|Home | About | Journals | Submit | Contact Us | Français|
Psychiatric comorbidities are common among patients with schizophrenia. Substance abuse comorbidity predominates. Anxiety and depressive symptoms are also very common throughout the course of illness, with an estimated prevalence of 15% for panic disorder, 29% for posttraumatic stress disorder, and 23% for obsessive-compulsive disorder. It is estimated that comorbid depression occurs in 50% of patients, and perhaps (conservatively) 47% of patients also have a lifetime diagnosis of comorbid substance abuse. This article chronicles these associations, examining whether these comorbidities are “more than chance” and might represent (distinct) phenotypes of schizophrenia. Among the anxiety disorders, the evidence at present is most abundant for an association with obsessive-compulsive disorder. Additional studies in newly diagnosed antipsychotic-naive patients and their first-degree relatives and searches for genetic and environmental risk factors are needed to replicate preliminary findings and further investigate these associations.
The clinical heterogeneity of schizophrenia is indisputable. Virtually no 2 patients present with the same constellation of symptoms. Moreover, even in the same patient, symptoms can show dramatic change over time, and there is significant interplay between different sets of symptoms: eg, “secondary” negative symptoms might be ameliorated with resolution of positive symptoms, while core “deficit” negative symptoms are more enduring but can worsen over the longitudinal course of illness. Such observations give way to considerations that these may even constitute groups of diseases of generally common phenotypic expression but of different underlying etiopathology.1
Further complicating the clinical picture of schizophrenia as well as understanding the boundaries and etiology of this condition is the substantial psychiatric comorbidity.2 Depression, anxiety, and substance abuse are common accompaniments of the schizophrenia condition, and they in turn perturb the clinical picture.3 For example, depression can cause secondary negative symptoms, panic attacks can drive paranoia, and cannabis abuse can worsen positive and disorganization symptoms. Conversely, depressive symptoms seen in the context of a florid psychotic relapse often resolve with treatment of the positive symptoms but may remerge in the “postpsychotic” state and in turn worsen the longitudinal course of the illness.4,5
Nosologists have great difficulty dealing with complex sets of symptoms.3,6,7 Generally, an implicit or explicit hierarchy is embraced, such that schizophrenia “trumps,” depression, and anxiety. Or, if no primacy can be determined, resort is made to labels such as “schizoaffective disorder” or even “schizoobsessive” subtype of schizophrenia.3,8 An alternative approach, reified in Diagnostic and Statistical Manual of Mental Disorders, is to consider these symptoms as part of another axis I diagnosis that is occurring alongside schizophrenia.9 Under this scenario, the patient has 2 major conditions, and these have co-occurred (perhaps for some etiological reason common to both disorders). This is very much the model considered—and clinically endorsed—when a patient with schizophrenia also has an alcohol dependence or drug addiction problem.10 Additionally, recent work on the potential biological vulnerability to cannabis abuse that might explain some variance in the risk of later developing schizophrenia raises again the proposition that the clinical associations that we commonly observe in schizophrenia may also have biological and potentially etiopathological significance. 11
Bermanzohn et al12 provocatively proposed that we “stake out the midground”; they suggest that psychiatric comorbidities are so common that they might be integral to schizophrenia. To a large extent, our current research in clinical trials and neurobiological studies is increasingly coming in line with this proposition because now such studies support broad inclusion criteria of “all comers” … the schizophrenia patients whom we see in everyday clinical practice, who have prominent anxiety symptoms, or may also have depressive symptoms, and also abuse drugs and alcohol.
The purpose of this article is to “take stock” of these (anxiety, depression, and substance abuse) comorbidities and their relationship to schizophrenia. Reviewing the relevant epidemiological, genetic/familial, neurobiological, and therapeutic literature, we ask whether comorbidities should be considered:
There is an increased prevalence of anxiety disorders among patients with schizophrenia compared with the general population.13 These include panic disorder, posttraumatic stress disorder (PTSD), obsessive-compulsive disorder (OCD), generalized anxiety disorder, and social anxiety disorder. Prevalence estimates are influenced by heterogeneity among definitions of symptoms and rating instruments used for diagnosis. Other diagnostic issues complicating the study of anxiety in schizophrenia are that symptoms may occur spontaneously, intermittently, in direct response to psychotic symptoms, and/or as a side effect of antipsychotic medications. Here, we will focus on panic disorder, PTSD, and OCD, the 3 anxiety disorders that have been most extensively studied in patients with schizophrenia.
Two studies from the National Institute of Mental Health Epidemiologic Catchment Area survey found a 28%–63% (across the 5 study sites)14 and 45%15 prevalence, respectively, of panic attacks in patients with schizophrenia. Two other studies from the ECA survey16,17 found a lifetime odds of ≥35 for having panic disorder in subjects with (compared with those without) a diagnosis of schizophrenia. In total, 27 published studies have investigated the epidemiology of panic symptoms among patients with schizophrenia.12,14–39 (see Table 1) The prevalence of panic attacks (7.1%–63%) and panic disorder (3.3%–29.5%) vary widely across these studies. A weighted average of the available data from these (heterogeneous) studies crudely estimates a 25% prevalence of panic attacks and a 15% prevalence of panic disorder in patients with schizophrenia (data not shown). By comparison, the lifetime prevalence of panic disorder in the US general population ranges from 2.0%–5.1%.17,40,41
The majority of studies of panic attacks/disorder focused on patients with chronic schizophrenia. Two studies, however, investigated the prevalence of panic symptoms in patients with first-episode psychosis (FEP), which reduces confounding by medications and other factors. Strakowski et al35 found that of 102 consecutive patients hospitalized FEP patients, 6% met Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised, criteria for panic disorder. The authors also found poorer initial outcomes in patients with comorbid panic disorder. In another study, Craig et al24 examined panic symptoms in 225 patients with first-admission psychosis (schizophrenia or schizoaffective disorder) participating in the Suffolk County (NY) Mental Health Project. Panic symptoms were present at baseline in 11.2% of patients. Furthermore, patients with baseline panic symptoms were significantly more likely to exhibit positive symptoms of psychosis after 24 months.
Evidence from 3 studies18,19,31 suggests that panic symptoms may be more common in patients with paranoid schizophrenia, compared with other schizophrenia subtypes or schizoaffective disorder. Furthermore, panic attacks were frequently related to paranoid ideations in these studies. Bermanzohn et al43 described the potential for a relationship between panic attacks to paranoia. Comorbid panic symptoms may be associated with more severe psychopathology,22,23,30,36,37 as well as increased risks of suicidal ideation and behavior,26,45,46 and may also increase vulnerability to comorbid substance use.28
In a contrasting approach to the epidemiology of this association, another study from the ECA survey by Tien and Eaton42 found that the presence of panic attacks was associated with a 2.28-fold increased risk of developing schizophrenia. The temporality of this association suggests that, for some patients, the presence of panic attacks may be part of the psychosis prodrome.
Weissman et al47 reported on a potential syndrome with genetic linkage to chromosome 13q32 (marker D13S779) in 34 families segregating for panic disorder. In addition to panic disorder, these patients also had an excess of urologic problems, headaches, thyroid problems, and/or mitral valve prolapse. This region of chromosome 13 (13q32–34) encodes the G72/G30 gene complex (G72 is also known as D-amino acid oxidase activator or DAOA), which has been linked to schizophrenia in multiple studies, as reviewed in a recent meta-analysis.48 Additional studies are needed to investigate potential shared genetic risk factors for schizophrenia and panic disorder.
Family and twin studies support the biological plausibility of an association between schizophrenia and panic disorder. Heun and Maier29 reported on the only family study of panic disorder in schizophrenia. They assessed a total of 1068 first-degree relatives of 59 patients with schizophrenia, 54 patients with panic disorder, 29 patients with panic disorder and schizophrenia, and 109 controls. They found a significantly increased prevalence of primary panic disorders among relatives of patients with schizophrenia (4.3%) compared with controls (0.9%) but not compared with subjects with panic disorder. Furthermore, there was not an increased risk of schizophrenia among relatives of patients with either panic disorder (0%) or control subjects (0.3%). In a sample of N = 24 twin pairs discordant for schizophrenia and N = 3327 twins without schizophrenia from the Vietnam Era Twin Registry, Lyons et al32 found that the nonaffected co-twins of schizophrenia probands had a 3-fold increased odds of panic disorder compared with control twins. This finding did not reach statistical significance, but the results were limited by the small sample size of the study.
One large family study found an increased prevalence of panic disorder in first-degree relatives of patients with schizophrenia. However, the present evidence supporting the hypothesis that panic disorder is part of the syndrome of schizophrenia is limited, subject to confounding, and findings are in need of replication. Additional studies in drug-naive patients and their first-degree relatives, controlling for potential confounding factors are also needed to further investigate this association. Twin studies might also contribute to our knowledge base in this regard.
Trauma histories are common in patients with schizophrenia, and childhood trauma is a risk factor for psychosis.49 Patients with schizophrenia may be at increased risk for exposure to trauma, due to illness-related features, environmental influences, and/or comorbid substance use. Many factors complicate the diagnosis and investigation of co-occurring PTSD and schizophrenia, including the presence of psychotic symptoms within the context of PTSD, or PTSD symptoms—such as reexperiencing the trauma—that may mimic psychotic symptoms. Furthermore, psychotic symptoms (eg, hallucinations and delusions) or experiences (eg, involuntary hospitalization, seclusion, restraint, forced medications) may themselves be a traumatic event contributing to PTSD,50–54 though they have not been uniformly considered as a potential precipitating stressor.
A total of 20 published studies have reported on the epidemiology of PTSD in schizophrenia. 32,34,36,39,51–66 (see Table 2) These samples, including those which considered psychosis-related symptoms or experiences, found a prevalence of PTSD among patients with psychosis 0%–67%. A weighted average of the available data from these (heterogeneous) studies crudely estimates a 29% prevalence of PTSD in patients with schizophrenia, compared with a 7.8% estimated lifetime prevalence of PTSD in the US general population.67
While the majority of these studies focused on patients with chronic schizophrenia, Strakowski et al66 found that 4 of 18 (22%) patients with a schizophrenia-spectrum disorder met criteria for PTSD antecedent to their first psychotic episode. The diagnosis of PTSD predated the onset of the psychotic disorder by more than 1 year in 2 of these 4 patients. In a cohort of 170 patients with a FEP, Neria et al62 found a 10% prevalence of PTSD.
The presence of PTSD has also been shown to be associated with more severe psychopathology (including cognitive impairments)56,58,70,71, higher rates of suicidal ideation and suicidal behaviors,68 and more frequent outpatient physical health visits and hospitalizations69 in patients with schizophrenia.
Although a majority of studies found an increased prevalence of PTSD in excess of that in the general population, including inpatients with both FEP and chronic schizophrenia, there is little other evidence to support the hypothesis that PTSD is part of the illness of schizophrenia. The increased prevalence may be largely accounted for by environmental factors, particularly increased rates of exposure to childhood trauma or as the direct result of psychosis-related trauma. Factors, both clinical and neurobiological, that confer increased vulnerability to PTSD in patients with schizophrenia have been largely unexplored. We are aware of no published genetic or family studies of patients with schizophrenia and PTSD. In the twin study of Lyons et al,32 nonaffected monozygotic co-twins of schizophrenia probands had a nonsignificant increased odds of PTSD compared with control twins.
Obsessive-compulsive symptoms (OCS) and OCD have been frequently studied in patients with schizophrenia (see Table 3), with the majority showing an increased rate of both OCS and OCD in schizophrenia. A study from the ECA survey14 found a 12.5-fold increased odds of having OCD given a diagnosis of schizophrenia. By contrast, another study from this survey42 found a 3.77-fold increased risk of schizophrenia among patients with OCD, suggesting that, for some patients, the presence of OCD may be part of the psychosis prodrome. A total of 36 studies have investigated the epidemiology of OCS/OCD among patients with schizophrenia.12,15,16,19–21,23–25,28,34–36,38–40,65,66,72–87 (see Table 3) The prevalence of OCS (10%–64%) and OCD (0%–31.7%) vary widely across these studies and may have been overestimated due to difficulties in distinguishing clinically obsessions and delusions. It may be very difficult to determine whether the patient is experiencing an obsession or a delusion, especially when an obsession is held with firm conviction. A weighted average of the available data from these (heterogeneous) studies crudely estimates a 25% prevalence of OCS and a 23% prevalence of OCD in patients with schizophrenia. OCS are present throughout the course of schizophrenia. Several studies have suggested that OCS manifest as part of the psychosis prodrome. 88–91 Two studies72,91 found that the presence of OCS was associated wither earlier age of onset of psychosis. Additionally, 3 studies35,66,84 have found an 11.0%–15.2% prevalence of OCD in patients with FEP.
Several studies have directly compared clinical features of schizophrenia with or without comorbid OCD. In a study of 22 adolescents with schizophrenia and OCD matched for age, gender, and number of hospitalizations with 22 adolescents with non-OCD schizophrenia, Poyurovsky et al91 found that patients with schizophrenia and OCD had earlier age of onset of illness and more OCD-spectrum disorders, including primary tic disorders. There was no difference in the severity of schizophrenia symptoms based on the presence or absence of OCD. In a majority of patients in this study, OCS either preceded or co-occurred with the onset of schizophrenia. Poyurovsky et al92 also compared 100 patients with schizophrenia and OCD, 100 patients with schizophrenia but no OCD, and 35 patients with OCD alone. They found an increased prevalence of OCD-spectrum disorders, including body dysmorphic disorder and tic disorder, among patients with schizophrenia and OCD vs non-OCD schizophrenia. There were no differences in affective, anxiety, and substance use disorders between these groups. The prevalence of OCD-spectrum disorders was similar between patients with schizophrenia and OCD and those with OCD. In a study of schizophrenia with and without OCD (n = 50 in each group), Rajkumar et al93 found that comorbid OCD was associated with greater paranoia and first-rank psychotic symptoms, less anergia, increased rates of depression and comorbid personality disorder, and perhaps less disability. Several studies have found more severe neuropsychological impairments in patients with schizophrenia and OCD.94,95 In comparison to patients with non-OCD schizophrenia and OCD, Whitney et al95 found that patients with schizophrenia and OCD had greater impairments across several domains of executive function.
The presence of OCD also has prognostic significance in patients with schizophrenia. Parenthetically, obsessive symptoms—as was the case with affective symptoms—were originally considered to be of favorable prognostic value. This does not appear to be the case. Braga et al39 found that patients with schizophrenia and comorbid OCD had greater disability as measured by Sheehan disability scale global scale, work subscale, and social life subscale scores. In a study of N = 102 patients with chronic schizophrenia, Berman et al72 found that the presence of OCS was associated with earlier age of illness onset, increased rates of hospitalization in the previous 5 years, and a decreased likelihood of being employed or married. Sevincok et al96 found that OCD was an independent risk factor for suicidal ideation and suicide attempts in patients with schizophrenia. OCD was also more prevalent among patients with (than those without) suicidal ideation. The authors also found that compulsions were a predictor of suicide attempts.
Poyurovsky et al.97 completed an important family study of schizophrenia probands with (n = 57) and without (n = 60) OCD and 50 controls. Relatives of probands with combined schizophrenia and OCD (n = 182) had significantly increased rates of schizophrenia and OCD, as well as obsessive-compulsive personality disorder, than probands with non-OCD schizophrenia (n = 210). There was also a trend for an increased risk of OCD in relatives of probands with schizophrenia and OCD compared wiith non-OCD schizophrenia. Furthermore, relatives of probands with schizophrenia did not differ in risk of schizophrenia-spectrum, mood, or substance use disorders, based on the presence or absence of OCD. The authors argued that the differential aggregation of obsessive-compulsive-spectrum disorders in first-degree relatives supports the validity of a putative “schizoobsessive” schizophrenia subtype. In the only published genetic study of patients with schizophrenia and OCD, Poyurovsky et al98 completed a case-control study of the COMT Val158 Met polymorphism. They found no differences in COMT allele and genotype distribution between patients with schizophrenia and OCD (n = 113), OCD (n = 79), and controls (n = 171).
Dopamine and serotonin are key neurotransmitters involved in the pathophysiology of both schizophrenia and OCD. While substantial overlap in neurobiology87,99 may well contribute to the association between these disorders, a potential confounding factor in epidemiological studies of this association is that second-generation antipsychotics (SGAs) with serotonergic 5HT2 receptor blockade may exacerbate or produce de novo OCS in patients with schizophrenia.100–103 By contrast, SGAs have also been effective as an adjunctive medication for treatment-refractory OCD.104–107 Additional studies are needed to this investigate the neurobiology of this paradox, which has important treatment-related implications.
Taken together, several lines of evidence suggest that patients with both schizophrenia and OCD may represent a “schizoobsessive” subtype of schizophrenia, with differences in psychopathology, course of illness, and response to treatment, as opposed to comorbid syndromes. This literature supports the possibility that, in a subgroup of patients, OCD may be part of the illness of schizophrenia. Further studies in large samples from diverse population, including in newly diagnosed antipsychotic-naive patients, designed to control for potential confounding factors, are needed to replicate preliminary findings. The search for shared environmental risk factors, such as low birth weight or prenatal stress, for schizophrenia and OCD represents another area for future investigation. The presence of such factors would not provide direct evidence for the relationship between these 2 conditions but would support the biological plausibility of their association. Additional family and genetic studies are needed to determine if schizophrenia and OCD constitute part of the heritable schizophrenia spectrum are accounted for by shared environmental risk factors or both.
The relationship between psychotic and affective symptoms has been central to the dilemma of psychiatric classification. Indeed, substantial evidence (not reviewed here) show that schizophrenia and bipolar disorder, in particular, may be distributed across a dimensional spectrum (or more apt, across multidimensional spectra).108–112 Furthermore, there has been an ongoing and robust debate about the nosological status of “schizoaffective” disorder,7,8,112 with varying definitions and approaches that make that literature very difficult to negotiate. These 2 aspects go way beyond the scope of this review yet are important aspects of nosology that are of relevance to the topic of psychiatric comorbidity. Here, we confine ourselves to reviewing studies of the co-occurrence of the symptoms of psychosis and unipolar depression, a phenomenon seen at some point in illness in the majority of schizophrenia sufferers, as well as in a substantial number of primary depressive patients. In this context, Möller poses the question: “whether these depressive symptoms are part of the rich psychopathological picture of schizophrenia, which, beside the core paranoid-hallucinatory syndrome, includes a negative syndrome, a cognitive syndrome and also a depressive syndrome, or whether depression and schizophrenia should be seen as separate conditions in terms of the concept of comorbidity.”113
Bartels and Drake114 suggested that depressive symptoms in schizophrenia be divided into 3 subtypes, including (1) depressive symptoms secondary to organic factors, (2) “nonorganic” depression intrinsic to the acute psychotic episode, and (3) depressive symptoms that are not temporally associated with the acute psychotic episode, such as symptoms associated with the prodrome, the postpsychotic interval, as well as those symptoms that resemble depression that may represent negative symptoms of schizophrenia. Such approaches offer a structure for considering the various relations of depressive symptoms in patients with schizophrenia and are addressed below.
Antipsychotic medications themselves produce neurological side effects like Parkinsonism (particularly bradykinesia, diminution of affective expression, masked facies, and verbal delays) and akathitic restlessness that may be confused with the psychomotor retardation or agitation of depression. Antipsychotic drugs may also produce a primary dysphoria, possibly due to dopamine blockade in reward pathways, and it has even been suggested that these drugs are innately depressogenic. People with schizophrenia are also prone to general medical morbidities115 and substance use disorders,116 some of which may also produce depressive symptoms. Certain negative symptoms, such as anhedonia, abulia, alogia, amotivational and avolitional states, and social withdrawal, can overlap with or spuriously suggest depression.117 Demoralization,118 disappointment, or loneliness119 following a psychotic episode may create lingering feelings of dysphoria.
The classic construct of depression in schizophrenia is that of postpsychotic depression (PPD), defined in an appendix of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision,8 as a major depressive episode that is superimposed on, and occurs only during the residual phase of schizophrenia. PDD has traditionally been formulated as a psychological reaction to loss or to the psychological trauma of the psychotic episode. Roth wrote that the “depressive reaction that follows (a psychotic episode) is part and parcel of a total psychobiologic reaction to a failure of the patient in some area of human relationships.”120 The relationship of PPD to the psychotic episode itself remains unclear, including the question of whether the depression is a reaction to psychosis, or represents an unmasking effect of the depression as the psychosis remits.121 This latter view is supported by observations that depressive symptoms are often associated with positive symptom scores122 and decrease with effective neuroleptic treatment. There is a long-standing literature that depression is a common symptom found during psychotic decompensation.123
In contrast to early views that depression was associated with favorable prognosis in schizophrenia, the evidence speaks otherwise. Mandel et al124 followed 211 schizophrenia patients in the community for a year after hospital discharge. The 25% of patients who suffered depression in the first few months after discharge had a notably greater burden of symptom chronicity. Johnson125 reported that chronic patients who developed depression more than a year after acute recovery experienced more relapses than other patients. An older study by Tsuang and Coryell126 and a more recent one by Sim et al127 both failed to reveal better outcomes in schizoaffective disorder than schizophrenia.
Patients with schizophrenia are at increased risk of developing depression relative to the already high lifetime prevalence of depression in the general population. Many investigators have reported rates of depressive psychopathology in psychotic patients,128–166 a summary of which is shown in table 4. As might be expected, the measured rates of depressive experience varied widely. There is methodological diversity in this literature due to varying definitions for schizophrenia (or psychotic illness), heterogenous study populations, and varying time intervals over which depressive occurrence was considered, ranging from point prevalence to many years. Nonetheless, as was noted in the outstanding review by Siris and Bench,4 the above-cited studies have convincingly indicated that patients with schizophrenia were prone to elevated rates of depression, with a modal frequency of about 25%.
It is instructive to examine the likelihood that major depressive episodes will evolve into psychosis, and indeed, depressed patients are at high risk for developing psychotic symptoms during the course of affective illness. As a whole, however, this has been less thoroughly studied than the likelihood of depression in schizophrenic patients. Ohayon and Schatzberg167studied the point prevalence of depression in a general population sample of 18,980 people surveyed in a multinational European study. About 16.5% of all subjects endorsed at least one key depressive criterion, and of those, 12.5% reported delusions and/or hallucinations. Of the 454 subjects diagnosed with a full Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, major depressive episode, 18.6% experienced delusions and/or hallucinations. This is consistent with findings that 14% of the ECA community sample diagnosed with major depression168 and 16.9% of an independent first-admission major depression sample169 experienced psychosis during the course of the depressive episode.
Studies of individuals at high risk and ultrahigh risk for developing schizophrenia have generally demonstrated a significant degree of depressive symptoms prior to and during the emergence of psychotic symptoms.170–172 Cornblatt et al173 identified affective disturbances and social isolation as part of an “underlying vulnerability core” in a group of 62 adolescents in various stages of emerging psychosis. Hafner et al169 obtained comprehensive histories from 232 mostly previously untreated, first-admission adult, and teenaged patients diagnosed with schizophrenia, as well as 130 healthy controls and 130 demographically matched, first-admission patients with a diagnosis of depression. In comparing a subset of 130 patients with schizophrenia with the primary depression group, the authors generated lists of the 10 most frequent initial symptoms of illness for depressed and schizophrenia groups, resulting in a combined and highly overlapping list consisting of 13 symptoms. Eight of the 13 most common symptoms did not differ significantly in the frequency. Both diagnostic groups had suffered from both a depressive core syndrome and negative syndrome (difficulties in thinking and concentration, loss of energy, social withdrawal) in the early course of illness, with closely paralleling courses. Patients with schizophrenia were most distinguished, not surprisingly, by markedly escalating positive symptoms leading up to the index admission. The authors concluded that that the initial symptoms of illness reflected a core psychopathology common to the very early stages of both illnesses. They also noted that the peak of depressive experience in patients with schizophrenia coincided with peak psychosis.
Concerning the neurobiology of schizophrenia and depression and evidence of etiologic and pathophysiologic coincidence or overlap, the majority of studies that might inform the topic utilize subjects with either depression or schizophrenia but rarely both. While functional imaging studies of depressed patients have shown decreased prefrontal metabolism or decreased regional cerebral blood flow (rCBF),174,175 those that directly compared rCBF changes during working memory tasks in schizophrenia and depressed groups showed substantial intergroup differences in the direction of greater reduction of rCBF in schizophrenia than depressed subjects.176,177 Reduced hippocampal volume has been reported in depression and schizophrenia,178–180 and a recent study examined the association between myelin related genes and the clinical characteristics of 280 schizophrenia subjects—specifically the presence or absence of depressive comorbidity.181 The investigators demonstrated an association between the glycoprotein M6A gene (GPM6A)—a modulator of the influence of stress on the hippocampus in animals—with the subgroup of schizophrenia patients who showed the highest degree of depression.
In contrast, one of the most consistent functional imaging findings in depression concerns increased rCBF or glucose metabolism in the amygdala of ill subjects relative to healthy controls.182 This finding is absent in patients with schizophrenia who, if anything, show diminished amygdalar activity relative to controls.183 What is missing in the literature is direct comparison of neurobiological variables between patients with schizophrenia with and without comorbid depression.
Much has been written about the newer generation antipsychotic drugs and the potential advantage in psychotic patients with depression, including avoidance of dopamine blockade dysphoria and extrapyramidal side effects (EPS). The newer drugs’ unique pharmacologic features, particular affinity at various serotonergic receptors, may well confer some direct and indirect advantages.184
Tollefson et al185 raised the possibility that SGAs—in this instance olanzapine—may have a direct effect on a depressive symptom domain in patients with schizophrenia. In a path analysis of depression in schizophrenia and its treatment with either olanzapine or haloperidol, they reported superiority for olanzapine and that 56% of this effect on depressive symptoms was on “primary” symptoms rather than secondary to negative symptoms, relief of EPS, etc. Similar analyses have been conducted for other SGAs.186,187 Moreover, clozapine's antisuicide effect did not appear to be related to (merely) better amelioration of symptoms because both clozapine and olanzapine fared equally well on positive, negative, and depressive symptom improvements.188 Moreover, the now widespread use of SGAs in bipolar disorder is additional indirect evidence for some independence of effects on mood and not simply an “antipsychotic” effect (to reduce positive symptoms, thereby lessening depression) in schizophrenia.
From a different vantage point, the role of antidepressant therapy in schizophrenia has received relatively little attention, particularly given the frequency of depressive symptoms and the regular copharmacy (in approximately 30% of patients) of antidepressants and antipsychotics when treating patients with schizophrenia. Interestingly, Siris et al189 showed that adjunctive imipramine improved depression and also resulted in fewer psychotic relapses. The information concerning adjunctive antidepressant therapy with SGAs is particularly scant. Whether, in view of their effect on neuroplasticity, these drugs might have broader clinical effects beyond treating comorbid depressive symptoms in schizophrenia is of interest. Cornblatt et al190 found that prepsychotic adolescents who received antidepressants did just as well as prepsychotic patients who were treated with antipsychotics. Although a naturalistic study, the potential that antidepressants may impact the development of psychosis is intriguing and provocative.
In concluding this section, the following observations can be made: (1) depressive symptoms are common in patients with schizophrenia; (2) they add further to the disability of schizophrenia, including being associated with a heightened risk for psychotic relapses; (3) PPD may be a particular “forme fruste” of major depression in schizophrenia; (4) there is some evidence, far from conclusive, that medications might directly impact depressive, mood, and suicidality to some extent that is not simply “less depression because of less psychosis”; and (5) although intuitively appealing, there is insufficient evidence in the literature (including a dearth of neurobiological studies) to support the proposition that this represents a distinct subgroup of schizophrenia.
The abuse of alcohol and/or illicit drugs by patients with schizophrenia is a remarkably common phenomenon …“the rule rather than the exception.10,191 In the ECA study, it was estimated that 47% of patients with schizophrenia also had a lifetime diagnosis of substance abuse disorder.191 This is consistent with findings from a variety of other epidemiological and clinical studies, both in the United States and worldwide.10,191–194 In general terms, substance abuse comorbidity is associated with a variety of negative consequences for the course of schizophrenia (see table 5), with medication nonadherence often appearing as a “final common pathway” for these effects. Description of the epidemiology and consequences of substance abuse and schizophrenia is beyond the scope of this article and is accounted well elsewhere.8 Suffice it to say that substance abuse comorbidity is common and is deleterious to the course and outcome of schizophrenia. The investigation of this co-occurrence has been (perhaps to a greater extent than the other comorbidities) hampered by a general effort to exclude patients with comorbid substance abuse when studying schizophrenia. Thus, sampling bias is an important consideration in evaluating this association.
Explanations for the common association of substance abuse with schizophrenia are highlighted. Firstly, one might consider this merely a chance co-occurrence, particularly because adolescents and young adults abuse drugs. Why should adolescents/young adults with schizophrenia be any less likely to do this? Epidemiological data and clinical experience suggest that this is not merely a “chance occurrence” and that it is frequent and beyond mere coincidence.191–193 Furthermore, it seems that both patterns of use and motivations for use are very similar in individuals with schizophrenia, as in the nonschizophrenia population.
The second notion is that alcohol or drugs actually caused schizophrenia and that this explains the co-occurrence. Heavy and protracted abuse of alcohol has been causally associated with a discrete alcoholic hallucinosis, but this is relatively rare and the longitudinal course is usually not the same as schizophrenia, such that, eg, affect and personality are relatively preserved. Any strong causal association between alcohol and schizophrenia per se is difficult to envisage given the ubiquity of alcohol use and the fact that the vast majority of people who do use alcohol to excess do not develop schizophrenia. The evidence for illicit drugs being causally associated with schizophrenia is at best mixed but has been most compelling argued for cannabis. A number of cohort studies have now established a temporal relationship between cannabis exposure in adolescence or early adulthood, and later schizophrenia, with an odds ratio of just over 2.0 (2.09, with confidence intervals of 1.54–2.84) in a recent metaanalysis of these studies.195 Additionally, some studies show a “dose-dependent” effect, such the more cannabis consumed the greater the likelihood of schizophrenia.196 However, again the prevalence of schizophrenia is disproportionate to the ubiquitous smoking of cannabis, there is no clear association between rates of schizophrenia and rates of cannabis use in any given population, and most people who imbibe cannabis do not develop schizophrenia. Thus, it seems that cannabis can be conceptualized as a cumulative causal factor in some individuals, acting in concert with other vulnerability factors to promote the manifestation of the illness in some individuals who might otherwise have remained schizophrenia free. The effect is small, with a population attributable fraction of 5%–7%. Also, it does not appear that patients with schizophrenia and comorbid cannabis have any higher genetic loading for schizophrenia than patients with schizophrenia alone.197
Caton et al198 examined the relationship between substance-induced psychosis and schizophrenia by longitudinally evaluating patients who presented acutely psychotic, all of whom had abused drugs or alcohol prior to this first ever presentation with psychosis. Forty-four percent of patients turned out over time to have had a drug-induced psychosis, while 56% of patients ultimately had schizophrenia as their primary diagnosis. Patients with a drug-related psychosis had marginally less positive and negative symptoms at initial presentation, they were more likely to have visual hallucinations, and their parents had a history of substance abuse. Caspi et al11 examined this issue from a different, complementary perspective. As part of a large epidemiological study of schizophrenia in New Zealand, they found that those adolescents who possessed the “faulty” allele (val 158 met) polymorphism of the COMT (cathechol -O-methyl-transferase) gene were the people who had the vulnerability to cannabis abuse. This might help explain this association, which appears robust from epidemiological data but is still a small effect. There is a recent study of brain imaging in nonpsychotic cannabis abusers that shows progressive brain changes with heavy and chronic cannabis abuse.199 The authors report some association between paranoid experiences in a subset of these patients and greater prominence of hippocampal changes. As a general observation, there have been few biological studies of this dual diagnosis patient population because substance abuse is more often than not an exclusionary criterion. On the other hand, there is a growing appreciation of potentially shared neurochemical vulnerability between substance abuse and schizophrenia.200 Animal neurochemical and now human brain imaging studies point to the role of dopamine in the amygdala as being key to understanding drug craving and reward behaviors. In schizophrenia, pleasure and reward are blunted as part of negative symptoms. It is plausible that dopamine dysregulation might predispose patients with schizophrenia to abuse drugs.194,200 It has also been explained that patients with schizophrenia who abuse drugs may actually have milder symptoms and that their poorer course is more attributable to the direct effect of drugs on worsening symptoms as well as the associated medication nonadherence. This is certainly intuitive in the sense that patients with more severe illness are less likely to have the opportunity and social context to acquire street drugs. It has also long been suggested that patients self medicate either to reduce their symptoms or to counteract the effects of antipsychotic medications.201,202 Either association is plausible and in accord with clinical experience. However, the rate of substance abuse comorbidity has not seemed to diminish in an era of treatment with SGAs that have less motor and secondary negative symptom effects.203 Regarding treatment of patients with substance abuse, these patients show similar responses to antipsychotic medications as nonabusing patients with schizophrenia—once they take their medication, a major challenge in this patient group.20,204 In the clinical antipsychotic trials of intervention effectiveness (CATIE) study, patients with comorbid substance abuse showed comparable responses with each SGA than patients without substance abuse.205 There is some evidence that dual diagnosis patients might do better on clozapine, with less relapse into abuse of drugs or alcohol.206
Overall, while substance abuse comorbidity is remarkably common in schizophrenia, the evidence is lacking that this represents some distinct subgroup of etiopathological significance. While explanations toward a shared neurochemical, dopamine-mediated vulnerability to both schizophrenia and substance abuse are intuitively appealing, at present the evidence base is scant.194 Moreover, the prevailing view in both the addiction field and in schizophrenia research is that this represents a co-occurrence of 2 conditions rather than some etiologically distinct subgroup of schizophrenia patients who are characterized by a proclivity to substance abuse.
There is clearly an increased prevalence of anxiety, depressive, and substance abuse disorders in patients with schizophrenia that occurs in excess of that in the general population. These comorbidities occur at all phases of the course of illness, including in the psychosis prodrome, FEP, and chronic schizophrenia. A limited body of evidence supports the plausibility of the hypothesis that anxiety disorders are part of the illness of schizophrenia, with the strongest evidence being for OCD. PTSD, and other anxiety symptoms, while common, do not appear to be etiologically linked to schizophrenia. Depressive symptoms are also intrinsic to the illness and import a poorer outcome, including more psychotic relapses. Understanding this relationship is important and is also additionally complicated by broader perspectives about the boundaries/overlap between psychosis and mood disorders. Substance abuse is particularly common and also worsens the course of illness, although here this effect is inextricably linked to treatment noncompliance. For each of these comorbidities, their presence is generally associated with more severe psychopathology and with poorer outcomes. What is conspicuous from this review is the relative lack of investigation toward a neurobiological basis of comorbidity among patients with schizophrenia. This is striking in view of how common and challenging these comorbidities are. There is a conspicuous absence of any “smoking gun” findings for etiological heterogeneity here. While there has been at least some headway in treatment studies of both pharmacology and nonpharmacology, it is rudimentary and in relation to OCD and schizophrenia there is the suggestion that antipsychotic medications might even aggravate these symptoms.3 There is also, on the other hand, evidence that antidepressants can not just improve depressive symptoms but perhaps also impact favorably negative and general psychopathology as well.5 These observations may contribute in part to the high rates of polypharmacy that are observed in the treatment of schizophrenia.207 At present, the therapeutic implications of this clinical heterogeneity are poorly understood and are largely manifested in “trial and error” treatment choices. The most parsimonious conclusion at the present time is that these comorbidities are certainly more common than chance in schizophrenia, but their etiopathological significance and treatment implications thereupon are poorly understood at the present time.