Course and outcome of obsessive-compulsive disorder.

Obsessive-compulsive disorder (OCD) is generally believed to follow a chronic waxing and waning course. The onset of illness has a bimodal peak - in early adolescence and in early adulthood. Consultation and initiation of treatment are often delayed for several years. Studies over the past 2-3 decades have found that the long-term outcomes in OCD are not necessarily bleak and that at least half the treatment-seeking patients with OCD show symptomatic remission over long term. A short duration illness, of low severity that is treated early and intensively, with continued maintenance treatment over long term possibly has a good outcome. Recent studies have also identified neuroimaging and neuropsychological correlates of good outcome, but these need further replication. This paper presents an overview of conceptual issues and studies on long-term outcome of OCD and predictors of outcome.

INTRODUCTION

Community prevalence estimates of obsessive–compulsive disorder (OCD) in adults range from 0.5% to 3%.[12] OCD ranks among the top 10 causes of disability, worldwide.[3] Nosological and therapeutic advancements have improved recognition and early initiation of treatment for this disorder. Patients coming in for treatment often have pertinent questions about what might happen to their illness in the long run – “What happens if treatment is not taken?,” “What will happen to my illness over time?,” “Do different treatments have different effects on the illness?” It is important to address these questions in order that clinicians and patients make informed decisions about management. This chapter presents an overview of literature on the course and outcome of OCD in adults. The focus will be more on the long-term course and outcome and predictors of outcome, given the greater utility of long-term outcome data in clinical discussions rather than short-term benefits seen in treatment trials.

CONCEPTUAL ISSUES

Course

Course refers to the behavior of an illness over time, from its onset to the sequential stages, and forms it takes. In the case of OCD, two major types of course are seen – chronic and episodic. A chronic course implies near persistent presence of symptoms. Symptom severity might wax and wane, with phasic exacerbations and incomplete remissions; however, there is never a complete relief from symptoms. In episodic course, symptoms are present only during an episode, and for the remaining time, symptoms remit, with or without treatment. It is unclear what the symptom-remission interval should be to qualify for an episodic course in OCD. Ravizza et al. proposed a symptom-free interval of at least a month in a year for episodic OCD.[4] There is a paucity of operational definitions for the course of OCD. Thomsen proposed a clinical classification that is useful to understand the possible presentations OCD may have over longitudinal course of illness.[5] This classification identifies “no OCD,” “subclinical OCD,” “episodic OCD,” “chronic OCD,” and “true remission.” No OCD implies complete recovery from the index episode. Subclinical OCD has symptoms lasting <1 h/day without any functional consequences. In episodic OCD, remissions (with no or subclinical symptoms) and relapses are seen. Patients with chronic OCD have a continuous illness with significant distress and functional impairment. A true remission implies a no OCD status at follow-up without being on any treatment. This classification while clinically useful lacks duration criterion. For example, how long a patient needs to be in remission without being on treatment to qualify for “true remission” is not clear. In addition, patients may move from one status to another over a period of time. For example, those who have achieved remission may relapse after several years.[6]

Outcome

Outcome refers to the condition of an illness, at a defined time point. While course is longitudinal, the outcome is cross sectional. Given the dimensional and time-varying nature of psychopathology, outcomes have to be operationalized for research. Moreover, the outcome can be understood differently from symptomatic and functional paradigms.

Operational definitions for outcome of obsessive–compulsive disorder

Researchers have operationalized definition of outcome using different levels of symptomatic improvement and severity rating. A cutoff score on the Yale–Brown Obsessive–Compulsive Scale (YBOCS) total score is the most commonly used outcome measure.[78] Ratings on the YBOCS are based on both the patient's report and clinical judgment. The scale has separate ratings for obsessions and compulsions, with a maximum possible score ranging from 0 to 40. While most studies employ a YBOCS cutoff of 16[9] to define remission, scores of 12 and 14[10] have also been used. Lower scores correlate better with functional recovery and quality of life. The increasing understanding about the nuances and multifactorial nature of remission, especially the important role of functionality, paved the way for the development of international expert consensus guidelines for defining response, remission, recovery, and relapse in OCD.[11] Notably, in these guidelines, a change in YBOCS severity score has to be maintained for a minimum specified time period to ascertain remission, recovery, and relapse. It is hoped that these consensus guidelines will make future research on outcome more meaningful and comparable across studies.

“Symptomatic” versus “functional” outcome

A “symptomatic outcome” implies the reduction in severity of symptoms, as assessed by scores on a symptom rating scale. This is the most common type of outcome discussed and reported. However, it is well established among clinicians that a mere reduction in symptom severity does not always translate into improved “functional outcome” and “quality of life.”[12] “Patient-reported outcomes” can be different from severity ratings assessed by clinicians, especially in a disorder like OCD where all the symptoms might not even be apparent.[13] A specific scenario in OCD would be when remission of compulsions occurs earlier leading to a reduction in total severity scores; however, the persisting obsessions continue to significantly impair functioning and quality of life. It has also been noted that obsessions, along with depressive symptoms, predict impairment in quality of life, even when obsessions and compulsions are of comparable severity.[14] A study using signal detection analysis[10] demonstrated that on the YBOCS,[78] a score of ≤14 predicted symptom remission in a sample of over 250 patients; however, a lower cutoff (YBOCS ≤12) predicted wellness, defined as a combination of symptom remission, good quality of life, and high level of adaptive functioning. In another study, it was seen that posttreatment three groups could be delineated – those with strong symptom reduction and good quality of life gains, those with significant symptom reduction but continuing impaired quality of life, and a third group with no symptom reduction and no change/worsening in quality of life.[15] Thus, the interaction between symptom severity and quality of life is not linear. Needless to say, clinical intervention goals should prioritize functionality and quality of life.

THE COURSE AND OUTCOME OF “UNTREATED” OBSESSIVE–COMPULSIVE DISORDER

Literature on the course of untreated OCD is largely from studies conducted before effective treatments – serotonin reuptake inhibitors (SRIs) and cognitive behavior therapy (CBT) – were available. The longest ever follow-up study in OCD over 40 years reported outcomes in a clinical sample of 144 patients.[6] In this sample, 83% of the patients “improved,” while only 20% attained “full remission.” In this study, although an early recovery was a harbinger of good prognosis, it did not guarantee against a late relapse, for a fifth of even the fully remitted patients suffered relapses after being in remission for nearly two decades. A review of 13 studies from the pre-SRI/CBT era broadly concluded that “obsessional neurosis has a favorable course.”[16] This review noted that spontaneous remissions were common, and certain variables – outpatient treatment (versus hospitalization) and lower baseline severity of illness – predicted better outcomes. While the findings from this era seem positive, they must be cautiously interpreted in the current scenario. Diagnostic criteria, assessment methods, definitions of response, relapse, and recovery have been significantly modified since. In addition, these studies may not have excluded treatment-resistant patients and those who underwent neurosurgical treatments.

Possibly, the only report of a truly naturalistic outcome in OCD from recent times comes from a community-based study in Zurich.[17] This study has followed up nearly 600 individuals over 30 years. OCD was present in 5% of this sample and 27% had nonsyndromic OC symptoms. Using stringent criteria for remission – absence of OCD symptoms for 3 consecutive years – the researchers found that >60% of the sample remitted. Only about a third of the patients with OCD were professionally treated in this community cohort. Thus, there is some evidence that untreated OCD may remit spontaneously over several years.

THE COURSE AND OUTCOME OF OBSESSIVE–COMPULSIVE DISORDER IN RECENT CLINICAL FOLLOW-UP STUDIES

Onset of obsessive–compulsive disorder

The age at onset in OCD has been defined as the age at which obsessive–compulsive symptoms were first associated with distress, as recalled by the patient.[1819] Age at onset in OCD varies from preschool to elderly years, with two peaks occurring between 9–11 and 20–23 years.[2021] A juvenile onset may be seen in 30%–50% cases,[2223] whereas onset beyond the fourth decade occurs in <10% cases.[181924] Age at onset accounts for at least some of the heterogeneity seen in the clinical presentation and outcome of OCD. There is a larger burden of developmental morbidity and predisposition associated with younger onset, whereas elderly onset may have a higher loading of environmental and neurological influences.[25]

Course of obsessive–compulsive disorder

A chronic course, spanning over several decades, with waxing and waning symptom severity, is typical of OCD.[2627] The initial consultation and diagnosis may itself be delayed from a few years to several decades. It is not uncommon to find people presenting in their mid-20s with a history of at least subclinical symptoms from early childhood.[17] Beliefs about self-control, accompanied by unawareness about illness and spontaneous remissions in the waxing-waning course, may contribute to the delay in help seeking.[28] In two studies from India, a truly chronic, unremitting course was seen in <30% patients, while 30%–40% patients had “no OCD” at long-term follow-up.[2930] Also notable were the rates of subclinical OCD at around 25%–35%. By definition subclinical symptoms would not cause functional impairment, however, it is conceivable that they might act as risk factors for relapse.

An episodic course may also be seen in OCD. Older studies, from the 1970s to 1980s, reported a 15%–50% rate of episodic OCD, whereas some later reports give a <2% prevalence.[31] A large case series of 135 patients, with mean illness duration over 10 years, specifically analyzed for the proportion of episodic OCD found that 27% of the sample had an episodic course with periods of remission extending up to 6 months.[32] Another study that looked at only pediatric cases also reported that almost one-third of the children with OCD had an episodic course.[5] Rates of episodic OCD are highly variable across studies. Moreover, the implications of episodicity on treatment and outcome are poorly understood. Episodic OCD has been associated with certain clinical features, as enumerated in Table 1. Given the high comorbidity of bipolar disorder with episodic OCD,[38] some studies compared OCD patients with and without bipolar disorder comorbidity.[39] Interestingly, this format of analysis also showed that an episodic course of OCD was more common in the group with bipolar disorder comorbidity. The bipolar-OCD group was characterized by higher family loading of mood disorders, comorbidity with anxiety disorders, lesser severity of OCD, and poorer insight.[3940] Another interesting observation was that OCD symptoms worsened with depression and improved with mania, in the majority. A recent meta-analysis that looked at the diagnostic validity of bipolar-OCD comorbidity concluded that an episodic course is seen in 75% of comorbid cases, compared to 3% in other OCD cases;[41] symptoms generally tend to worsen with depression and improve with mania/hypomania; and antidepressant-induced switches are more common in the comorbid group. The authors commented that OCD with bipolar disorder appears to be a part of the bipolar diathesis, rather than an independent entity. Research on course of OCD, especially episodic OCD, needs further attention to understand prognostic and management implications.

Table 1

Clinical features associated with episodic (versus chronic) obsessive-compulsive disorder

The long-term outcome of obsessive–compulsive disorder

There have been several long-term naturalistic treatment outcome studies in the past 2–3 decades [Table 2]. There are significant variations in sampling, clinical characteristics, follow-up, and outcome assessments in these studies. All except one of these studies have followed up treatment-seeking patients, over a duration varying from 1 to 15 years. A majority of studies report outcomes on treatment with a combination of SRIs and CBT,[30424346474849515354555658596061626365] only a few with SRIs alone[294445505264] and only one study with CBT alone.[57] Outcome has been assessed based on YBOCS ratings. The operational definition for outcome varies across studies – a percentage reduction in YBOCS score,[4344] YBOCS score <16 at the final follow-up,[306063] YBOCS score <12 at the final follow-up,[51] YBOCS score <8 at the final follow-up,[61] and no OCD symptoms for 3 consecutive years.[17] These studies aimed to study not only the naturalistic treatment outcomes over long term but also the impact of different treatment strategies, the effect of comorbidities on outcome,[58] the role of family accommodation,[65] etc. The report of comorbidities in these studies is particularly notable. Comorbidities have been reported in one-third to four-fifths of the sample, with depression and anxiety disorders being the most common. It is possible that the variations in comorbidity account for the varying outcomes across studies. It is surmised, therefore, that while these studies shed light on the multiple considerations in understanding the long-term outcome of OCD, the heterogeneity across studies poses synthesis and interpretation challenges.

Table 2

Long-term follow-up studies in obsessive-compulsive disorder

A meta-analysis pooled the results of 17 long-term (follow-up duration of 1 year or more) outcome studies on OCD.[23] Studies were included if they reported proportion of patients with YBOCS <16 at the final follow-up, i.e., the operational definition of remission used in this meta-analysis. The meta-analysis found a reasonably high remission rate of 53% in a pooled sample of 1265 individuals followed-up for a mean duration of about 5 years. The majority of this pooled sample was moderately ill and on outpatient treatment. Over 60%–90% patients were on some form of evidence-based treatment in a majority of the included studies. Higher remission rates were found in this meta-analysis in prospective studies and in studies from India, compared to retrospective studies and those from the west. Inherent characteristics of prospective studies such as higher follow-up rates and more diligent and frequent follow-up may account for the higher rates. Patient characteristics in the Indian studies possibly explain the observed higher remission rate. Patients in the Indian studies were largely self-referred, drug-naive, outpatients, with shorter illness durations, low comorbidity rates, and moderate illness severity. This first meta-analysis of long-term outcome studies in OCD is, therefore, quite optimistic in reporting that over half the patients achieved remission. There are, however, some caveats to this finding. First, symptomatic remission might not always mean functional recovery. Unfortunately, none of the studies included had data on quality of life, disability, or functioning. For a given patient, clinically and functionally meaningful improvement might require a much lower score on YBOCS. Second, important mediating variables such as age at onset of OCD, presence of comorbidities, and insight into illness could not be studied due to the absence of data on these variables in a majority of studies. Third, this meta-analysis included studies conducted at established treatment centers that cater to a treatment-seeking population. Long-term outcome for patients in the general community may differ.

Perhaps, the only community-based follow-up study over 30 years has lent further credibility to the idea of a favorable long-term outcome in OCD.[17] Even with a stringent definition of remission, i.e., a complete absence of OC symptoms over 3 consecutive years, the study found 63.5% people completely remitted at follow-up. Interestingly, the median age at remission lay in the mid-30s for participants recruited at 19–20 years of age. Thus, it took about at least 10 years for the illness to remit. Only about one-third of the patients in this study were professionally treated. If the researchers had used the YBOCS <16 definition for remission and if a larger proportion of patients took professional treatment, the remission rate may have been even higher.

The outcome of OCD is highly variable across individual studies, from as low at 12%[45] to 76%.[30] Illness characteristics are notably different across studies and possibly explain at least some of the variation in remission rates. In general, a short duration illness, of low baseline severity, with low comorbidity rates, drug-naive, and with a robust initial response to treatment shows a higher remission rate.[25] Remission rates also tend to improve with follow-up duration. For instance, the 15-year follow-up from the Harvard–Brown Anxiety Disorders Program found the remission rate to increase from a mere 16% at 1-year to 42% at 11-year follow-up.[60] Other clinical features associated with a better outcome are scattered across studies. These will be further discussed under prognostic factors.

Does quality of life change with treatment in obsessive–compulsive disorder?

Quality of life is impaired in OCD.[66] A reduction in symptom severity may not always better quality of life.[15] The cutoff scores used for symptom remission might not correspond to a better quality of life; or variables other than obsessions and compulsions, such as depression and other comorbid disorders, may influence quality of life. Studies over short follow-up durations have reported improvements in quality of life with treatment.[6768] In one of the few long-term follow-up studies examining change in quality of life with the treatment of OCD, over 12 months, 73 patients were assessed for depression and quality of life.[69] Treatment produced significant changes in quality of life, and depression was a major mediating variable; however, even at follow-up, patients continued to have quality of life scores lower than the population norms. There is a need for more systematic research on quality of life in OCD, including identifying predictor variables other than symptom severity.

What happens to treatment nonresponders in the long term?

There is little literature on long-term outcomes of treatment nonresponders. A clinic-based study followed up 36 individuals, who had not responded to two adequate trials of OCD, for about 4.5 years.[55] Thirty-nine percent of the patients remitted at follow-up. Those who did not remit had not received CBT in the interval period, had poorer quality of life at baseline, and had shorter duration of follow-up and a later age at onset. In another study from the same center, 31 SRI nonresponders were treated with CBT and followed up over 1 year.[59] Seventy-four percent of patients showed a substantial reduction in YBOCS rating, and 48% had remission. Remission from illness was also associated with a reduction in depressive symptoms and improvement in quality of life. It appears from these studies that aggressive treatment over long term has the potential to improve outcomes even for initially treatment nonresponsive OCD.

PROGNOSTIC FACTORS IN OBSESSIVE–COMPULSIVE DISORDER

For a clinician, knowledge of prognostic factors is essential for communicating likely prognosis to a patient. Across OCD studies, there is significant variability in reported predictors. Sociodemographic variables, illness characteristics, and comorbid disorders have all been identified as prognostic factors.[25] Since the recent outcome studies included treated patients, these predictors might best be considered predictors of treatment response, rather than predictors of naturalistic outcome. Predictors might be related inherently to the illness, or may influence treatment seeking through sociocultural mechanisms, or may be associated with both. While several factors have been associated with outcome of OCD in individual studies, perhaps the most consistent of these include – age at onset, baseline illness severity and duration, and early treatment response.

Age at onset seems to have a complex relationship with outcome. Onset in early or middle childhood has been associated with a good outcome[70] and a high rate of spontaneous remission,[71] whereas an adolescent or later onset may lead to a more persistent illness.[6172] Among other illness-related characteristics, a low severity[5560] and a shorter duration of illness[525561] have been repeatedly associated with a better outcome; these observations support the idea of early recognition and intervention for OCD. A robust initial treatment response,[566173] achieving full remission on treatment,[6264] continuation of treatment in the long term,[2374] and combined treatment with SRIs and CBT[50] have been associated with a better prognosis. Discontinuation after prolonged treatment has lower relapse rates[75] than seen in short-term treatment studies.[7677] Therefore, early intensive treatment that is continued in the long term may help achieve a better outcome in OCD. In the meta-analysis of 17 long-term outcome studies,[23] age at onset, gender, duration of illness, and baseline illness severity stood out as predictors of outcome.

Other factors such as symptom dimensions,[304763647278] insight,[79] comorbid axis I disorders such as depression,[306061] developmental disorders such as tic disorders,[80] personality disorders,[8182] schizotypal disorder,[52] and course of illness (continuous versus episodic)[74] have also been found in individual studies as predictors of outcome; however, these associations have been inconsistent and require further inquiry. In patients on CBT, the quality of homework compliance plays a key role in long-term outcome.[59] The family's involvement with and reaction toward a patient with OCD has also been associated with the outcome. Higher accommodation and high expressed emotions worsen the long-term outcome in OCD.[6583] There is, thus, a need for comprehensive assessment at baseline, early identification, and family-focused interventions in OCD.

Neurobiological prognostic factors in obsessive–compulsive disorder

Inconsistency across studies is a limitation of clinical prognostic factors. There have been a handful of studies that reported neurobiological predictors of outcome. Structural neuroimaging data showed anterior cingulate cortex (ACC) volumes to predict short[84] and long-term[85] response to SRIs and cingulotomy.[86] The role of ACC in error monitoring may explain its mechanistic role in OCD. Better neuropsychological performance at baseline on tests for selective attention, verbal memory, and cognitive flexibility was found to predict response to both fluoxetine and CBT over a 12-week trial.[87] A functional neuroimaging study found elevated activity in the right caudate to predict response to a 12-week treatment with paroxetine in OCD; a comparator depression group had lower baseline amygdala and higher prefrontal cortex activity.[88] Findings on neurobiological predictors are currently preliminary at best and need replication. However, they shed light on plausible mechanistic underpinnings of the etiopathogenesis and efficacy of evidence-based treatments in OCD.

CONCLUSIONS AND FUTURE DIRECTIONS

This chapter illustrates the many facets of course and outcome of OCD in adults. Meta-analytic observations from clinic-based studies, as well as findings of long-term community-based follow-ups, suggest that the overall prognosis for OCD might not be bleak and that more than half the patients remit in the long run. Certain predictors of remission have been identified; these highlight the crucial role of early intensive intervention and long-term treatment in improving outcomes. However, the complex nature and heterogeneity in the disorder beget many unanswered questions. The role of comorbid disorders, especially depression; definitions of response, remission, recovery, and relapse; factors associated with quality of life impairment; and the role of family are all significant aspects that need further systematic examination. These inquiries would have high clinical relevance to individuate prognosis.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.