The natural history of lifetime psychiatric disorders in patients with obsessive‐compulsive disorder followed over half a century

Few long‐term studies have examined the life‐time prevalence of comorbid psychiatric conditions in patients with obsessive‐compulsive disorder (OCD). We therefore studied the frequency of comorbid psychiatric disorders, and their relation to onset and prognosis, in patients with OCD who were followed for almost half a century.


| INTRODUCTION
Lifetime history of obsessive-compulsive disorder (OCD) has been suggested to increase lifetime risk for a large number of other psychiatric disorders. 13][4][5][6][7][8] Two exceptions are the Zurich population study with 30 cases followed over 30 years, and an Indian study with 75 patients followed over 11-13-years. 9,103][4][5][6][7][8][9] A recent meta-analysis regarding psychiatric comorbidity in OCD, based on both community and clinical samples, reported that 69% of cases presented comorbidities, with mood disorders (48%) and anxiety disorders (32%) being the most prevalent comorbid conditions. 11Additionally, an international study from OCD research centers using retrospective information reported lifetime comorbidity of major depressive disorder (MDD) in 50%, dysthymia in 23%, social phobia in 26%, generalized anxiety disorder (GAD) in 24%, specific phobia in 26%, and psychotic disorder in 4%. 3 There is a controversy whether comorbid psychiatric conditions should be regarded as primary or secondary to OCD.3][14][15] In addition, it has been suggested that comorbidity could be used for subclassification of OCD. 14 Long-term prospective studies on the implication of psychiatric comorbidity on prognosis of OCD are scarce. 9t has been suggested that number of comorbidities is related to an unfavorable prognosis, 4 while anxiety is related to a better prognosis. 16e followed 144 OCD patients over a mean of 47 years since onset 17 with the same psychiatrist conducting all examinations during a time when, for most part, an effective treatment for OCD was lacking.We have previously reported that half of these patients remitted spontaneously from OCD. 17 The aim of this study was to examine lifetime frequency of comorbid psychiatric conditions in patients with OCD, their time of onset in relation to OCD and their prognostic significance.

Significant outcomes
• All patients had at least one comorbid condition during the course of OCD when followed for almost five decades.• One third had neither OCD nor a comorbid condition at the end of the study, even though no effective treatment for OCD was available at the time of the study.• GAD, psychotic disorder, and substance abuse worsened the prognosis of OCD as did the number of comorbidities.

Limitations
• The study was completed almost three decades ago, and as a result, some findings may not be applicable today.Nevertheless, replicating this naturalistic study today is not possible due to the availability of effective treatment for OCD.• The diagnostic criteria used for OCD are from the 1950s.However, 86% also met the criteria for OCD according to DSM-IV, despite that the questions were not originally designed for that purpose.

| Sample
In 1947-1953, 5732 patients were admitted as inpatients to the Department of Psychiatry, Sahlgrenska Hospital, Gothenburg, and 285 of these had OCD according to case records.Among these, 251 patients (88%), 158 women and 93 men, were examined with a comprehensive psychiatric examination 1954-56. 18his report is based on those who survived at least 30 years after the first examination, and took part in a follow-up examination 1989-1993.At that time, 75 had died, and 32 were lost due to other causes, leaving 122 for re-examination (96 with a comprehensive telephone interview, 23 by personal examination in the subject's home and 3 by letter).For another 22 patients, information was obtained from close informants and case records.The study thus comprises 144 patients (103 women, 41 men; response rate 82% among survivors). 17Mean age was 36.0 years (range 19-52 years) in 1954-1956 and 71.1 years (range 53-87 years) 1989-1993.The mean follow-up time from onset of the OCD to the examination 1989-1993 was 47.0 years (range 37-69 years). 17nformed consent was obtained from all participants and/or their relatives.The study was approved by the Ethics Committee for Medical Research at Göteborg University.

| Examinations
The same psychiatrist (Gunnar Skoog) performed the comprehensive psychiatric examinations in 1954-1956  and in 1989-1993.The semi-structured examination in 1954-1956 was performed in the patient's home or at the outpatient clinic at the Department of Psychiatry, Sahlgrenska Hospital, and included questions about background factors, personality, and psychiatric symptoms. 17he re-examination in 1989-1993 was based on the questionnaire from 1954 to 1956 and focused on the present state and on changes since the first examination. 17Questions included obsessive-compulsive symptoms (including time and type of onset), reported and observed depressed and elevated mood, dysphoria, anxiety, sensitive and paranoid reactions, psychotic symptoms (e.g., hallucinations, delusions, thought disturbances, derealisation, confusion), and reported alcohol and substances abuse.

| Definitions
The diagnosis of OCD in the 1950s was based on Kurt Schneider's criteria, 19 which were widely used at that time.According to these criteria, obsessions and compulsions are "contents of consciousness that manifest themselves with the experience of obsession or compulsion and cannot be forced aside, although in conditions of relative calm they are admitted as absurd or as dominating without ground."These criteria, which were used also in the re-examinations 1989-1993, are broader than those of the DSM-IV 20 (encompassing a more generous inclusion of phobic symptoms).However, a retrospective analysis showed that 124 (86%) of the 144 patients also fulfilled the criteria of DSM-IV for OCD already at baseline. 17Clinical symptoms were diagnosed only when they caused obvious distress and interfered with normal activities.Among the 144 cases of OCD, 69 had recovered (defined as absence of clinically relevant symptoms the last 5 years) at the end of the study, and 75 had remaining clinically relevant OCD. 17 The diagnosis of comorbid mental disorders was based on the DSM-IV criteria. 20Paranoid ideation was classified according to the DSM-IV Glossary of Technical Terms.Beliefs involving persecution, harassment, or unfair treatment that did not reach delusional proportions were classified as paranoid ideation.
Onset was defined as the time at which clinically relevant symptoms of a mental disorder first occurred as determined by self-report/informants and from case records.Information was obtained from the participants' initial period of care as inpatients at the psychiatric hospital, the comprehensive interview 1954-56, medical records between follow-ups and the comprehensive interview 1989-93.Symptoms of comorbid mental disorders occurring before 1954-56 were collected at the first interview and from case records.

| Statistical methods
A two-tailed Fisher's exact test (p < 0.05) was used to compare differences between groups.Two logistic regressions were carried out using the sum of comorbid disorders during lifetime and end of the study in relation to recovery, while controlling for age and sex.
The lifetime frequency of psychotic syndromes was 15.3% (N = 22; 17 schizophrenia-related symptoms, three only delusional syndrome, two psychosis not otherwise specified).An additional six patients had moodcongruent psychotic symptoms only in connection with MDD.

| Onset of comorbid conditions
The mean age of onset for OCD was 24 years.As seen in Table 2, first onset of comorbid psychiatric conditions occurred throughout the course of OCD.Two thirds had comorbid psychiatric conditions prior to OCD onset.Onset >1 year before OCD was found in 59 patients (41.0%), including 50 with specific phobia.After excluding specific phobias, 18 (12.6%)patients had onset of comorbid psychiatric conditions >1 year before onset of OCD, but the proportion with onset during the year before OCD increased to 40.6%, giving at total of 53.2% with onset before OCD.
T A B L E 1 Lifetime prevalence of DSM-IV psychiatric disorders in patients with obsessive-compulsive disorder followed for 47 years.

| Lifetime history of comorbid psychiatric disorders in relation to outcome of OCD
Lifetime history of GAD, psychotic disorder, and substance abuse were related to worse prognosis of OCD (Table 3).Number of lifetime comorbid conditions were associated with less recovery from OCD (OR, 0.80; 95% CI, 0.66-0.98).Among those with >7 comorbid conditions (n = 6), none recovered from OCD.Among those with only one comorbid condition (n = 5), all had recovered from OCD at the end of the study.

| Frequency of comorbid psychiatric conditions at the end of the study
Most OCD patients recovered from their comorbid conditions.The prevalence of comorbid psychiatric conditions at the end of the study was 45.1% (Table 4).The most common were depression (24.3%),GAD (19.4%),MDD (13.2%), specific phobias (6.3%), and paranoid conditions (6.3%).Depression was more common in men than in women at the end of the study (39.0% vs. 18.4%, p < 0.05).Remaining OCD at the last examination was strongly associated with remaining comorbidity (Table 5).At the last examination, 64% of those with remaining OCD had a comorbid condition compared to 25% of those who had recovered.In general, reduction of comorbid conditions during the later course of OCD followed improvement of OCD-symptoms, and vice versa.Number of comorbid disorders at the end of the study were associated with less recovery from OCD (OR, 0.26; 95% CI, 0.14- 0.47).Among those with ≥3 comorbid disorders at the end of the study (n = 10), only one had recovered from OCD.Among those with no remaining comorbid disorder at the end of the study (n = 84) 66.7% had recovered from OCD.

| Main findings
We followed 144 OCD patients for nearly half a century after onset and found that all patients had at least one other mental disorder during their life.The onset of these comorbid conditions occurred throughout the life span, and the course of OCD.More than half of the patients had at least one mental disorder before onset of OCD.However, when each specific mental disorder was considered separately, onset was generally after OCD onset, with the exception of specific phobias.At the end of the study, one third had neither OCD nor another mental disorder.Presence of GAD, psychotic disorder, substance abuse, and number of comorbidities worsened prognosis of OCD.
The finding that all patients had at least one lifetime comorbid psychiatric disorder is in line with previous studies, reporting a prevalence of 90%. 7,8It seems like OCD is involved and integrated with nearly all other psychiatric conditions.These conditions seem more as a part of OCD than independent disorders.In fact, in a longtime perspective, OCD appears as a crystallization of psychiatric nosology in general.Explanations may be that OCD shares biological vulnerability with several other mental disorders, related to for example, shared risk factors, personality traits, or disturbance in the serotonergic system and circuit-based networks (involving e.g., prefrontal cortex, basal ganglia, and thalamus). 21ur frequency of specific disorders was in general higher than in other studies, for example specific phobias (65% The two figures under the two columns all in a row always adds up to the total number of participants in the study (N = 144).
8][9] Psychotic syndromes were found in 15%, similar to Eisen & Rasmussen, 22 but considerably higher than the 2%-8% reported by others. 3,4,7,23The higher frequencies may be due to the long follow-up, the high mean age at the end of the study (71 years), and because episodes occurring before OCD onset were included.Also, retrospective information may underrate lifetime mental disorders, while multiple followup examinations give higher rates. 24On the other hand, our frequency of MDD, 3,4,7-9 alcohol abuse, 3,8,9 substance abuse, 3,7-9 delusional syndrome, 22,25 and schizophreniarelated conditions 2 are within ranges reported by others.Several authors emphasize the overlap between GAD and obsessive anxiety, 26,27 with ruminative worrying and need to keep everything under control as shared features of the two disorders, 26 with corresponding difficulties of setting diagnostic boundaries between worrying in GAD and obsessions in OCD. 28This may explain the disparate prevalence figures in previous studies.OCD and GAD are proposed to be part of the same phenotypic spectrum, as relatives of OCD patients have more often GAD than non-OCD relatives, even in the absence of OCD, and that both disorders have similar age of onset and treatment. 28he high comorbidity figure regarding GAD in our study gives support to this proposition.
We found that 40% of our OCD patients had paranoid conditions, with 11% having delusional disorder and 29% paranoid ideation.The latter characterized by sensitive, near-obsessional experiences of reference.Distinguishing obsessions from delusions and paranoid experiences are challenging, 29 with a continuum from obsessional doubt to delusional convincement.The phenomenon of doubt is on the border between insight and non-insight, and between obsessive superstitions and paranoid suspiciousness, and sometimes exceeds the border in a psychotic direction.There are diagnostic problems in the presence of uncertainty and doubt, when the patient has an openness, even if little, that the belief might be false.The clinician's ability to psychologically understand or not understand the symptoms is then of importance to differentiate the psychotic from the obsessive. 30ew studies have reported on sex differences in the prevalence of comorbid psychiatric disorders in OCD.We found few such differences, as also noted previously. 31owever, there was a substantially larger sex difference in alcohol abuse (39% in men and 3% in women) than previously reported. 32However, during the study period, there was a substantial sex difference in self-reported alcohol consumption in the general population of Sweden 33 and a greater social stigma regarding alcohol abuse among women than men, 34 making comparisons with later studies less relevant.The higher frequency of PAD in women is in agreement with others. 31epressive disorders had variable onsets in relation to OCD, as also reported by others. 8Our finding that 62% of the depressions began after onset of OCD, and nearly a quarter before OCD, suggests that OCD could be related to depression both as a primary and a secondary disorder.It has been proposed that OCD and depression are so integrated that depression within OCD should be classified as part of OCD, 35 while others consider depression as a reaction to OCD. 4,13,28 Alternatively, both disorders may have similar pathogenesis or similar vulnerability.Social phobia occurred after OCD onset in more than three-quarters of the cases, in contrast to the populationbased ECA-Replication Study, where four-fifths had onset before OCD, 8 and the early age of onset in the general population. 36The late onset of social phobia in our sample supports the suggestion that social phobia in OCD differs from that in the general population and might be a reaction to or part of OCD. 28During most part of the last century, specific phobias were considered deeply connected with OCD and even an integrated part of the syndrome. 19,37More than 90% of specific phobias occurred before or early in OCD, as previously reported by others, 8 suggesting that it is part of the etiological pathway, but phobias has an early onset also in the general population. 36GAD had its onset mainly at or after OCD debut, giving no support to the suggestion that OCD is a secondary or adaptive reaction to GAD. 38 In our study, onset of agoraphobia was at or after onset of OCD, in contrast to the variable onset of PAD in relation to OCD.It has been suggested that agoraphobia and GAD, in contrast to PAD, are part of the same spectrum as OCD and that the tendency to worry and ruminate in GAD and OCD leads to agoraphobia. 28Most psychotic symptoms (77%) had onset at or after OCD onset, in line with studies in patients with schizophrenia 27 and with a register study reporting that OCD is a risk factor for schizophrenia. 39hus, OCD may trigger onset of schizophrenia-like symptoms or be a response to early subclinical psychotic symptoms. 15,29,39epression has been related to chronicity, 12 severity 40 and better prognosis of OCD. 41In line with a previous study, 42 we found that comorbid depression did not influence outcome of OCD.However, depression was the most common remaining comorbid condition at the end of the study, affecting almost one fourth of the patients, and connected with less recovery of OCD.GAD was associated with worse prognosis of OCD, contrary to a report that prominent anxiety is related to better prognosis of OCD. 16Very few patients reported alcohol abuse or substance abuse at the end of the study.However, lifetime substance abuse, but not alcohol abuse, was related to less recovery from OCD.This indicates that substance abuse had a negative effect on the prognosis, or that these patients had more severe OCD symptoms at baseline.Psychotic symptoms predicted poorer prognosis of OCD, in line with an earlier report that psychotic features predict poor treatment response and unfavorable outcome. 25ne reason for our finding may be that psychotic symptoms reflect more severe OCD.However, in contrast to the prognosis of primary schizophrenia, the prognosis of psychotic symptoms in our OCD patients was good, and rarely had a chronic course, as also reported previously. 25emaining OCD at the last examination was strongly associated with remaining comorbidity.At the last  examination, 64% of those with remaining OCD had a comorbid condition compared to 25% in those recovered.Thus, in a long-time perspective of almost half a century, 36% of patients with OCD were free from both OCD and other mental disorders.
Among the strengths of this study are the high response rate, and examinations conducted with similar instruments by the same psychiatrist over almost four decades during a time when there were mostly no specific OCD treatment. 17It is thus a naturalistic study, which will not be possible to replicate.Some limitations must be noted, as discussed previously. 17First, the study was finished almost three decades ago and findings may not be applicable to the present situation.However, as mentioned, the study gives a unique possibility to examine the naturalistic course of OCD. 17 Second, the study was based on psychiatric in-patients with OCD, who probably have more severe OCD and more comorbidities, which might have led to hospitalization.Third, in Schneider's criteria for OCD, 19 inclusion criteria in the 1950s, 18 phobias are included in OCD as "obsessively experienced fears." 19Our high frequency of specific phobias may partly be explained by the use of these criteria for inclusion.In addition, in DSM-IV, OCD is defined more strictly and detailed compared to Schneider criteria, which are broader and does not have exclusion and inclusion criteria in the same way as DSM-IV.Thus, DSM-IV might capture a more pure OCD disorder, but might on the other hand exclude too many cases and thus underrate comorbidity.However, 86% of our cases also fulfilled the DSM-IV criteria, 17 despite that the questions used in 1954-56 were not designed to diagnose OCD according to the DSM-IV system.Therefore, patients who appeared in retrospect not to fulfill DSM-IV criteria for OCD may in fact have satisfied these criteria if more DSM-IV specific questions were used.We believe that the broader criteria of Schneider capture the core symptomatology of OCD even better than DSM-IV.Schneider stated regarding the definition of OCD that "only its core can be defined." 19evertheless, this demands more from the interpreter.However, all cases were at inclusion and follow-up comprehensively interviewed by an experienced psychiatrist (G.S.).Fourth, other mental disorders were diagnosed according to DSM-IV, which did not exist when the study started.These diagnoses are therefore not based on a questionnaire constructed to yield DSM-IV diagnoses.Our estimates of specific comorbidities may therefore most likely be underestimations.The reason why we used Schneider's criteria for inclusion, but DSM-IV criteria to diagnose comorbidity, was that we wanted to use newer diagnostic criteria for comorbid conditions, but believed that it was not appropriate to change inclusion criteria retrospectively.Fifth, the study was performed over a long time period.The participants might thus have had difficulties to remember events between examinations, leading to underestimation of comorbid conditions.We tried to alleviate this problem by also including examination of medical records.Sixth, the examination procedures were based on practices of the 1950s, although new and possibly better instruments are now available.However, we chose to follow "the golden rule" for longitudinal studies, i.e. to keep the original schedule.Seventh, we only included persons who survived to the second examinations, which might have influenced the results.If anything, this probably underrated comorbidity, as it is possible that individuals with comorbid conditions would have less survival than those without.
Abbreviation: NOS, not otherwise specified.a A two-tailed Fisher's exact test was used to compare difference in prevalence of comorbid conditions between those with and those without OCD at the end of the study: *p < 0.05; **p < 0.01; ***p < 0.001.
T A B L E 2 Onset of other mental disorders in relation to onset of obsessive-compulsive disorder (OCD).
Note: Any mental disorder when specific phobias is excluded.The figures do not always add up due to overlapping.
T A B L E 5 Prevalence of comorbid conditions at the end of the study in relation to remaining obsessive-compulsive disorder (OCD).