Course of insight in manic episodeA Kumar1, S Kumar2, NM Khan3, S Mishra2
1 State Mental Health Institute, Selaqui, Dehradun, Uttarakhand, India
2 Institute of Mental Health and Hospital, Agra, Uttar Pradesh, India
3 Department of Psychiatry, Rama Medical College Hospital and Research Centre, Kanpur, Uttar Pradesh, India
Background: Insight is an important factor associated with non compliance and poor outcome. Poor level of insight has been described as a characteristic in patients with acute bipolar disorder with more unawareness in social consequences with increasing severity in manic episode. Aim: Main aim of study was to see the baseline and longitudinal relationship between dimensions of insight with improvement in psychopathology. Setting and Design: Forty four patients diagnosed with mania, were selected from an inpatient setting at Institute of Mental Health and Hospital, Agra with mean age of 31.07(±9.00) years. They were assessed at base line and were followed up weekly or psychopathology and insight. Materials and Methods: The Young's mania rating scale for psychopathology and insight was assessed on three dimensions of SUMD. Results: Twenty five patients eventually completed the study. There was a positive correlation with global insight and with psychopathology consistent in longitudinal follow-up (P<0.05), but not correlating for awareness for achieved effect of medication and social consequences. Linear regression showed a positive relationship at the first and second week of assessment of SUMD and YMRS scores (P=0.001; 0.019). Conclusion: Improvement in insight is graded in a manic episode as compared to psychopathology. There is slower improvement in awareness of social consequences of mental disorder. It means that improvement in psychopathology may not necessarily indicate remission and need further supervision to improve insight and hence monitoring.
Keywords: Insight, mania, psychopathology
Insight has been the focus of interest in the study of mood disorders. It is associated with noncompliance and poor outcome. ,, Recent studies have focused on insight in relation to quality of life, cognitive functions, psychotic symptoms, and number of episodes in mood disorders. ,,,, Although insight is clearly affected by mood-related psychopathology, mood symptoms and insight seems to relate to differential phenomenological domains.  It is severely impaired in euphoric mania in comparison to mixed episode, and in bipolar II as compared with bipolar I disorder. ,, In a meta-analysis, insight was found significantly compromised during the manic phase of the illness, and it was concluded that insight is a state dependent phenomenon in bipolar disorder. 
The study of insight and its correlates are important because awareness of illness may have diagnostic, nosological, and prognostic value, and can affect treatment and rehabilitation efforts.  A poor level of insight is described as a characteristic in patients with acute bipolar disorder. , Awareness seems to grow partially with clinical improvement and appears to diminish progressively with repetition of episodes. , Unawareness in social consequences was found greater with severity of manic symptoms but studies about insight in mood disorder have shown contradictory findings. ,,,
Previous studies were either cross-sectional, lacking appropriate case definitions or did not assess the dimensions of insight longitudinally. ,, Further, there is evidence suggesting that the phenomenology of depression and schizophrenia in Indian patients differs from those in the developed world. , Since the duration of poorer insight may impact on the burden on caregivers, it would be relevant to see prospectively the course of insight and hence the present study which aimed at investigating the relationship between improvement in insight and it is dimensions with the improvement in psychopathology.
The study protocol was approved by the Institutional Review Board. Consecutive patients in the age range of 18-45 years with manic episodes as per International Statistical Classification of Diseases and Related Health Problems (ICD-10) criteria were selected.  Those having major co-morbid psychiatric or physical illness, history of a head injury with loss of consciousness, history of seizures or other co-morbid substance dependence except nicotine and caffeine were excluded. Written informed consent was obtained from all patients. Consent for those with poor insight were taken from their legally acceptable representative. Written consent for the latter was signed by them at a later stage when they were able to understand their illness and study process. Those who declined consent were alos excluded. A total of 44 patients were taken and 25 eventually completed the study.
Severity of illness was assessed on the Young mania rating scale (YMRS) and for insight on the Scale to Assess Unawareness of Mental Disorder (SUMD) at the time of admission and then monitored at weekly intervals for a duration of 2 weeks. , YMRS also contains an insight item that was taken into consideration while calculating the total score. Patients received mood stabilizers and electroconvulsive therapy when indicated. Sodium valproate and lithium were used asmood stabilizers in combination with either Chlorpromazine, Trifluperazine, Haloperidol, Olanzapine, or Risperidone. Lorazepam was administered for agitation.
Data was analyzed with the help of Statistical Package for Social Sciences (SPSS) (version 10.0).  Descriptive analysis was for demographic variables (age, years of education, illness duration, duration of current episode, number of past episodes, and previous hospitalization), and psychopathology scores (YMRS and SUMD). Comparison of psychopathology scores with respect to variables (sex, marital status, residence, family-type, use of nicotine/caffeine, and positive family history of psychiatric illness) was analyzed by independent t-test. Serial comparison of YMRS and SUMD scores were analyzed by analysis of variance (ANOVA). Pearson's Correlation was computed to see the relation between YMRS and SUMD scores. Further, linear regression analysis was done. All analyses were done at 5% significance.
The mean age of subjects (primarily male, 68.2%) at the time of the study was 31.07 (±9.00) years. The mean age at onset of illness of subjects was 24.31 (±7.25) years with 5.48 (±4.94) years of education. Total duration of illness was 346.63 (±391.26) weeks and mean duration of current episode was 6.99 (±6.08) weeks. There were 3.43 (±2.50) previous episodes with 2.09 (±2.12) undergoing previous hospitalization. There was no significant difference between male and female subjects with respect to the above mentioned variables. There were 29 (65.9%) married, 14 (31.8) single, and one (2.3%) separated patient. Twenty-eight (63.6%) subjects belonged to rural background. Eighteen subjects (40.9%) belonged to a nuclear family and the rest (59.1%) to a non-nuclear family. Use of nicotine/caffeine was present in 17 (38.6%) subjects. Twenty-one (47.7%) subjects had a positive psychiatric history in their family. There was no significant difference in baseline score of YMRS or SUMD with respect to sex, marital status, residence, family-type, use of nicotine/caffeine, and positive family history of psychiatric illness.
When patients with a positive family history of psychiatric illness were analyzed, they were seen to more likely to have more number of episodes (P=0.001) and to have undergone previous hospitalization (P=0.009) after controlling for total duration of illness. The base line YMRS score was 32.80 (±6.56) and SUMD score was 12.86 (±2.46). The YMRS score after the second week showed a positive correlation (P<0.05) with education. There was gradual improvement in psychopathology as well as in insight both weeks. (P<0.001). The individual component of insight showed similar patterns except for awareness of achieved effect of medication which showed deterioration in the first week (P<0.001) and then improvement in the second week (P<0.001). [Table 1] and [Table 2] show the correlation between total YMRS score with individual components of insight and the linear regression analysis respectively.
[Table 1] shows the significant correlation (Pearson's r) between YMRS scores and total scores of SUMD. There is also positive correlation for awareness of mental disorder at the end of 1 st and 2 nd week; and with awareness of achieved effects of medication and awareness of social consequences of mental disorder at the end of 1 st week.
[Table 2] shows regression analysis to predict SUMD score from YMRS score. It shows that there is positive relationship at 1 st and 2 nd week of assessment of SUMD and total YMRS scores.
The present study prospectively assessed the course and progression of insight in a group of patients with mania, It also correlated this with age, education, and gender.
The mean age of the patients (31.07 years) in the present study was similar to a previous study.  The dominance of male patients in this study was consistent with findings in other Indian studies. ,, However, contrary findings in earlier studies may be due to socio-cultural differences. , A larger proportion of married subjects were comparable again to previous findings. , This was also supported by the demographic profile in India where 58% of men are seen to be married by the age of 25, the proportion being higher in women (74% married by 20 years of age). 
The rural proportion in the study was again similar to the Indian rural population 72.18% conforming to norms of general population.  There is a growing trend towards nuclear families in India. These families constitute a married couple or a man or a woman living alone or with unmarried children (biological, adopted, or fostered), with or without unrelated individuals. , There is under-representation of non-nuclear family in the study population (40.9%) relative to the national data (60.5%).  Good family support of patients belonging to non-nuclear family may lead to lesser number of admissions as in-patient explains this finding.
The mean age of onset of illness in the study was similar to other studies done in India. ,,, Age of onset was lower than earlier studies. A potential reason could be restriction of age limit to 45 years. 
The mean episode duration in the study sample was lower than in previous studies conducted developing countries. , Treatment of previous episodes might have increased awareness about the illness compelling caregivers to bring the patient early in the course of subsequent episodes. The mean number of past episodes of 3.34 may be equated with 4.4 in other studies which took the total number of episodes rather than just the past one. ,
Despite alcohol being the commonest substance used (82.5%), followed by cannabis (16.1%), there was a very small number of patient with substance use.  This is likely due to the exclusion of substance dependence other than nicotine/caffeine. Use of nicotine or caffeine only may be less in manic patients and most of them might have been using other substances like alcohol or cannabis which were excluded from the study.
Study findings consistently show genetic vulnerability of bipolar disorder. , Positive family history (47.7%) in the study is comparable to the figures in literature. , Presence of family history leads to a recurrence of greater number of episodes (P=0.001) and hospitalizations (P=0.009), which was consistent with previous finding. 
Results of the scores showed gradual improvement in psychopathology and insight both globally as well as in its dimensions. However, there was a paradoxical deterioration in awareness to achieved effect of medication which may be explained by the construction of the scale itself. The higher end of score in SUMD reflects poorer insight. However, zero (0) is scored when the item is not applicable. At the time of admission most of the subjects were not receiving any medication, thus scored zero. With improvement in subsequent weeks their score increased from 1 to 15. Thus a higher score was reflected when compared to previous rating.
The findings did not show any significant difference in baseline psychopathology (severity of manic episode and insight) with any of demographic and clinical variables. This indicates that there was no effect of mentioned variables on the psychopathology at the time of hospitalization. Positive correlation of education with improvement in psychopathology indicates a good prognosis with education. This is only a preliminary finding, which needs to be confirmed.
Findings of positive correlation between the psychopathology and insight in longitudinal follow-up were well supported by the previous meta-analysis.  However, individual subscales of insight did not correlate with psychopathology. It is due to the fact that the ceiling score of insight subscales in SUMD would be 15 irrespective of increasing severity in psychopathology. This explains the findings of the regression analysis when impairment in global insight can be predicted from psychopathology with symptoms improvement (at the end of 1 st and 2 nd week) but not at baseline, when psychopathology was severe. It was well correlated at the end of 2 nd week when severity lessened. Positive correlation with awareness of effect of medication may again be paradoxical because of reason mentioned in the earlier part of the discussion. However, after 2 weeks, there was awareness of mental illness but not to the effect of medication or social consequences. [Table 2] shows that deterioration was more in the awareness of social consequences of mental disorder as compared to the other two dimensions. The finding is supported by previous research where partial improvement in insight was present when psychopathology improved. , Lack of awareness for social consequences with psychopathology was supported by recent findings. 
Thus, it may be concluded that improvement in insight is graded in manic episode with the improvement of psychopathology; awareness of social consequences of mental disorder being slower. This means that improvement in psychopathology may not necessarily indicate remission and it may take longer to improve insight and hence compliance. One of the limitations of the study is that the same rater rated the insight as well as psychopathology thus rater's bias could not be excluded. Another limitation is the period of observation which is short, which can be overcome in future studies. Finally, instead of a descriptive approach, more analytical perspective may be considered.
I owe my sincere thanks to N. Manjunatha, National Institute of Mental Health and Neurosciences, Bangalore, India who assisted for preparation and proof-reading of the manuscript.
[Table 1], [Table 2]