Sociodemographic and clinical factors associated with relapse in schizophrenia

Authors


Ajit Avasthi, MD, Department of Psychiatry, Postgraduate Institute of Medical Education and Research, Chandigarh-160012, India. Email: pratapsharan@yahoo.com

Abstract

Abstract  The aim of the present study was to examine sociodemographic and clinical factors associated with relapse in schizophrenia. The study group consisted of a convenience sample of 40 schizophrenia patients (20 patients each in relapse and remission). Relapse and remission were defined based on clinical criteria (ICD-10 criteria, course since last episode, and duration of remission) and psychometric criteria (scores on Socio-Occupational Functioning Assessment Scale [SOFAS] and Positive and Negative Syndrome Scale for Schizophrenia [PANSS]). The index group was evaluated after the occurrence of current relapse but within 6 months of its onset. Sociodemographic, current psychopathology (PANSS) and functioning (SOFAS), and other (mainly retrospective) variables were assessed with a specifically designed clinical profile sheet, Schedule for Affective Disorders and Schizophrenia Lifetime version, Presumptive Stressful life Events Scale, and World Health Organization Life Chart Schedule for Assessment of Course and Outcome of Schizophrenia. Patients who had relapsed were more symptomatic and exhibited greater dysfunction in comparison to remitted patients. Relapse in schizophrenia was significantly associated with unemployment, number of psychotic episodes, side-effects of medication, and life events score. The present findings suggest that a severe illness (no. psychotic episodes, unemployment), psychological stress and inappropriate treatment (side-effects of medicines) may be causally related to relapse in schizophrenia. However, the possibility that these variables may be caused by relapse or may be explained by a common underlying variable needs to be assessed prospectively.

INTRODUCTION

Relapses in schizophrenia predict poor prognosis, bring about deterioration in social, occupational and financial status and increase the burden of care on the family. Relapse has been defined variably by researchers, for example clinical deterioration of such magnitude that hospitalization was imminent;1 a return to medication;2 a return of the disorder with a re-emergence of florid psychotic features and gross social decomposition leading to a change in treatment and hospitalization.3 This has contributed to inconsistency in findings.

Mortensen and Eaton found that readmission risk decreased with increasing age at first schizophrenia admission till the fifth discharge but no effect was found at later admissions.4 Moller et al. reported that age >40 years reduced the risk of re-hospitalization, and age >30 years reduced the rate of relapse in patients on maintenance treatment.5 Angermeyer et al. reported that in one-third of 18 comparison studies on schizophrenia, women had lower rates of rehospitalization than men,6 whereas only two studies found the opposite result.4,7 Some studies have shown that living in a partnership predicted a better outcome.5,8 However, in many other studies marital status did not predict readmission risk.9–11

Readmission risk was lower after the first discharge for the simple, paranoid, and latent subtypes than for other groups.4 Diagnosis of residual type (ICD-9) decreased the risk of relapse.7 Comorbid substance use disorder was associated with early readmission.12 Subjective feeling of depression during the first admission, number of depressive episodes, depression developing within 1 year of recovery from a schizophrenia episode, and suicide attempts were associated with either early relapse or readmission during follow up in various studies.7,13–15 However, the presence of depressive delusions during a schizophrenia episode seemed to protect against an early relapse.13

Patients in families with high expressed emotion are more likely to relapse and to relapse more often,16,17 but expressed emotion, in all likelihood, is a weak predictor of relapse.18–20 Similarly, while some studies have shown a significant association between number of independent life events and schizophrenia relapse,21,22 other studies have found no such association.23,24 Ventura et al. found life events to be significant predictors of relapse in patients on medication but not in patients off medication.25

Medication discontinuation is a risk factor for schizophrenia relapse.26,27 However, Prien et al. reported that only 6% of patients relapsed on high doses (2000 mg/day) and 13% on low doses (300 mg/day) of chlorpromazine.27 Other treatment variables may also be important. Early readmission was associated with four or more previous hospitalizations and absence of a family meeting with inpatient staff.12 Functional status before onset of schizophrenia (poor school adaptation in childhood and late adolescence) and after the onset of schizophrenia (period of unemployment) has also been observed to be associated with relapse and readmission.5,28–30

Identification of factors associated with relapse is important to develop modalities for relapse prevention. The present study was conducted to examine sociodemographic and clinical factors associated with rigorously defined relapse in schizophrenia.

METHODS

The sample for the present study was taken from psychiatric outpatient and inpatient services of the Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, a tertiary care center located in north-western India. The index group consisted of a convenience sample of 20 patients with schizophrenia presenting in relapse, and the control group consisted of 20 patients with schizophrenia presenting in remission.

The Institute Ethics Committee approved the protocol for the research project and the work was undertaken within its guidelines, and it conforms to the provisions of the Declaration of Helsinki in 1995 (as revised in Edinburgh 2000). All subjects or their legal guardian (in the case that the subject was incompetent to give consent) gave written informed consent and patient anonymity was preserved.

Relapse was defined as (i) presence of ≥1 criterion from the A–D group or presence of ≥2 criteria from the E–H group of ICD-10 criteria for schizophrenia lasting for at least 1 week;31 (ii) Socio-Occupational Functioning Assessment Scale (SOFAS) level of ≤70;32 (iii) Positive and Negative Syndrome Scale for Schizophrenia (PANSS) positive subscale score ≥13;33 (iv) following a period of remission of at least 2 months in which there was absence of any A–D criteria and ≥2 E–H criteria for schizophrenia in ICD-10; (v) with the current relapse not being the first expression of schizophrenia psychosis. Remission was defined as (i) absence of A–D criteria and absence of ≥2 E–H criteria for schizophrenia in ICD-10; (ii) SOFAS >70; (iii) PANSS positive subscale score ≤12; (iv) no change in medication by >50% and no hospitalization; and (v) duration of at least 2 months. All the five criteria had to be fulfilled for the identification of relapse or remission.

Patients in both groups were of either sex between the ages of 18 and 55 years. Patients with clinical evidence of concomitant major psychiatric illness, substance abuse (excluding tobacco), organic brain syndrome, mental retardation, major physical illness or incapacitation were excluded from the study. Patients in the index group were evaluated after the onset of current relapse but within 6 months of its onset. Information was collected on the following classes of variable: (i) sociodemographic variables; (ii) current psychopathology (PANSS) and functioning (SOFAS); and (iii) other (mainly retrospective) variables with the help of a specifically designed clinical profile sheet, and also Schedule for Affective Disorders and Schizophrenia Lifetime version (SADS-L),34 Presumptive Stressful life Events Scale (PSLES),35 and World Health Organization (WHO) Life Chart Schedule for Assessment of Course and Outcome of Schizophrenia (WHO-LCS).36

The PANSS consists of three subsets of items, that is, for positive symptoms, negative symptoms and general psychopathology.32 SOFAS focuses exclusively on the individual's level of social and occupational functioning, and is not directly influenced by the overall severity of the individual's psychological symptoms.33 The clinical profile sheet specifically designed for the present study was used to record the following variables: age at onset, duration of illness, family history of schizophrenia, subcategory of schizophrenia, and medication (status, type, dose, side-effects). SADS-L generates diagnoses of major psychiatric disorders based on information relevant to diagnosis, prognosis and phenomenology of current and past episodes of illness.34 PSLES is an inventory of 51 items from which scores on desirable, undesirable and ambiguous and also ‘personal’ and ‘impersonal’ categories can be derived.35 Normative data from north-western India are available for comparisons. WHO-LCS provides information about specific areas such as work, residence, symptoms and treatment since first treatment contact. It is a valid and reliable test for assessment of course and outcome of patients with schizophrenia.36

Unpaired Student's t-test was applied for group comparisons on continuous variables. Fisher's exact test was applied for categorical variables. Significance was set at P < 0.05 (two-tailed test).

RESULTS

Demographic variables

The majority of the patients in the relapsed and remitted groups of schizophrenia, respectively, were male (>60%), literate (>80%), employed (>50%), Hindu (>70%); and from an urban background (>65%). Only a minority were married (<45%). Monthly income in the majority (>70%) was below Rupees 3000 (approximately US$60). The mean age of the patients in the relapsed and remitted schizophrenia groups was 33.8 ± 10.2 years and 33.85 ± 11.02 years, respectively. The two groups did not differ significantly on sociodemographic variables (Table 1).

Table 1.  Subject characteristics
VariableRelapsed group (n = 20)
n (%)
Remitted group (n = 20)
n (%)
Fisher's exact
P
Gender  0.74
 Male14 (70)12 (60) 
 Female6 (30)8 (40) 
Marital status  0.32
 Single15 (75)11 (55) 
 Married5 (25)9 (45) 
Occupation  1.00
 Employed11 (55)10 (50) 
 Unemployed9 (45)10 (50) 
Education  0.34
 Illiterate4 (20)1 (5) 
 Literate16 (80)19 (95) 
Income (Rupees/month)  1.00
 ≤300014 (70)15 (75) 
 >30006 (30)5 (25) 
Religion  0.45
 Hinduism14 (70)17 (85) 
 Others6 (30)3 (15) 
Locality  0.27
 Urban17 (85)13 (65) 
 Rural3 (15)7 (35) 

Current psychopathology and functioning

The relapsed group had significantly higher positive syndrome; general psychopathology; negative syndrome, anergic and paranoid feature scores in comparison to the remitted group (Table 2). The majority (55%) of patients in the relapsed group had moderate impairment in socio-occupational function, while the majority in the remitted group had functioning in the superior to mild impairment range. The difference between the two groups was statistically significant (Table 2).

Table 2.  Psychopathological dimensions and stressful life events
VariablesRelapsed group
(n = 20)
Mean ± SD
Remitted group
(n = 20)
Mean ± SD
Statistic
  • P < 0.05,

  • ** 

    < 0.01,

  • *** 

    P < 0.001.

  • † 

    t (d.f. = 38);

  • ‡ 

    χ2 (d.f. = 2).

  • PANSS, Positive and Negative Syndrome Scale; SOFAS, Socio-Occupational Functioning Assessment Scale.

PANSS score
 Positive syndrome19.50 ± 5.35 8.85 ± 2.068.31***
 General psychopathology31.70 ± 10.824.90 ± 5.732.48*
 Negative syndrome20.15 ± 6.9214.35 ± 4.033.24*
 Anergia10.40 ± 2.91 8.40 ± 2.462.34*
 Thought disturbance 9.30 ± 3.54 6.70 ± 8.881.21
 Activation 5.10 ± 2.85 3.90 ± 0.971.85
 Paranoid 8.05 ± 3.61 3.95 ± 1.154.84***
 Depression 6.65 ± 1.95 6.30 ± 2.410.52
SOFAS categoriesn (%)n (%) 
 Superior functioning to mild impairment0 (0)17 (85)30.57***
 Moderate impairment11 (55)3 (15) 
 Severe impairment to incapacitation9 (45)0 (0) 

Other (mainly retrospective) variables

There was no significant difference between the relapsed and remitted groups, in terms of presence or absence of psychopathology elicited on the SADS-L: persecutory delusions (95%, each), somatic, grandiose, religious or other delusions without persecutory or jealous content, lasting at least 1 week (35% and 30%, respectively), delusions of being controlled (or influenced), multiple delusions, or other bizarre delusions (20% and 30%, respectively), delusions of any type accompanied by hallucinations of any type lasting for at least 1 week (25% and 20%, respectively), hallucinations of any type throughout the day for several days, or intermittently for at least 1 month (80% and 65%, respectively), visual hallucinations (25% and 15%, respectively), auditory hallucinations (85% and 75%, respectively), auditory hallucinations commenting/discussing (70% and 75%, respectively), non-affective verbal hallucination spoken to the subject (30% and 5%, respectively), thought broadcasting/withdrawal/insertion (20% and 50%, respectively), definite instances of formal thought disorder (10%, each), catatonic motor behavior (10% and 15%, respectively) and grossly bizarre behavior (30% and 10%, respectively; data not shown).

The interview, assisted by the specifically designed clinical profile sheet, showed that significantly more patients in the relapsed group had moderate to severe side-effects in comparison to remitted patients (Table 3).

Table 3.  Clinical subject features
VariableRelapsed group
(n = 20)
n (%)
Remitted group
(n = 20)
n (%)
Statistic
  • < 0.05.

  • † 

    t (d.f. = 38);

  • ‡ 

    Fisher's exact test (P).

  • § 

    n = 16 in relapsed group.

  • CPZ, chlorpromazine.

Age at onset (years), mean ± SD24.50 ± 7.6226.05 ± 6.950.671
Duration of the illness (years), mean ± SD10.15 ± 6.87 7.85 ± 7.880.98
Family history of schizophrenia  0.06
 Present8 (40)2 (10) 
 Absent12 (60)18 (90) 
Subtype of schizophrenia  0.54
 Paranoid11 (55)13 (65) 
 Others9 (45)7 (35) 
Medication  0.11
 Receiving16 (80)20 (100) 
 Not receiving4 (20)0 (0) 
Type of medication§  0.73
 Antipsychotic only9 (56.25)13 (65) 
 Antipsychotic + other drugs7 (43.75)7 (35) 
Dosage (CPZ equivalent) mg/day§  0.30
 ≤3007 (43.75)5 (25) 
 >3009 (56.25)15 (75) 
Compliance§  0.15
 Good10 (62.5)17 (85) 
 Poor6 (37.5)3 (15) 
Side-effects§  0.03*
 Minimal10 (62.5)19 (95) 
 Moderate to severe6 (37.5)1 (5) 

The relapsed group also reported significantly higher stress scores than the remitted patient group on PSLES (Table 4).

Table 4.  PSLES scores
PSLES scoreRelapsed group
(n = 20)
Mean ± SD
Remitted group
(n = 20)
Mean ± SD
t (d.f. = 38)
  • P < 0.05.

  • PSLES, Presumptive Stressful life Events Scale.

Undesirable items score27 ± 46.836.70 ± 20.621.77
Desirable items score7.25 ± 17.954.40 ± 13.630.56
Ambiguous items score39.55 ± 41.2421.95 ± 26.741.60
Total score73.00 ± 56.3929.65 ± 33.332.95*

The relapsed group had significantly more unemployment due to mental illness and higher mean number of psychotic episodes in comparison to the remitted group (Table 5).

Table 5.  Life course variables
VariableRelapsed group
(n = 20)
n (%)
Remitted group
(n = 20)
n (%)
Statistic
  • < 0.05.

  • † 

    t (d.f. = 38);

  • ‡ 

    Fisher's exact test (P).

  • § 

    n = 8 in relapsed group, n = 4 in remitted group;

  • ¶ 

    n = 10 in relapsed group, n = 6 in remitted group.

Psychotic episodes (mean ± SD)4.42 ± 1.242.92 ± 1.322.90*
Suicide attempts  1.00
 Yes17 (85)16 (80) 
 No3 (15)4 (20) 
Months in hospital for psychiatric problems§ (mean ± SD)3.10 ± 3.032.17 ± 1.170.70
Hospital admission for psychiatric problem  1.00
 <49 (90)6 (100) 
 ≥41 (10)0 (0) 
Unemployed  0.30
 Sometimes8 (40)4 (20) 
 Never12 (60)16 (80) 
Unemployment related to mental illness§  0.02*
 Not at all/a little0 (0)3 (75) 
 Some/a lot8 (100)1 (25) 
Social relations (5 years)  1.00
 Fair to superior16 (80)17 (85) 
 Poor/grossly inadequate4 (20)3 (15) 
Level of functioning (5 years)  0.14
 Absent/minimal/mild18 (90)14 (70) 
 ≥Moderate impairment2 (10)6 (30) 

DISCUSSION

The relapsed and remitted patient groups did not differ significantly in terms of sociodemographic variables. The occurrence of significant difference between relapsed and remitted groups in positive syndrome score and current level of functioning was mandated by our definition of relapse. The higher general psychopathology, negative syndrome, anergic and paranoid feature scores on PANSS in the relapsed patient group as compared to the remitted patient group, although not mandated by our definition of relapse, attest only to the likelihood that the symptom cluster of schizophrenia is of a syndromal nature.

In the present study, subtype of schizophrenia did not differ between the relapsed and remitted groups, unlike in some previous studies.4,7 The small sample size in the present study and the risk of type II error, limit the confidence that can be placed on negative statistical findings.

The significant difference in the relapsed and remitted groups in PANSS scores and the absence of such differences in psychopathology in SADS lifetime interview could be due to variation in period of assessment (cross-sectional vs lifetime), variation in rating methods (dimensional vs categorical) and possibility of recall bias (accurate assessment vs retrospective falsification).

The relapsed group had significantly greater number of psychotic episodes in comparison to the remitted group. However, the present study failed to find any association between the number of previous hospitalizations (>4) and relapse, unlike some previous studies.7,12 The apparent paradox becomes less puzzling when we consider the fact that most psychotic relapses are treated on an outpatient basis in India, unlike in the West, where a closer correspondence between relapse and admission would be seen.

Relapsed patients reported significantly more moderate–severe side-effects of medication as compared to patients in the remitted group. Presence of greater side-effects may be the result of increase in dose of antipsychotics to combat relapse, or the cause of non-compliance that may have led to the relapse. Earlier studies have reported discontinuation of medication and ultra low dosage of antipsychotics to be significant predictors of relapse in patients with schizophrenia.26,37,38 However, there was no significant difference between the two groups in terms of medication intake during the 6 months prior to evaluation in the present study.

The occurrence of significantly higher number of life events during the previous 6 months in the relapsed group in comparison to the remission group was consistent with previous reports.21,22,25

A longer period of unemployment was reported to be a reliable predictor of re-hospitalization rate in some studies.5,29,30 Significantly more relapsed patients had become unemployed due to their mental illness in comparison to the remitted group. It is likely that (repeated) relapse(s) had interfered with their occupational functioning. However, it is also possible that the relapsed group may be retrospectively magnifying their occupational dysfunction.

The recruitment of a small sample resulted in difficulties in interpreting the findings (type II errors). Use of a convenience sample could have resulted in skewed data.

The present findings suggest that a severe illness (number of psychotic episodes, unemployment), psychological stress and inappropriate treatment (side-effects of medicines) may be causally related to relapse in schizophrenia. However, prospective studies are needed to establish the causal independence of these variables, and multivariate techniques may lead to the identification of common underlying variables that may explain the effect of many predictor variables.

Ancillary