.” Rather than addressing emotional needs, Gesell believed children engage in rituals to master the tasks of a specific developmental epoch, for instance, matters of feeding, toileting, and dressing. By the age of 3 years more than 80% of parents of children report the presence of a bedtime ritual (4) . Less well known is that for most children a regular sequence exists in which a need to arrange things “just right” or in a symmetrical pattern appears next, followed by the child’s being very concerned with dirt and germs, and finally, the need to collect and store objects. With the exception of hoarding, each of these behaviors peaks at age 3 years. Hoarding, in contrast, shows a monotonic increase, at least until the age of 6 years, when more than 60% of normal children display this trait (4, 5) . The earlier the age at which these concerns appear, the more advanced is the child’s developmental level. It is also striking that children with severe intellectual disabilities whose mental age remains in the low range show a persistence of many of these obsessive-compulsive behaviors (7, 8) . This larger body of data has led some to conclude that obsessive-compulsive behaviors are evolutionarily conserved (9, 10) and have a tendency to reappear normally at particular developmental transitions, such as falling in love or the advent of a firstborn infant (10, 11) .

These developmental data have also supported the view that OCD is composed of temporally stable symptom dimensions (12) and that clinically meaningful OCD arises as a result of a developmental sequence in which familial factors, both genetic and environmental, play critical roles. Within clinical populations the existence of these symptom dimensions appears beyond doubt, as confirmed by our recent meta-analysis that examined the results of more than 20 studies involving close to 5,000 OCD cases (13) . Briefly, these dimensions include 1) forbidden thoughts, 2) obsessions with symmetry and exactness and related ordering and arranging rituals as well as counting and repeating rituals, 3) contamination obsessions and washing compulsions, and 4) hoarding obsessions and compulsions. How useful to the clinician are these dimensions? Do they inform clinical practice? For example, do they reflect patterns of clinical response or longer-term outcome? More to the point, are they aligned with specific vulnerability genes, specific neural circuits, and neuropsychological traits? Although we advocate, with many of our colleagues (14), that a combined category-and-dimension approach makes the most sense for DSM-V and that specific subtypes and dimensions would be useful specifiers, more data are needed. The one exception to this conclusion is compulsive hoarding.

As presented in this issue by Pertusa and his colleagues in their naturalistic study of compulsive hoarding (15) and by Samuels et al. in their recent population-based study of hoarders (16), OCD patients with prominent hoarding symptoms differ from nonhoarding OCD patients across many important respects, including an earlier onset but later presentation for treatment. The pattern of comorbidity is also distinct, with hoarders of all types showing higher levels of social phobia and a greater number of personality disorders. Hoarding symptoms are consistently associated with poor response and premature dropout from cognitive-behavioral therapy for OCD. The picture with regard to pharmacological inventions is less clear-cut. Most reports indicate that hoarding is a strong predictor of poor response. However, in one recent study compulsive hoarders responded as well to paroxetine as nonhoarding OCD patients (17) . However, this study was limited by the fact that the compulsive hoarders took paroxetine for approximately 20–25 days longer than the OCD patients without hoarding. Furthermore, the percentage improvement in the OCD group (about 30% were responders) was lower than would be typically expected. Compulsive hoarders have lower global functioning and more severe family and social disability. Hoarding patients, with or without OCD, are more likely to be living alone and to have significantly lower incomes. As Pertusa et al. show in this issue, hoarding is a highly familial disorder, with approximately 50% of patients reporting a first-degree relative with compulsive hoarding. However, Pertusa et al. also found a high rate (26%) of OCD in the first-degree family members in their group of hoarders without OCD, which would argue for keeping compulsive hoarding closely aligned with OCD in the DSM-V structure.

The most remarkable part of the hoarding story is that there have been a number of reports that demonstrate that compulsive hoarding can newly emerge as a result of brain injury. The greatest lesion overlap in hoarders is in the right medial prefrontal cortex, orbitofrontal cortex, anterior cingulate cortex, and adjacent white matter (18) . Although we do not fully understand the relationship between brain lesions and patterns of brain activity, both lesion and functional imaging studies have indicated the involvement of specific frontolimbic brain regions in compulsive hoarders (19) . These areas are largely distinct from the circuits often implicated in OCD proper (20) . This difference speaks the most directly to the fact that a separate set of neural circuits may underlie our evolutionarily conserved, and developmentally expressed, tendency to be collectors.

This brings us to the two remaining OCD papers in this issue. The first point of note is that both groups specifically excluded OCD patients with prominent hoarding symptoms as well as patients with a diagnosis of Tourette’s disorder. This increasingly common practice suggests that experts in the field are beginning to consider that patients with a personal or family history of tics and those with prominent hoarding symptoms are fundamentally different from other OCD patients. Both nonhoarding OCD patients and compulsive hoarders may be better served by diagnostically separating these two conditions. Our scientific understanding and treatment of both conditions may be held back by the lumping together of these two conditions that appear phenotypically, genetically, and neurobiologically distinct.

As new imaging technologies emerge, we can be confident that they will contribute to our knowledge of how best to lump and/or split OCD into component parts. Regrettably, the story is far from clear. This month Menzies et al. (21) report a diffusion tensor imaging study that included both adult patients with OCD and an equal number of unaffected first-degree relatives. Remarkably, both the patients and their relatives showed white matter changes: a significant increase in fractional anisotropy in a right medial frontal region and significantly reduced fractional anisotropy in a large region of the right inferior parietal white matter. A similar reduction in fractional anisotropy within an overlapping area within the right inferior parietal region was previously reported by Szeszko et al. in a group of patients with childhood-onset OCD (22) . The existence of similar abnormalities in the unaffected first-degree relatives suggests that these are not compensatory changes but rather potential trait markers that could serve as markers of vulnerability. A good deal of additional work is needed to further replicate these findings and evaluate their potential value as endophenotypes. First, how interrelated are these two white matter findings? Are they separate vulnerabilities or does one predict the other? These questions are important as they will suggest either a single aberrant circuit or possibly two or more, potentially separable circuits. Second, is either finding related to the deficits in inhibitory control, cognitive flexibility, visuospatial ability, nonverbal intelligence, and executive planning that are frequently reported in OCD patients and unaffected family members with OCD (19, 23, 24) ? Given the Cambridge group’s interest in this area, we will likely know the answer to this question shortly. Indeed, their recent report in Science (25), in which they identified reduced activation of several cortical regions, including the lateral orbitofrontal cortex, in both OCD patients and a matched group of unaffected first-degree relatives, during a reversal learning task adds support to this being a potentially valuable endophenotype.

Third, given the fact that the majority of patients had either prominent contamination obsessions and washing compulsions or prominent checking compulsions, it will be fascinating to see if either of these white matter findings is closely related to any OCD symptom dimension. Our a priori hypothesis is that the spatial/attentional circuit proposed by Menzies et al. (20) that involves the parietal regions will be associated with deficits in inhibitory control and related to obsessions of symmetry and a need for things to feel or look “just right” and to compulsive checking, ordering, and arranging. Likewise, the circuit described by Menzies et al. as the “affective” circuit we would expect to be more related to contamination obsessions and washing compulsions as well as to obsessions about harm and the need for related compulsive checking.

Finally, the new report by Szeszko et al. (26), also in this issue, indicates that in addition to white matter changes, some pediatric patients with OCD have structural changes in regions comprising cortical-striatal-thalamic-cortical circuits. Specifically, using voxel-based morphometry, this team of investigators found that, compared to healthy volunteers, pediatric OCD patients had significantly more gray matter bilaterally both in the orbital frontal cortex and in the putamen. Both of these structures have been repeatedly implicated in previous studies (19) . Two great strengths of this study were that all of the patients were drug naive, so the data are free of the potential confound of prior use of selective serotonin reuptake inhibitors, and that the patient’s predominant domain of OCD symptoms is documented. The patients with either predominant contamination/cleaning or aggressive/checking symptoms had significantly more gray matter bilaterally in both the putamen and orbital frontal cortex, suggesting that these structures may contribute to the circuitry that supports these OCD symptom dimensions. In line with this finding, neuroimaging studies have found that OCD patients with predominantly washing symptoms show increased neural responses to washing-related stimuli and to disgusting photographs in specific brain regions, including the ventrolateral prefrontal cortex, anterior insula, and putamen/globus pallidus (27 – 29) .

The completion of human genome sequencing provides an extraordinary opportunity to identify the genetic basis of disorders of brain function. Although none of these reports in this issue of the Journal addresses the role of genes in OCD, we await the integration of genetic and neuroimaging research. It is likely that the genetic structure of common disorders differs substantially from current diagnostic classifications based on symptoms (30) . Our prediction is that some genes will be specific to certain subtypes or dimensions, while others will be “generalist” genes that influence the expression of OCD and closely related disorders, possibly including major depression and other internalizing disorders as well as body dysmorphic disorder, trichotillomania, tic disorders, and compulsive hoarding (31) . These generalist genes may exist within modules of coexpressed genes that are functionally related. By using this framework, it will be worthwhile to determine whether overlapping transcriptional networks underlie the expression of the obsessive-compulsive spectrum of normal phenomena as they are expressed within specific evolutionarily conserved neural networks. Then when these networks become dysregulated, for whatever reason, OCD and related disorders emerge as disorders of mind, brain, and behavior.