Comparison of Putative Intermediate Phenotypes in Schizophrenia Patients With and Without Obsessive-Compulsive Disorder: Examining Evidence for the Schizo-Obsessive Subtype

Abstract

Obsessive-compulsive symptoms or obsessive-compulsive disorder (OCD) are estimated to occur in up to 30% of patients with schizophrenia. Whether this subgroup of patients is cognitively, affectively, or physiologically distinct remains unclear. 204 schizophrenia patients, 15 who also met criteria for a diagnosis of OCD, and 147 healthy controls were examined on several intermediate phenotypes. The patient groups did not differ from each other except that the co-morbid group exhibited an elevated rate of eye-tracking dysfunction. Results suggest that OCD-co-morbid patients did not comprise a distinct subgroup based on the measures studied here, although systematic assessment of larger cohorts is warranted.

1. Introduction

Obsessive-compulsive symptoms (OCS) or obsessive-compulsive disorder (OCD) occur in approximately 30% of patients with schizophrenia (Byerly et al., 2005; Tibbo et al., 2000)--a much higher prevalence than is observed in the general population (1.2-2.4%; Foa et al., 1995). OCS are also increased in the psychosis prodrome (Niendam et al., 2008), are associated with poorer prognosis (Fenton & McGlashan, 1986), and may reflect reduced functional connectivity in basal ganglia and prefrontal cortex (Harrison et al., 2009).

Although some evidence suggests that OCD-co-morbid schizophrenia is a possible subtype of schizophrenia (e.g., Aoyama et al. 2000; Iida, et al. 1998; Levine et al. 1998), the clinical relationship between schizophrenia and OCD is poorly understood (Bottas et al. 2005). The two most common designs are to use categorical diagnostic criteria to establish co-morbidity, or to evaluate severity of OCS using symptom scales and checklists. Indeed, the frequency of OCS in schizophrenia seems to vary as a function of how extensive the OCS screening methods were (Byerly et al. 2005; Nolfe et al 2010).

The use of intermediate phenotypes may help to identify schizophrenia subtypes with distinct underlying pathophysiologies. In the present study, we examined neurocognitive phenotypes associated with genetic liability to schizophrenia and severity of psychotic symptoms in schizophrenia patients with and without co-morbid OCD in order to determine whether any of these measures distinguished between the schizophrenia subgroups. These phenotypes include eye tracking dysfunction (Levy et al. 1993; Levy et al. 2010a, thought disorder (Holzman et al. 2005; Levy et al. 2010b), and craniofacial dysmorphology (Deutsch et al. 2000; Deutsch et al. 2012).

2. Methods

2.1. Participants

A total of 189 patients with a diagnosis of schizophrenia or schizoaffective disorder, 15 schizophrenia-OCD-co-morbid patients, and 147 healthy comparison subjects participated. Normal controls did not meet criteria for a diagnosis of psychosis (lifetime), an Axis II schizophrenia-related personality disorder, or nonpsychotic bipolar disorder, and did not have a family history of psychosis, psychiatric hospitalization, or suicide. Information concerning demographics, medications, and global level of functioning is presented in Table 1.

Table 1. Demographic Characteristics, Global Assessment of Functioning, and Antipsychotic Medication: Schizophrenia Patients (with and without OCD) and Controls.

	SZ	SZ-OCD	NC
N	189	15	147
Age (Years)	47.2 (0.0)	41.3 (8.0)	39.9 (14.9)
Gender (% Female)	45	50	56
Race (% Caucasian)	91	100	85
SES	2.5 (1.0)	2.5 (1.3)	2.3 (.9)
Est. VIQ	98.5 (22.1)*	100.7 (10.2)	105.8 (11.2)
Years Education	13.9 (2.3)*	15.0 (2.8)	15.0 (2.3)
GAS	37.7 (10.0)	34.1 (4.8)	75.7 (11.8)
Antipsychotics
Typical only	53	1
Atypical only	129	13
Typical and Atypical	23	0

SZ = schizophrenia patients, SZ-OCD = schizophrenia patients with co-morbid obsessivecompulsive disorder diagnosis, NC = normal controls; SES = socio-economic status based on a revised version of the Hollingshead Scale (1965), Est VIQ = estimated verbal IQ based on the vocabulary subtest of the Wechsler Adult Intelligence Scale - Revised (WAIS-R) (Wechsler, 1981), GAS = global assessment of functioning score. Mean (standard deviation)

^*p < 0.001, comparison with control participant group

2.2. Diagnostic Assessment

Axis I disorders were assessed in patients and controls using the Structured Clinical Interview for DSM-IV (SCID, First et al., 2002). Schizotypal, schizoid, and paranoid personality disorders were assessed in controls using the Structured Interview for Schizotypal Symptoms (Version 1.5; Kendler, 1989). An experienced clinician administered the interviews. An independent group of senior diagnosticians reviewed the interview materials and all available hospital records and assigned consensus DSM-IV (APA, 1994) Axis I and Axis II (nonpsychotic individuals only) diagnoses based on best-estimate methods (Leckman et al., 1982). These subject groups are part of a larger sample that also includes individuals who meet criteria for a diagnosis of bipolar disorder with psychotic features and relatives of the three subject groups. The Brief Psychiatric Rating Scale (BPRS; Overall & Gorham, 1962) was used to assess symptom severity in the patient groups. The Global Assessment of Functioning score (GAF; DSM-IV, 1994) was used to estimate level of functioning in all subjects. Chlorpromazine equivalents were calculated using widely adopted methods (Davis, 1974; Woods, 2003).

2.3. Craniofacial Dysmorphology

Quantitative dysmorphology methods were used to assess craniofacial asymmetries arising along an embryologically derived region implicated in schizophrenia maldevelopment: the junction of two embryonic primordia, the frontonasal and the maxillary prominence derivatives. Using the methods of Farkas and Deutsch (1996), standard anthropometric techniques were used to measure these regions, including the orbits, surrounding architecture, and alar conformation. Higher scores indicate more abnormal measurements. This quantitative measure has been shown to be elevated among probands with schizophrenia and their unaffected first-degree relatives, but not among those with bipolar affective disorder (Deutsch et al., 2012).

2.4. Thought Disorder

Thought disorder was assessed using the Thought Disorder Index (TDI; Johnston & Holzman, 1979; Solovay et al, 1986). Verbatim transcripts of responses to ten cards of the Rorschach were scored by two experienced raters. Deviant verbalizations and combinatory thinking, two components of the total score, were also calculated. Higher scores indicate greater deviance. Each record was also assigned a TDI “diagnosis,” indicating whether the thought disorder that was present was characterized by features consistent with schizophrenia, affective illness, or neither.

2.5. Eye Tracking Dysfunction

A target subtending 1.25 degrees of visual angle moved horizontally in a 0.4 Hz sine wave for 30 seconds. The subject was instructed to follow the target as carefully as possible. Eye movements were recorded using infrared reflectometry. Eye position tracings were assigned a qualitative rating (normal/abnormal) and were scored for the following quantitative variables using an interactive software program:

• Closed-loop peak gain - the ratio of eye velocity to target velocity during the period of peak target velocity;

• Frequency of catch-up saccades – frequency of saccades in the same direction as the target that re-positioned the eyes on the target when eye velocity lagged behind that of the target;

• Frequency of compensatory saccades - sum of the frequency of catch-up and back-up saccades; and

• Frequency of intrusive saccades - the sum of the frequency of square-wave jerks and anticipatory saccades.

The interview, diagnostic review, collection and processing of phenotype measures, and the clinical ratings were carried out blind to group membership.

2.6. Statistical Analysis

Shapiro-Wilk tests of normality were used to examine the distributions of the dependent measures within groups. Group comparisons of continuously distributed demographic, clinical, and intermediate phenotype variables were performed using Wilcoxon tests. This nonparametric test was used instead of the usual 2-sample t-test, because it is essentially as powerful as the t-test, but does not require that the variable be normally distributed within groups and is therefore a conservative choice. Chi square tests were used to assess group comparisons on qualitative variables. An α level of 0.01 was used as the minimum requirement for statistical significance.

3. Results

Demographic characteristics of the schizophrenia, SZ-OCD, and control groups are presented in Tables 1 and 2. OCD and non-OCD patient groups did not differ in medication dosage in chlorpromazine equivalence units (t[13] = 1.05, P = 0.31). Non-OCD schizophrenia patients had significantly fewer years of formal education (Z = 4.40, P < 0.001) and a significantly lower estimated verbal IQ than controls (Z = 4.77, P < 0.001); SZ-OCD patients did not differ from controls in year of education (Z = 1.47, P = 0.14) or estimated verbal IQ (Z = 0.27, P = 0.79). No other significant differences between groups on the demographic variables were observed. Tables 2 and 3 present data comparing the two schizophrenia groups on demographic and clinical variables. The two subgroups of patients did not differ significantly on any measures of symptom severity (all Ps >0.01).

Table 2. Comparisons of SZ and SZ-OCD Groups on Demographic Variables.

	SZ Mean (SD)	SZ-OCD Mean (SD)	p	d
Age	38.5 (9.6)	41.3 (8.0)	ns	−0.32
SES	2.5 (1.0)	2.5 (1.3)	ns	0.08
Race			ns	−0.13
Handedness			ns	0.02
Est. VIQ	98.5 (22.1)	100.7 (10.2)	ns	0.10
Education	13.9 (2.3)	14.9 (2.8)	ns	0.44
GAS	38 (10)	34 (5)	ns	0.36
CPZ	617 (490)	945 (1076)	ns	0.39

d = Cohen’s effect size for group contrast

SES = socioeconomic status, Est VIQ = estimated verbal IQ, GAS = global assessment of functioning score, CPZ = chlorpromazine equivalent dose.

Table 3. Comparisons of SZ and SZ-OCD Groups on Measures of Symptom Severity based on the BPRS.

	SZ (M)	SZ-OCD (M)	p	d
Somaticism	2.603	1.648	0.04	−0.50
Anxiety	3.027	1.814	0.06	0.54
Emotional Withdrawal	3.627	1.704	0.22	0.30
Disorganization	3.107	1.898	0.28	−0.32
Guilt	2.509	1.744	0.31	−0.21
Tension	2.201	1.593	0.04	0.55
Mannerism	2.342	1.707	0.53	−0.16
Grandiose	2.156	1.717	0.55	−0.17
Depressed Mood	2.719	1.609	0.96	−0.00
Hostility	1.862	1.648	0.28	0.29
Suspiciousness	3.924	1.483	0.04	0.50
Hallucinations	2.955	1.995	0.71	0.12
Motor	2.182	2.251	0.52	−0.18
Uncooperativeness	1.741	1.420	0.46	−0.13
Unusual Thought Content	4.121	1.334	0.34	0.28
Blunted Affect	3.831	2.130	0.20	0.33
Excitement	1.629	1.267	0.72	0.07
Disorientation	1.277	1.026	0.32	−0.27
BPRS Total	47.759	14.788	0.49	0.16
BPRS Factor Scores
BPRS Affect	10.857	4.421	0.94	−0.05
BPRS Positive	12.339	5.920	0.90	−0.01
BPRS Negative	9.640	4.187	0.36	0.21
BPRS Resistance	7.527	3.320	0.11	0.38
BPRS Activation	6.156	3.562	0.18	0.20

d = Cohen’s effect size for group contrast

Participants with a diagnosis of schizophrenia, with or without OCD, differed from control participants on all phenotype measures, with the exception of frequency of intrusive saccades. Participants with schizophrenia exhibited significantly more deviance involving the frontonasal-maxillary junction, more thought disorder, and worse performance on all quantitative indices of eye-tracking.

SZ-OCD patients showed a significantly increased rate of ETD (72%) compared with non-OCD schizophrenia patients (44%; X² = 6.3, P < 0.01; Controls = 12%). However, the SZ and SZ-OCD groups did not differ on any of the quantitative eye-tracking measures. Patients with and without OCD did not differ in frequency of SZ-related TD (χ² = 0.003, P > 0.96), quantitative thought disorder measures, or frontonasal maxillary junction scores.

4. Discussion

SZ and SZ-OCD patients closely resembled each other on clinical variables and on the intermediate phenotype measures used in this study, suggesting that SZ-OCD is not a distinct subtype. The one exception to this pattern was the significantly higher rate of eye-tracking dysfunction (ETD), assessed qualitatively, in schizophrenia patients with co-morbid OCD.

Although schizophrenia has been systematically associated with abnormalities in pursuit initiation and maintenance (for a review, see Levy et al., 2010), findings on oculomotor function in OCD have been much less consistent. In a recent review of eye movement studies in OCD, the authors concluded that pursuit deficits in this clinical population “are much less robust and systematic” than in schizophrenia, and speculated that this difference may be due to the greater involvement of dorsolateral prefrontal cortex in schizophrenia vs. orbitofrontal prefrontal cortex in OCD (Jaafari et al., 2011).

The superiority of the qualitative rating of eye tracking in distinguishing OCD-co-morbid schizophrenia patients from non-OCD patients is consistent with other reports that this measure discriminates patients and relatives from controls better than specific quantitative measures (e.g., Friedman et al. 1995; Keefe et al. 1997; Levy et al. 2000). A recent meta-analysis of eye tracking dysfunction in schizophrenia quantified the results of studies that used specific quantitative measures of pursuit performance and global measures (O’Driscoll and Callahan, 2008). They found that effect sizes for global variables were uniformly large, and that the largest effect size was obtained for qualitative ratings (d =1.55). Global measures are less physiologically informative than specific measures, however, underscoring the need to systematically compare schizophrenia patients with and without OCD and non-schizophrenic-OCD patients using eye movement tasks that can help to determine whether the specific processes that underlie eye tracking dysfunction differ among these groups.

Limitations

The sample size of the OCD-co-morbid schizophrenia patients was small, which may have reduced power to detect differences between this group and the non-OCD-co-morbid group. Also, we compared the two groups of schizophrenia patients on a select group of measures that may not be sensitive to subgroup distinctions. Thus, we cannot conclusively rule out the possibility that OCD-co-morbid schizophrenia is a subtype. Similarly, the OCD co-morbid subgroup was identified on the basis of meeting diagnostic criteria for both disorders. Although this is a particularly conservative criterion for co-morbidity, it is possible that detailed screening of OCS (i.e., Y-BOCS scores) would have yielded significant associations between number or severity of OCS and the intermediate phenotypes studied here.

Table 4. Schizophrenia Group Comparisons: Craniofacial Dysmorphology, Thought Disorder Variables, and Quantitative Eye Tracking Dysfunction.

	Group	N	Mean (SD)	P	d
Craniofacial Dysmorphology	SZ	176	11.49 (5.7)	0.76	−0.15
	SZ-OCD	11	10.64 (5.9)
Thought Disorder
Combinatory Thinking	SZ	194	3.46 (3.1)	0.45	0.08
	SZ-OCD	14	3.71 (2.3)
Deviant Verbalizations	SZ	194	5.51 (5.2)	0.18	−0.40
	SZ-OCD	14	3.43 (2.7)
Thought Disorder Index Score	SZ	194	22.86 (22.7)	0.83	−0.15
	SZ-OCD	14	19.52 (15.6)
Eye-Tracking
Catch-Up Saccade Frequency (/min)	SZ	187	82.33 (44.6)	0.45	−0.25
	SZ-OCD	14	71.34 (29.2)
Compensatory Saccade Frequency (/min)	SZ	187	93.54 (45.6)	0.67	−0.18
	SZ-OCD	14	85.38 (27.5)
Intrusive Saccade Frequency (/min)	SZ	187	14.61 (19.4)	0.23	0.20
	SZ-OCD	14	18.49 (17.4)
Gain	SZ	183	0.80 (0.2)	0.49	−0.11
	SZ-OCD	14	0.78 (0.1)

d = Cohen’s effect size for group contrast

SZ = schizophrenia patients, SZ-OCD = schizophrenia patients with co-morbid obsessive-compulsive disorder diagnosis.