Negative Symptoms in Schizophrenia: Their Longitudinal Course and Prognostic Importance

Transcription

1 VOL 11, NO. 3, 1985 Negative Symptoms in Schizophrenia: Their Longitudinal Course and Prognostic Importance 427 by Michael F. Pogue-Geile and Martin Harrow Abstract Negative and positive symptoms were investigated longitudinally in 39 young schizophrenic patients at two assessments approximately 2.5 and 5 years after hospital discharge. Negative symptoms, such as flat affect and poverty of speech, which were assessed at the first, were found to be effective prognostic signs in schizophrenic patients for predicting later poor role functioning at the second. The prognostic importance of negative symptoms was predominantly due to their tendency to occur in patients who were already functioning poorly in social and instrumental areas at the first, and who tended to continue doing poorly at the second. Contrary to some current hypotheses, positive symptoms, such as delusions and hallucinations, were also found to be prognostic of later deficits in role functioning at the second. Negative symptoms appeared to be generally persistent over time, although there was some tendency toward remission. Potential models of the etiology of negative symptoms and their role in schizophrenia are proposed. The construct of negative symptoms has come to play an important role in recent theorizing about schizophrenia (e.g., Strauss, Carpenter, and Bartko 1974; Crow 1980; Andreasen and Olsen 1982; Lewine, Fogg, and Meltzer 1983; Pogue-Geile and Harrow 1984a). Negative symptoms, such as flat affect and poverty of speech, have been proposed to be important behavioral individual differences among schizophrenic patients that have both pathophysiological and psychopathological correlates. Much of this theorizing concerning negative symptoms has been in the general tradition surrounding other individual differences that also have been hypothesized to be important within schizophrenia, such as the paranoid/nonparanoid dimension (e.g., Kendler and Davis 1981) and the process/reactive dimension (e.g., Garmezy 1968; Zubin et al. 1961; Stofflemayr, Dillavaou, and Hunter 1983). More specifically, negative symptoms have been proposed by several investigators (e.g., Strauss, Carpenter, and Bartko 1974; Crow 1980) to be persistent over time, prognostic of poor outcome, associated with intellectual deficits, and unresponsive to antipsychotic medication. Crow (1980) has further proposed an anatomical model for negative symptoms to mediate the relationships hypothesized above, in which negative symptoms (his type II symptom syndrome) represent a behavioral manifestation of some structural brain abnormality, such as cortical atrophy. Several studies have explored this hypothesized association between negative symptoms and ventricular enlargement on computed tomography scans (Johnstone et al. 1976; Andreasen et al. 1982; Nasrallah et al. 1983; Luchins, Lewine, and Meltzer 1984; Pearlson et al. 1984) and other biological variables, such as platelet monoamine oxidase (Lewine and Meltzer 1984), with somewhat mixed results. Studies have been generally more unequivocal in finding negative Portions of this research were presented at the 137th Annual Meeting of the American Psychiatric Association, Los Angeles, California, May 5-11, Reprint requests should be sent to Dr. M.F. Pogue-Geile, Department of Psychology, University of Pittsburgh, Pittsburgh, PA

2 428 SCHIZOPHRENIA BULLETIN symptoms to be associated with intellectual deficits in chronic schizophrenic patients (Johnstone et al. 1978, 1981; Owens and Johnstone 1980; Pogue-Geile and Harrow, in press). Similarly, negative symptoms appear to be generally unresponsive to antipsychotic medication (Johnstone et al. 1978; Angrist, Rotrosen, and Gershon 1980). Although the results of these studies have generally supported the potential utility of further study of the negative symptom construct, few studies have investigated either the longitudinal course or the prognostic value of negative symptoms, both of which are key hypotheses in theory concerning negative symptoms. Earlier studies have suggested that restricted affect during the inpatient phase may be a poor prognostic sign for later outcome in schizophrenia (e.g., Carpenter et al. 1978). However, more recent work on negative symptoms has focused almost exclusively on schizophrenic inpatients (cf. Johnstone et al. 1979). Therefore, one of the primary purposes of the present research was to investigate these issues of the longitudinal course and prognostic importance of negative symptoms within the context of a prospective study of young schizophrenic patients. Positive psychotic symptoms, such as hallucinations and delusions, are generally distinguished from negative symptoms. In contrast to the hypothesized persistence and poor prognostic implications of negative symptoms, positive symptoms during the acute phase have been theorized to fluctuate over time and to have relatively less prognostic importance (Strauss, Carpenter, and Bartko 1974; Crow 1980). Zubin and colleagues have also proposed a model for the course of schizophrenia in which positive psychotic symptoms are predominantly episodic (Zubin and Spring 1977; Zubin, Magaziner, and Steinhauer 1983). Crow (1980) has suggested that positive symptoms (his type I symptom syndrome) may be caused by a fluctuating, functional dopamine excess, in contrast with the structural brain abnormality hypothesized to underlie negative symptoms. Because of these differing predictions concerning the longitudinal characteristics of negative and positive symptoms, the present study also investigated the prognostic importance of positive symptoms. Thus, the present prospective, longitudinal study of young schizophrenic patients early in the course of their disorder is among the first to investigate the prognostic importance in schizophrenia of negative symptoms and their relationship with positive symptoms. The following specific research questions are addressed: Are negative symptoms prognostically important for the future adjustment of schizophrenia? If negative symptoms are prognostically important, do they identify schizophrenic patients before a decline in functioning, or do they characterize patients who are already functioning poorly? Do some of the major positive symptoms, such as hallucinations and delusions, have less prognostic utility than negative symptoms? What is the relationship between negative and positive symptoms in predicting later functioning? What is the longitudinal course of negative symptoms in schizophrenia? Methods Patient Sample. Thirty-nine young schizophrenic patients who had been prospectively assessed at three points during the early course of their disorder participated in this study. Subjects were selected from consecutive inpatient admissions between the ages of 18 and 30 years old, to either Michael Reese Medical Center (n = 21), a private hospital, or to the Illinois State Psychiatric Institute (ISPI) (n - 18), a State research institution. The subjects were part of a larger project, the Chicago Followup Study, based at Michael Reese Medical Center and The University of Chicago, which is investigating the longitudinal course of negative symptoms, thought disorder, psychosis, and functioning in schizophrenia and other major psychopathology (Harrow and Quinlan 1985; Harrow, Silverstein, and Marengo 1983; Pogue-Geile and Harrow 1984a). Within several weeks of their index admission, patients were assessed with the Schedule for Affective Disorders and Schizophrenia (SADS) (Spitzer and Endicott 1978) or with the Schizophrenia State Inventory (SSI), a taped, structured interview (Grinker and Holzman 1973). Based on these interviews and inpatient charts, the patients were diagnosed by a trained diagnostic team, unaware of other test and outcome data, using the Research Diagnostic Criteria (RDC) (Spitzer, Endicott, and Robins 1978). The current sample of 39 schizophrenic patients diagnosed according to RDC was composed of patients who had been followed up twice after their index hospitalization and who had negative symptom behavior ratings available at the first. These subjects had been followed up for the first time a median of 2.6 years after their index hospitalization discharge. The second took place a median of 2.1 years after the first, for a median total of 4.9 years after their

3 VOL. 11, NO. 3, index hospitalization discharge. All s involved in-person interviews averaging 3-4 hours. Both first and second s consisted of an interview with a revised version of the SADS, negative symptom behavior ratings, the Harrow Functioning Interview (used to rate outcome measures), and select other measures of symptomatology and thought disorder. Twenty-nine schizophrenic patients in the current sample were also included in a previous report on negative symptoms at the first (Pogue-Geile and Harrow 1984a). The current study includes 10 schizophrenic patients in addition to the initial 29 and presents data on the second. The 39 schizophrenic patients in the present sample were from the following RDC subtypes: 19 paranoid, 1 disorganized, 18 undifferentiated, and 1 catatonic. All except four of these patients also met DSM-III criteria for schizophrenia (American Psychiatric Association 1980). The remaining four were diagnosed as DSM-III schizophreniform disorders. Table 1 presents further characteristics of the patient sample. In general, this was a group of young patients who were in the early course of their disorder. Their average age at index hospitalization was 23.8 years old. Thirty-nine percent (15) of the patients were first admissions and an additional 31 percent (12) had only one or two prior hospitalizations. Sixty-seven percent of the patients were male and most were from middle-class homes. The average amount of education by index hospitalization was approximately 12 years. Fifty-four percent of the subjects at the first and 56 percent at the second were taking antipsychotic medication. Five subjects were rehos- Table 1. Sample characteristics Characteristic Sex (% male) Parental SES' At index admission Age at Index Education (years) Number of hospitalizations before index At 1st Age at 1st Rehospitalized at 1st Taking antipsychotic medication at 1st At 2nd Age at 2nd Rehospitalized at 2nd Taking antipsychotic medication at 2nd Schizophrenic patients (n = 39) 26 (67%) 3.2 (1.4) (4.1) 12.2(2.0) 1.9(2.3) 26.7 (4.8) 5 (13%) 21 (54%) 29.1 (4.7) 0 (0%) 22 (56%) ' Socloeconomlc status scaled according to Holllngshead and Redllch's (1958) Index, where 5 Is the lowest social class. 1 Means (standard deviations). pitalized at the time of the first and none of them were rehospitalized at the second. Measures of Negative Symptoms. Negative symptoms were rated by trained interviewers from behavior shown during the 3- to 4-hour assessment interviews. Behavior rating scales based on items from Strauss and Carpenter's Psychiatric Assessment Interview (PAI) (Carpenter et al. 1976) were used to rate the three negative symptoms (1) poverty of speech, (2) flat affect, and (3) psychomotor retardation. The item content of the scales is similar to that proposed by Andreasen (1982). Table 2 presents the item content of the scales. Each of the 13 scale items was rated as either: 0, absent; 1, present but in mild form; or 2, present in marked form. In addition to the individual subscales of poverty of speech, flat affect, and psychomotor retardation, a negative symptom total score was also computed as an estimate of the severity of the overall negiitive symptom syndrome, which was a sum of the subscale scores. In order to estimate the prevalence of the negative symptom syndrome in addition to its severity, a categorical negative symptom index was also developed. For this purpose, the negative symptom syndrome was considered to be present whenever one or more of the three subscales was scored 2 or greater. The three subscales were quite internally consistent, with Cronbach alphas ranging from.48 to.75. Similarly, the subscales were all significantly intercorrelated and the Cronbach alpha for the total score was.81, indicating that the items

4 430 SCHIZOPHRENIA BULLETIN Table 2. Item composition of negative symptom behavior rating scales Poverty of speech Flat affect Psychomotor retardation 1. Long lapses before replying to questions 2. Restriction of quantity of speech 3. Patient fails to answer 4. Speech slowed 5. Blocking 1. Avoids looking at interviewer 2. Blank, expressionless face 3. Reduced emotion shown when emotional material discussed 4. Apathetic & uninterested 5. Monotonous voice 6. Low voice, difficult to hear 1. Slowed In movements 2. Reduction of voluntary movements No/e. Items are from the Behavior Rating Inventory of the Psychiatric Assessment Interview (Carpenter et al. 1976). loaded on a common factor. The negative symptom scales also showed satisfactory interjudge reliability. Intraclass correlations between two independent raters for a sample of 20 videotaped interviews ranged from.67 to.88. Further details concerning the development of the negative symptom behavior rating scales can be found in Pogue-Geile and Harrow (1984a). Measures of Positive Symptoms. In addition to these measures of negative symptoms, two major types of positive symptoms were also assessed at both interviews. The presence of (1) hallucinations or (2) delusions of any type during the past month was rated from the revised SADS interview. Both of these symptoms were rated on a 3- point scale of: 1, absent; 2, mild or equivocal; or 3, definitely present. Further details on this rating scheme are available in Harrow and Silverstein (1977). As a parallel to the negative symptom total score, an overall positive symptom index was developed based on the presence of either hallucinations or delusions. This index was scaled as: 1, hallucinations or delusions absent; 2, at least one mild or equivocal, but no definite ratings; or 3, at least one definite rating of a hallucination or delusion. Measures of Functioning at Followup. The measures of role functioning at both s were based on our structured Functioning Interview (Harrow et al. 1978). The Levenstein, Klein, and Pollack (1966) (LKP) scale was used to rate overall functioning in the past year. This scale was originally constructed for use in a large study of the posthospital adjustment of discharged psychiatric patients. It is based on employment, symptomatology, social functioning, independence of functioning, and rehospitalizations. In addition, the multidimensional scales developed by Strauss and Carpenter (1972) for instrumental work functioning, social functioning, and rehospitalization were also used to provide more detailed assessments of functioning at both s during the year preceding each. Results Negative Symptoms at the First Followup: Prevalence and Cross- Sectional Correlates. In a previous study, we examined several major issues concerning negative symptoms during the posthospital phase of schizophrenia (Pogue-Geile and Harrow 1984a). In order to clarify the comparability with this previous work and to partially replicate our prior results, we have repeated the main analyses conducted on our earlier sample (Pogue-Geile and Harrow 1984a). These analyses were also performed on the current sample at the first and are presented in summary form herein. Although the majority of subjects (29) are shared in both studies, the current sample includes 10 additional subjects and data from the second. The level of negative symptoms at the first in the current sample was very similar to that in our earlier schizophrenic sample (Pogue-Geile and Harrow 1984a). The mean negative symptom total score was 2.44 (SD ) and the prevalence of overall negative symptoms, using our dichotomized negative symptom index described above, was 41 percent (16). In the current sample, the negative symptom total score was significantly associated with Hollingshead and Redlich's (1958) index of parental social class (r.43, p <.01), such that negative symptoms were more severe in schizophrenic patients from

5 VOL 11, NO. 3, lower socioeconomic status families. Similar to our previous results, negative symptoms did not differ significantly between those schizophrenic patients who were taking antipsychotic medications at the first and those who were not. There was also a nonsignificant trend for the negative symptom total score to be more severe in schizophrenic patients who were currently rehospitallzed at the time of the first compared with those who were living in the community (t = 2.00, df = 36, p <.06). As before, negative symptoms at the first in schizophrenic patients were not significantly associated with the demographic variables of: age at first, sex, number or duration of hospitalizations before index admission, number of years between first hospitalization and index admission, or length of time between index hospitalization and the first. In terms of measures of functioning before index hospitalization, the negative symptom total score at the first was significantly associated with the number of years of education completed before index admission (r =.43, p <.01) and with ratings of poor social functioning before index admission (r ".38, p <.02). Negative symptoms at the first were also associated with concurrent measures of poor functioning at the first in the current sample. The negative symptom total score was significantly associated with poor functioning at the first for the LKP scale, which measures overall functioning (r.35, p <.05), and the Strauss- Carpenter individual indexes of instrumental work functioning (r D.39, p <.02) and rehospitalization (r , p <.05). Analyses in the current sample also produced similar results to those from our previous sample concerning the cross-sectional relationship within schizophrenia between negative and positive symptoms at the first. We found no significant association between the total negative symptom score and our positive symptom index based on the presence of delusions or hallucinations (r -.02, p <.90). Prognostic Importance of Negative Symptoms for Later Functioning. The results that follow represent the next step in our investigation of negative symptoms, focusing on the prognostic importance of negative symptoms, and build upon our previous work in this area (Pogue- Geile and Harrow 1984a). Table 3 presents data on the prognostic relationship between negative symptoms assessed at the first and later outcome functioning at the second. Overall, there were a number of significant associations, although the relationships were generally of modest size. For the individual negative symptom subscales, psychomotor retardation assessed at the first was significantly predictive of both the Strauss-Carpenter rehospitalization and instrumental functioning scales at the second. Flat affect and poverty of speech also significantly predicted the Strauss-Carpenter rehospitalization and instrumental functioning ratings at the second. In addition, there was a nonsignificant trend for poverty of speech to be associated with later overall functioning as measured by the LKP scale. Generally more important than the individual subscales are the overall indexes of the negative symptom syndrome. In terms of the prognostic value of the severity of the negative symptom syndrome, the negative symptom total score was significantly Table 3. Prognostic relationships between negative symptoms at 1st and outcome functioning at 2nd Negative symptom at 1st Psychomotor retardation Flat affect Poverty of speech Negative symptom total score Dichotomized negative symptom index Overall functioning (LKP scale).19 (.13).14 (.19).24 (-07).24 (.07).31 (.03) Functioning at 2nd Strauss-Carpenter scales.65 (.0001).26 (.05).38 (.008).46 (.002).25 (.06) Renos- pltal- Ization Instrumental work Social functioning functioning.11 (.26).06 (.36).11 (.25).04 (.40).09 (.29).26 (.05).27 (-05).26 (.05).33 (02).30 ( 03) More. Schizophrenic patients, n = 39. Pearson correlations (probabilities). High scores on functioning scales Indicate poor functioning (Strauss-Carpenter scoring reversed).

6 432 SCHIZOPHRENIA BULLETIN correlated with later poor functioning at the second on the Strauss-Carpenter scales of rehospitalization and instrumental functioning, with a nonsignificant trend on the LKP scale of overall functioning. As an estimate of the prognostic importance of the prevalence of the overall negative symptom syndrome, table 3 also presents the predictive relationship between the dichotomized negative symptom index and later functioning. This categorical negative symptom index also significantly predicted later poor overall functioning, as measured by the LKP scale and Strauss-Carpenter instrumental functioning ratings, with a nonsignificant trend to predict later Strauss- Carpenter rehospitalization rating. In order to evaluate any potential confounding effects of chronicity that may have been already established before index admission (Strauss and Carpenter 1974, 1977) on the prognostic relation of negative symptoms to later functioning, control analyses were performed based only on that subsample of schizophrenic patients with no hospital admissions before index admission (n = 15). Among these first-admission schizophrenic patients, the dichotomized negative symptom index at the first continued to be a significant predictor of later poor functioning at the second on the overall LKP scale (r -=.54, p <.02), and on the Strauss-Carpenter social functioning scale (r.59, p <.01). There were also nonsignificant trends toward predicting instrumental functioning (r =.28, p <.16) and rehospitalization (r.21, p <.22). There were also no significant differences in the prognostic value of negative symptoms between those schizophrenic patients who were taking antipsychotic medication at the first and those who were not. In addition, for the current sample, there were no significant differences in the prognostic importance of negative symptoms between males and females. Table 4 presents a closer examination of the prognostic relationship between negative symptoms and later role functioning. This table details the association between the dichotomized negative symptom index at the first and the LKP scale of overall functioning at the second. It again illustrates the significant prognostic importance of negative symptoms (Kendall's Tau C =.29, p <.04), as 69 percent of those schizophrenic patients with negative symptoms at the first showed severe overall functioning impairment at the second. The severe impairment range on the LKP scale of overall functioning indicates very poor functioning in the past year with continuous marked symptoms, very low level of self-support, and possible rehospitalization. However, table 4 also indicates that negative symptoms represent only one of several possible pathways to such poor functioning in schizophrenia, as 39 percent of schizophrenic patients who did not initially have negative symptoms at the first also showed severe functioning impairment at the second. Thus, although the risk for later poor functioning in schizophrenic patients without initial negative symptoms was quite high (39 percent), the presence of negative symptoms at the first almost doubled this risk for later poor functioning (69 percent). Detailed Analysis of the Prognostic Importance of Negative Symptoms. Given the prognostic importance of negative symptoms for predicting later functioning, it becomes important to explore possible factors that may influence this association over time. Specifically, we were interested in whether negative symptoms at the first were occurring in schizophrenic patients who were currently functioning adequately but then declined in functioning by the second. In this case, negative symptoms might predate a later decline in functioning. Alternatively, even in this sample of young schizophrenic patients who are in the relatively early course of their disorder, Table 4. Categorical association between overall negative symptoms at 1st and overall functioning at 2nd Overall negative symptom index at 1st Absent Present Overall Good (1-2) 6 (26%) 2 (12%) functioning at 2nd (LKP scale) Equivocal (3-6) 8 (35%) 3 (19%) Note. Schizophrenic patients, n = 39. Row percentages presented. Poor (7-8) 9 (39%) 11 (69%)

7 VOL. 11, NO. 3, negative symptoms may be occurring primarily in subjects who were already functioning poorly at the first and would tend to remain at this level by the second. Stated another way, does the presence of negative symptoms at the first still contribute to the prediction of outcome functioning at the second after controlling for the initial level of functioning at the first? To investigate this question, we completed partial correlations between the overall negative symptoms syndrome at the first and the LKP scale of overall functioning at the second, after removing the shared variance of overall outcome at the first. Trie partial correlation between the negative symptom total score at the first and the LKP measure of overall functioning at the second, controlling for the initial level of overall functioning at the first, was not significant (r «.02, p <.45). This result suggests that the prognostic importance of negative symptoms during the early posthospital course of schizophrenia predominantly overlaps with the longitudinal stability of poor functioning. The prognostic value of negative symptoms seems to be largely enhanced by their tendency to occur in schizophrenic patients who are already functioning poorly at the first and who tend to continue to function poorly by the second. Table 5 illustrates these interrelationships in cross-tabular form. The upper portion of the table presents the association between negative symptoms at the first and subsequent overall functioning at the second for those schizophrenic patients who were initially functioning adequately at the first. Among these schizophrenic patients who were functioning adequately at the first. negative symptoms were relatively infrequent (26 percent; n =» 5), and although they increased the risk for later poor functioning at the second (40 percent vs. 14 percent) the relationship was not significant. Thus, although there may be some tendency during this phase of the disorder for negative symptoms to predate a decline in functioning, it is probably not the major factor accounting for their prognostic importance. The lower portion of table 5 presents the association between negative symptoms at the first and overall functioning at the second for those schizophrenic patients who were initially functioning poorly at the first. Among these patients, negative symptoms were more frequent (present in 55 percent) than in patients who showed better functioning at the first. However, the presence of negative symptoms in this group of already poorly functioning schizophrenic Table 5. Prognostic relationship between negative symptoms at 1st and overall functioning at 2nd (controlling for initial overall functioning at 1st ) A. Schizophrenic patients with good/equivocal (1-6) overall functioning (LKP scale) at 1st (n = 19) 2nd functioning (LKP scale) Negative symptoms Good/equivocal Poor at 1st (1-6) (7-8) Absent Present 12(86%) 3 (60%) Fisher's Exact Test, p <.27. Row percentages. 2(14%) 2 (40%) B. Schizophrenic patients with poor (7-8) overall functioning (LKP scale) at 1st (n = 20) 2nd functioning (LKP scale) Negative symptoms at 1st Absent Present Good/equivocal (1-6) 2 (22%) 2 (18%) Fisher's Exact Test, p <.63. Row percentages. Poor (7-8) 7 (78%) 9 (82%)

8 434 SCHIZOPHRENIA BULLETIN patients did not significantly increase the risk for later poor functioning at the second (82 percent vs. 78 percent). Therefore, these analyses suggest that much of the prognostic importance of negative symptoms reflects the fact that they tend to occur in a subgroup of schizophrenic patients who are already functioning poorly at the first and who then tend to continue to function poorly at the second. It is possible that at earlier phases of the schizophrenic disorder, such as the inpatient phase, negative symptoms may more frequently predate declines in general functioning. However, it may be even more likely that negative symptoms and general role functioning deficits are both influenced by some shared factors and thus tend to occur most often concurrently. Longitudinal Course of Negative Symptoms. Table 6 presents the longitudinal course of the severity of each individual negative symptom subscale at both the first and second s. Each negative symptom subscale showed a decline from the first to second s, although these declines were not statistically significant. Table 7 presents the longitudinal course of the prevalence of the dichotomized overall negative symptom index. There was a significant longitudinal association for this categorical index of the negative symptom syndrome (x 2 " 9.97, df «= 1, p <.002) and the oddsratio measure of association was Among those schizophrenic patients who experienced negative symptoms at the first, 55 percent also showed negative symptoms at the second. In contrast, only 5 percent of those patients who did not initially show negative symptoms at the first Table 6. Longitudinal course of severity of negative symptoms from 1st to 2nd Negative symptom Psychomotor retardation Flat affect Poverty of speech Negative symptom total score Note. Schizophrenic patients, n. 1 Means (standard deviations). showed later negative symptoms at the second. However, within the context of this significant stability, it also appears that some schizophrenic patients showed a reduction in negative symptoms. Forty-five percent of those schizophrenic patients with initial negative symptoms at the first showed a remission of negative symptoms by the second. In contrast, only 5 percent of those schizophrenics without initial negative symptoms developed them by the second. This overall pattern of remission of negative symptoms for some schizophrenic patients, with few patients increasing in negative symptoms, 1st 1.26 (.68).97(1.78) 1.06(1.81) 2.29 (3.52) 2nd.16 (.45).68(1.28).55(1.21) 1.48(2.55) 31; 8 subjects missing negative symptom data at 2nd may differ at later phases of schizophrenia. This sample is still in the relatively early stages of disorder, and perhaps during later phases negative symptoms may become irreversible in more patients. Alternatively, it may be that negative symptoms tend to fluctuate in some schizophrenic patients and to be relatively irreversible in only a subgroup. Prognostic Importance of Positive Symptoms. Because of the contrasts that have been drawn at times between positive and negative symptoms, we also examined the prognostic importance of the positive psychotic symptoms of delusions and Table 7. Longitudinal course of the prevalence of negative symptoms Negative symptoms at 1st Absent Present Negative symptoms at 2nd Absent 19 (95%) 5 (45%) Present 1 (5%) 6 (55%) Note. Schizophrenic patients, n = 31. Dichotomized Overall Negative Symptom Index. Row percentages.

9 VOL 11, NO. 3, hallucinations. As was the case for negative symptoms, the presence at the first of the positive symptoms of hallucinations or delusions, using our overall index of positive symptoms, was significantly predictive of later overall poor functioning at the second (r =.60, p <.0001). Table 8 illustrates the specific nature of this prognostic relationship. The majority (78 percent) of schizophrenic patients who experienced definite hallucinations or delusions at the first showed poor functioning at the second. However, as was also the case for negative symptoms, a subgroup of schizophrenic patients (29 percent) who did not experience positive symptoms at the first nevertheless showed later poor functioning. Thus, although positive symptoms occurring at the first also have significant prognostic importance for later functioning, they also do not represent the only pathway to poor functioning during this phase of schizophrenia. Prognostic Relationships Between Positive and Negative Symptoms. To explore further the relationship between the prognostic importance of positive and negative symptoms, schizophrenic patients were divided into the following four groups depending on their joint positive/ negative symptom profile at the first : (1) mixed symptoms (both positive and negative symptoms); (2) pure negative symptoms (without definite positive symptoms); (3) pure positive symptoms (without definite negative symptoms); and (4) neither type of symptom (without definite positive or negative symptoms). Assignment to groups was based on the definite cutting scores for the overall positive symptom index and the overall negative symptom index described above. The means on the LKP scale of overall functioning at the second for the four groups were: (1) mixed symptoms, M = 7.50; (2) pure negative symptoms, M = 4.83; (3) pure positive symptoms, M = 5.85; and (4) neither symptom, M = Higher scores on the LKP scale indicate poorer functioning. The twoway (negative by positive symptoms) analysis of variance of these data found significant main effects for both negative (F = 3.96; df - 1, 35; p <.05) and positive symptoms (F ; df - 1, 35; p <.001) in predicting later outcome. However, Table 8. Prognostic relationship between overall positive symptoms at 1st and overall functioning at 2nd Overall functioning at 2nd (LKP scale) Overall positive symptoms Good Equivocal Poor at 1st (1-2) (3-6) (7-8) Absent/mild Definite 7 (33%) 1 (5%) 8 (38%) 3 (17%) Note. Schizophrenic patients, n = 39. Kendall's Tau C =.52, p <.002. Row percentages presented. 6 (29%) 14 (78%) the interaction of positive and negative symptoms in predicting later overall functioning was not significant (p <.76). These results suggest that the prognostic effects of positive and negative symptoms are additive and do not statistically interact with one another. As can be seen, those schizophrenic patients who experienced both positive and negative symptoms at the first (mixed symptoms group) were functioning more poorly at the second than those schizophrenic patients who had only negative or only positive symptoms at the first. Implications The present study is one of the few to investigate negative symptoms in relation to later outcome in schizophrenia using a prospective research design. The following major results emerged: Negative symptoms were more frequent at among schizophrenic patients who had poor educational achievement and poor social functioning before index hospitalization. Negative symptoms were also more common among schizophrenic patients from lower social class parental homes. Negative symptoms in schizophrenic patients were associated at with concurrent poor functioning in other areas (e.g., poorer instrumental work functioning and more rehospitalizations). Negative symptoms in schizophrenia during the postacute phase were prognostically important for later overall functioning and for functioning in the specific areas of instrumental work performance and rehospitalization. Negative symptoms tended to characterize schizophrenic patients

10 438 SCHIZOPHRENIA BULLETIN who were already functioning poorly, rather than to precede a later decline. Negative symptoms were persistent during the early course of the disorder in many schizophrenic patients, although in some others negative symptoms remitted. Negative symptoms represented only one of several possible pathways to later poor functioning in schizophrenia. The presence of negative symptoms in this early phase of schizophrenia did not imply the absence of positive symptoms. Negative and positive symptoms were independent phenomena, and actually appeared to be positively associated from a longitudinal perspective. Positive symptoms during the postacute phase of schizophrenia were also prognostically important for later poor functioning. The prognostic importance of negative and positive symptoms was additive during the early course of schizophrenia. The finding of the prognostic importance of negative symptoms for later functioning during the early posthospital period is consistent with theory concerning negative symptoms (Strauss, Carpenter, and Bartko 1974; Carpenter et al. 1978; Crow 1980). This result, in a sample of young schizophrenic patients, suggests that the occurrence of negative symptoms early in the course of the disorder may, in fact, delineate a subgroup of patients with a chronic course. Negative symptoms appeared to occur primarily in a subgroup of schizophrenic patients who had a history of poor functioning and who subsequently tended to remain impaired in their role performance. It is important to note that not all chronically poor functioning patients showed negative symptoms during this early phase of the disorder. The results also suggested a degree of overlap between negative symptoms and the traditional prognostic factors of poor "premorbid" functioning. The finding of a tendency toward remission from negative symptoms in some schizophrenic patients is counter to theory which hypothesizes negative symptoms to be persistent over time. It may be that at later stages of the disorder negative symptoms will be more persistent. Perhaps during the earlier stages of the disorder, negative symptoms in some vulnerable patients may represent signs that remit especially slowly following an acute episode. In general, these results suggest that negative symptoms appear to be important individual differences in schizophrenic symptomatology with longitudinal and cross-situational correlates. More research is needed to investigate the potential role that such negative symptoms might play in nonschizophrenic disorders. Preliminary data suggest that negative symptoms also are present in other psychotic disorders, such as schizoaffective disorder and perhaps to some extent in psychotic depression (Pogue-Geile and Harrow 1984b). In contrast, our results suggest a more complex picture concerning positive symptoms than is usually hypothesized. Negative symptoms were not the only pathway to later poor role functioning. Schizophrenic patients who experienced positive symptoms during the postacute phase also were characterized by later functioning impairment. This finding contrasts with some hypotheses, which present positive symptoms as fluctuating and as having less prognostic value within schizophrenia (e.g., Strauss, Carpenter, and Bartko 1974). However, most such hypotheses are concerned with the presence of positive psychotic symptoms during the acute inpatient phase and the current study investigated the prognostic importance of positive symptoms that were present at. The prognostic importance of positive symptoms within schizophrenia may differ between inpatient and posthospital phases. Our results on positive symptoms generally suggest that they are not opposed to negative symptoms. Negative and positive symptoms were cross-sectionally independent and both made independent and additive contributions to predicting subsequent outcome functioning. A number of other investigators have also reported this cross-sectional independence of negative and positive symptoms in schizophrenia (Johnstone et al. 1981; Lewine, Fogg, and Meltzer 1983; reanalysis of Opler et al. 1984; reanalysis of Lindenmayer, Kay, and Opler 1984; Pogue- Geile and Harrow 1984a; Rosen et al. 1984; but cf. Andreasen and Olsen 1982). The distinction between the two groups of symptoms seems to be a useful one, although the terminology of "positive" and "negative" may imply an opposition between the two that does not appear to exist. Similarly, the terms can also lend themselves to an overly simplified dichotomous view of psychopathological symptoms. Clearly, it is unlikely that the broad range of psychopathological phenomena should be grouped into only two categories. Empirical approaches investigating the cooccurrence of signs and symptoms should provide a useful guide in grouping symptoms together (e.g., Farmer, McGuffin, and Spitznagel 1983).

11 VOL. 11, NO. 3, Conjectures of Potental Etiologies of Negative Symptoms in Schizophrenia. While Crow (1980) has proposed several valuable hypotheses concerning the possible organic pathology of negative symptoms, less discussion has centered on the potential etiology of negative symptoms and their potential role in schizophrenia. Although there are few studies specifically relevant to this issue, the evidence to date appears to indicate that negative symptoms have enough important correlates that some attention to their etiology is warranted. Since severe negative symptoms are probably only relevant to a subgroup of schizophrenic patients, as they are currently diagnosed, we will outline three general etiological models that would be applicable to individual differences within schizophrenia such as negative symptoms (Pogue-Geile and Harrow, in press). First, in a modifying influence model, negative symptoms would arise as the result of an interaction between some independent trait, such as low intelligence, and a more specific schizophrenic (or psychotic) diathesis. In this case, the independent trait would not actually increase the risk for schizophrenia itself, but would only serve to modulate the phenotypic manifestation of schizophrenia. The etiology of the independent trait itself could be genetic and/or environmental. In the case of such a modifying influence model, negative symptoms, although identifying a subgroup of schizophrenics, would not generally be considered to have identified a subtype of schizophrenia with a specific etiology. A second etiological alternative is a contributing influence model in which negative symptoms represent the behavioral manifestation of some "negative symptom specific" factor that not only leads to negative symptoms, but also directly contributes to the risk of schizophrenia. This negative symptom specific factor could itself be due to genetic and/or environmental influences. For example, within the context of a polygenic theory of schizophrenia in which three of five "schizophrenia specific" genes (e.g., genes A, B, C, D, E) are necessary to produce clinical schizophrenia, then any schizophrenic patient who has the "negative symptoms" gene (e.g., gene "D") as one of his three genes leading to schizophrenia will show negative symptoms. In the extreme case where the "D" gene is sufficient to produce a schizophrenic picture, then negative symptom schizophrenia would represent a distinct type of schizophrenia with its own specific etiology. A similar example could be fashioned for a putative environmental factor, such as viral infection, that could both lead to negative symptoms and also contribute to the risk of schizophrenia. In the third, or threshold model, negative symptoms would represent a behavioral manifestation of a more severe threshold along a continuum of multifactorial liability to schizophrenia. In this case, negative symptoms would not be due to any specific influence, but would reflect only the accumulation of negative symptom nonspecific influences present for a given individual. Within the context of the polygenic example given above, where three schizophrenia specific genes were necessary to produce clinical schizophrenia, it could be that four of the five genes would be necessary to produce schizophrenia with negative symptoms (the negative symptom threshold). In this case, no one of the five genes would be specific for negative symptoms, but only the total number of genes present would be important. To date, few studies addressing these issues have been performed. The two most relevant studies, which have involved reanalyses of twin samples, appear generally most supportive of the view that negative symptoms may represent a severity threshold on a continuum of liability to schizophrenia, although much more research is necessary in this area (Dworkin and Lenzenweger 1984; Farmer, McGuffin, and Gottesman 1984). These three potential models for the etiology of negative symptoms and their role in schizophrenia do not exhaust all the possibilities. Although they are quite general, they can nevertheless provide a framework for investigating these issues and are specific enough to make different predictions in the context of family and twin studies. These models also have been discussed in reference to the traditional subtypes of schizophrenia (e.g., Gottesman and Shields 1982). Although much more needs to be understood concerning the behavioral and biological correlates of negative symptoms, studies of etiological questions may provide new insights into the role that negative symptoms might play in schizophrenia and other disorders. References American Psychiatric Association. DSM-lIl: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. Washington, DC: The Association, Andreasen, N.C. Negative symptoms in schizophrenia: Definition and reliability. Archives of General Psychiatry, 39: , Andreasen, N.C, and Olsen, S.

12 438 SCHIZOPHRENIA BULLETIN Negative v positive schizophrenia: Definition and validation. Archives of General Psychiatry, 39: , Andreasen, N.C.; Olsen, S.A.; Dennert, J.W.; and Smith, M.R. Ventricular enlargement in schizophrenia: Relationships to positive and negative schizophrenia. American Journal of Psychiatry, 139: , Angrist, B.; Rotrosen, J.; and Gershon, S. Differentia] effects of amphetamine and neuroleptics on negative v positive symptoms in schizophrenia. Psychopharmacology, 72:17-19, Carpenter, W.T., Jr.; Bartko, J.J.; Strauss, J.S.; and Hawk, A.B. Signs and symptoms as predictors of outcome: A report from the International Pilot Study of Schizophrenia. American Journal of Psychiatry, 135: , Carpenter, W.T., Jr.; Sacks, M.H.; Strauss, J.S.; Bartko, J.J.; and Rayner, J. Evaluating signs and symptoms: Comparisons of structured interview and clinical approaches. British Journal of Psychiatry, 128: , Crow, T.J. Molecular pathology of schizophrenia: More than one disease process7 British Medical Journal, 280:1-9, Dworkin, R.J., and Lenzenweger, M.F. Symptoms and the genetics of schizophrenia: Implications for diagnosis. American Journal of Psychiatry, 141: , Farmer, A.E.; McGuffin, P.; and Gottesman, I.I. Searching for the split in schizophrenia: A twin study perspective. Psychiatry Research, 13: , Farmer, A.E.; McGuffin, P.; and Spitznagel, E.L. Heterogeneity in schizophrenia: A cluster-analytic approach. Psychiatry Research, 8:1-12, Garmezy, N. Process and reactive schizophrenia: Some conceptions and issues. In: Katz, M.M.; Cole, J.O.; and Barton, W.E., eds. The Role and Methodology of Classification in Psychiatry and Psychopathology. Washington, DC: Superintendent of Documents, U.S. Government Printing Office, pp Gottesman, I.I., and Shields, J. Schizophrenia: The Epigenetic Puzzle. New York: Cambridge University Press, Grinker, R.R., and Holzman, P.S. Schizophrenic pathology in young adults. Archives of General Psychiatry, 28: , Harrow, M.; Grinker, R.R.; Silverstein, M.L.; and Holzman, P. Is modern-day schizophrenic outcome still negative? American Journal of Psychiatry, 135: , Harrow, M., and Quinlan, D. Disordered Thinking and Schizophrenic Psychopathology. New York: Gardner Press, Harrow, M., and Silverstein, M.L. Psychotic symptoms in schizophrenia after the acute phase. Schizophrenia Bulletin, 3: , Harrow, M.; Silverstein, M.; and Marengo, J. Disordered thinking: Does it identify nuclear schizophrenia? Archives of General Psychiatry, 40: , Hollingshead, A.B., and Redlich, F.C. Social Class and Mental Illness. New York: John Wiley A: Sons, Johnstone, E.C.; Crow, T.J.; Frith, CD.; Carney, M.W.P.; and Price, J.S. Mechanism of the antipsychotic effect in the treatment of acute schizophrenia. Lancet, 1: , Johnstone, E.C.; Crow, T.J.; Frith, CD.; Husband, J. ; and Kreel, L. Cerebral ventricular size and cognitive impairment in chronic schizophrenia. Lancet, 11: , Johnstone, E.C.; Crow, T.J.; Frith, CD.; Stevens, M.; Kreel, L; and Husband, J. The dementia of dementia praecox. Ada Psychiatrica Scandinavica, 57: , Johnstone, E.C.; Frith, CD.; Gold, A.; and Stevens, M. The outcome of severe acute schizophrenic illnesses after one year. British Journal of Psychiatry, 134:28-33, Johnstone, E.C; Owens, D.G.C; Gold, A.; Crow, T.J.; and MacMillan, J.F. Institutionalization and the defects of schizophrenia. British Journal of Psychiatry, 139: , Kencfler, K.S., and Davis, K.L. The genetics and biochemistry of paranoid schizophrenia and other paranoid psychoses. Schizophrenia Bulletin, 7: , Levenstein, S.; Klein, D.F.; and Pollack, M. Followup study of formerly hospitalized voluntary patients: The first two years. American Journal of Psychiatry, 10: , Lewine, R.R.J.; Fogg, L.; and Meltzer, H.Y. Assessment of negative and positive symptoms in schizophrenia. Schizophrenia Bulletin, 9: , Lewine, R.R.J., and Meltzer, H.Y. Negative symptoms and platelet monoamine oxidase activity in male schizophrenic patients. Psychiatry Research, 12:99-109, Lindenmayer, J.P.; Kay, S.R.; and Opler, L. Positive and negative subtypes in acute schizophrenia. Comprehensive Psychiatry, 25: , Luchins, D.J.; Lewine, R.R.J.; and Meltzer, H.Y. Lateral ventricular

13 VOL 11, NO. 3, size, psychopathology, and medication response in the psychoses. Biological Psychiatry, 19:29-44, Nasrallah, H.A.; Kuperman, S.; Hamra, B.J.; and McCalley-Whitters, M. Clinical differences between schizophrenic patients with and without large cerebral ventricles. Journal of Clinical Psychiatry, 44: , Opler, L.A.; Kay, S.R.; Rosado, V.; and Undenmayer, J. Positive and negative syndromes in chronic schizophrenic inpatients. Journal of Nervous and Mental Disease, 172: , Owens, D.G.C., and Johnstone, E.C. The disabilities of chronic schizophrenia Their nature and the factors contributing to their development. British Journal of Psychiatry, 136: , Pearlson, G.D.; Garbacz, D.J.; Breakey, W.R.; Ahn, H.S.; and DePaulo, J.R. Lateral ventricular enlargement associated with persistent unemployment and negative symptoms in both schizophrenia and bipolar disorder. Psychiatry Research, 12:1-9, Pogue-Geile, M.F., and Harrow, M. Negative and positive symptoms in schizophrenia and depression: A. Schizophrenia Bulletin, 10: , 1984a. Pogue-Geile, M.F., and Harrow, M. "Negative Symptoms in Schizophrenia and Other Psychoses." Presented at the Annual Meeting of the American Psychological Association, Toronto, Ontario, Canada, 1984b. Pogue-Geile, M.F., and Harrow, M. Negative symptoms, affective deficit states, and intellectual dysfunction in chronic schizophrenia. In: Clark, D., and Fawcett, ]., eds. Anhedonia and Affect Deficit States. New York: S.P. Medical ic Scientific Publications, in press. Rosen, W.G.; Mohs, R.C.; Johns, C.A.; Small, N.S.; Kendler, K.S.; Horvath, T.B.; and Davis, K.L. Positive and negative symptoms in schizophrenia. Psychiatry Research, 13: , Spitzer, R.L., and Endicott, J. Schedule for Affective Disorders and Schizophrenia (SADS). New York: New York State Psychiatric Institute, Spitzer, R.L.; Endicott, J.; and Robins, E. Research Diagnostic Criteria (RDC) for a Selected Group of Functional Disorders. 3rd ed. New York: New York State Psychiatric Institute, Stofflemayr, B.E.; Dillavaou, D.; and Hunter, J.E. Premorbid functioning and outcome in schizophrenia: A cumulative analysis. Journal of Consulting and Clinical Psychology, 51: , Strauss, J.S., and Carpenter, W.T., Jr. The prediction of outcome in schizophrenia: I. Characteristics of outcome. Archives of General Psychiatry, 27: , Strauss, J.S., and Carpenter, W.T., Jr. The prediction of outcome in schizophrenia: II. Relationships between predictors and outcome variables. Archives of General Psychiatry, 31:37-42, Strauss, J.S., and Carpenter, W.T., Jr. Prediction of outcome in schizophrenia: III. Five year outcome and its predictors. Archives of General Psychiatry, 34: , Strauss, J.S.; Carpenter, W.T., Jr.; and Bartko, J.J. Speculations on the processes that underlie schizophrenic symptoms and signs. Schizophrenia Bulletin, 1 (Experimental Issue No. ll):61-69, Zubin, J.; Magaziner, J.; and Steinhauer, S.R. The metamorphosis of schizophrenia: From chronidty to vulnerability. Psychological Medicine, 13: , Zubin, J., and Spring, B. Vulnerability A new view of schizophrenia. Journal of Abnormal Psychology, 86: , Zubin, J.; Sutton, S.; Salzinger, K.; Salzinger, S.; Burdock, E.I.; and Peretz, D. A biometric approach to prognosis in schizophrenia. In: Hoch, P.H., and Zubin, J., eds. Comparative Epidemiology in the Mental Disorders. New York: Grune & Stratton, pp Acknowledgment The research reported was supported, in part, by grant MH from the National Institute of Mental Health and grants from the John D. and Catherine T. MacArthur Foundation and the Carnegie Corporation. The Authors Michael F. Pogue-Geile, Ph.D., is Senior Psychologist, Michael Reese Hospital and Medical Center, and Research Associate, Assistant Professor, Department of Psychiatry, The University of Chicago. Martin Harrow, Ph.D., is Director of Psychology, Michael Reese Hospital and Medical Center, and Professor, Department of Psychiatry and Department of Behavioral Sciences, The University of Chicago, Chicago, IL. Dr. Pogue-Geile has recently accepted a position as Assistant Professor, Department of Psychology, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, PA.