IN RECENT DECADES we have witnessed the emergence of a gradual consensus over the presence of cognitive deficits in the active phase as well as during remission of bipolar disorder (BD).[1, 2] Efforts are being directed to ascertain the type and extent of these cognitive deficits. Remitted patients with BD have been shown to have deficits in attention, processing speed, memory and executive functions.[2, 3] The deficits in executive functions and verbal memory appear to be especially marked. At this point, it is unclear whether the observed cognitive deficits reflect a single core impairment or multiple cognitive deficits.
Of particular importance is to know the clinical and functional significance of these cognitive impairments for clinicians and patients. The clinical relevance cannot be fully appreciated without assessing their impact on the day-to-day functioning and quality of the patient's life. It has been observed previously that these deficits could be negatively related to functional outcomes of the disorder.[4, 5]
The concept of quality of life (QoL) has become increasingly relevant in chronic, recurring disorders like BD. The QoL of BD patients has been found to be better than that of schizophrenia patients, but poorer compared to controls; however the results have been mixed.[6, 7] Many studies have shown that there is a definite, unexplained gap in the clinical and functional recovery of some patients with BD.[4, 5, 8] Although residual mood symptoms and a host of other factors have been studied as possible explanations, the role and contribution of cognitive impairments towards suboptimal QoL have not received much attention.
The present study was planned for two reasons. First, we wanted to test the hypothesis that poor cognitive functioning is likely to predict a poor QoL in euthymic BD patients. While cognitive functions and QoL have been independently demonstrated to be poor in several studies on BD, the nature and strength of the correlation between these factors merits a separate investigation. Second, most of the available studies have been conducted in Western settings and the current volume of work is not representative of a large part of the developing world, especially the Indian subcontinent.
Therefore, the present study was planned in order to: (i) assess the neuropsychological performance in euthymic patients and healthy controls from India; (ii) assess QoL and global functioning in patients and controls; and (iii) explore the correlation, if any, among cognitive impairments, QoL and global functioning of patients to support our hypothesis.
- Top of page
There was no significant difference in the sociodemographic profile (age, sex distribution, education, verbal IQ, HMSE) of patients and controls as shown in Table 1. The illness characteristics of the patient group are also shown in Table 1, along with the nature of current medication(s). The majority of patients were of middle socioeconomic status (76.7%), while the rest were of lower socioeconomic status. Nearly 83.3% were currently employed or working productively at home.
Table 1. Sample characteristics: patients and controls
|Mean ± SD||Mean ± SD|
|Age||33.53 ± 10.31||33.25 ± 10.41||0.095||0.93|
|Sex|| || || || |
|Male||63.3% (19)||60% (12)||γ (48)= 0.876||0.35|
|Female||36.7% (11)||40% (8)|
|Education||11.87 ± 2.93||13.05 ± 3.79||−1.243||0.22|
|Verbal IQ||102.37 ± 13.53||107.20 ± 10.72||−1.340||0.19|
|HMSE||27.90 ± 1.90||28.60 ± 1.23||−1.560||0.13|
|Duration of illness (years)||10.27 ± 6.44; Median 9.0 (2–28)|
|Number of episodes|| |
|Total||6.23 ± 5.33; Median 4.0 (2–20)|
|Depressive||1.70 ± 2.81; Median: 0.5 (0–12)|
|Manic||4.57 ± 4.35; Median: 3.0 (1–20)|
|Euthymic period prior to inclusion (months)||16.90 ± 16.23 Median: 12 (2–60)|
|Family history|| |
|Lithium||21 (70%), 1085.71 ± 211.63 mg/day|
|Valproate||11 (36.7%), 1204.55 ± 218.47 mg/day|
|Lamotrigine||2 (6.7%), 125.00 ± 35.36 mg/day|
|Olanzapine||5 (17.7%), 4.50 ± 2.09 mg/day|
Table 2 shows the performance of patient and control group on the tests of attention (Digit span, TMT-A), psychomotor processing speed (TMT-A, Mental Balance subtest, SCWT word and color cards), executive functions (TMT-B, SCWT) and working memory (N-Back). As seen from the table, there is a significant difference in the TMT-A and TMT-B tasks, with patients performing poorly compared to controls. On average, the patients took more time to complete both the tasks as compared to healthy controls (P < 0.01). The psychomotor processing speed was found to be slow among patients as evident from time taken to read the Stroop color card (P < 0.01), with moderate effect size (0.439). Table 3 shows the results of memory test in patients and controls, in which the patients performed poorly on subtests for delayed recall and verbal retention for similar pairs.
Table 2. Neuropsychological performance in patients and controls: Test for attention, psychomotor processing speed, executive functions and working memory
| ||Patients||Controls||Z||P||Effect size|
|Mean ± SD||Mean ± SD|
|Digit span (PGI memory subtest )|| || || || || |
|Digit forward + digit backward||9.30 ± 1.78||9.15 ± 1.31||−0.122||0.903||0.017|
|Mental balance (PGI memory subtest)|| || || || || |
|Psychomotor speed||7.43 ± 1.43||7.75 ± 1.25||−0.669||0.504||0.095|
|TMT|| || || || || |
|TMT-A, time taken (s)||62.40 ± 30.90||47.70 ± 27.65||−2.656||0.008**||0.376|
|TMT-B, time taken (s)||173.63 ± 99.20||120.70 ± 65.12||−2.777||0.005**||0.393|
|TMT B-A||111.23 ± 82.63||73.00 ± 59.97||−2.289||0.022*||0.324|
|Stroop Color and Word Test|| || || || || |
|Word card, time taken||60.60 ± 15.84||53.55 ± 11.52||−1.842||0.066||0.260|
|Color card, time taken||120.73 ± 34.11||91.40 ± 29.77||−3.108||0.002**||0.439|
|Interference score||149.93 ± 141.06||143.35 ± 80.25||−0.535||0.593||0.076|
|Verbal Working Memory N-Back Test|| || || || || |
|1-Back hits||8.00 ± 1.17||8.45 ± 0.83||−1.404||0.160||0.199|
|1-Back errors||1.23 ± 1.45||1.10 ± 1.89||−1.048||0.295||0.148|
|2-Back hits||5.70 ± 1.56||6.15 ± 1.95||−0.699||0.485||0.098|
|2-Back errors||3.87 ± 1.69||3.00 ± 1.89||−1.422||0.155||0.201|
Table 3. Neuropsychological performance in patients and controls: Tests for memory
|PGI Memory scale subtests||Patients||Controls||Z||P||Effect size|
|Mean ± SD||Mean ± SD|
|Delayed recall||8.30 ± 1.62||9.25 ± 1.16||−2.400||0.016*||0.339|
|Immediate recall||10.20 ± 1.19||10.85 ± 1.14||−1.845||0.065||0.261|
|Verbal retention for similar word pairs||4.53 ± 0.68||4.90 ± 0.31||2.146||0.032*||0.479|
|Verbal retention for dissimilar word pairs||11.90 ± 3.34||11.85 ± 3.48||−0.130||0.896||0.018|
|Visual retention||10.37 ± 2.79||11.00 ± 2.25||−1.668||0.504||0.236|
|Recognition||8.90 ± 2.09||9.75 ± 0.55||−1.773||0.095||0.251|
Table 4 shows the group differences in QoL and global functioning. Patients were found to have a significantly lower QoL in psychological and social domains. The difference in global functioning was highly significant (P < 0.01), with patients having poorer scores compared to controls.
Table 4. Quality of life and global functioning in patients and controls
| ||Patients||Controls||t||P||Effect size|
|Mean ± SD||Mean ± SD|
|World Health Organization Quality of Life domains|| || || || || |
|Physical||12.71 ± 2.91||14.50 ± 1.96||−1.718||0.090||0.496|
|Psychological||12.62 ± 2.68||14.33 ± 2.52||−2.404||0.020*||0.694|
|Social||18.93 ± 4.03||20.70 ± 4.01||−2.263||0.028*||0.653|
|Environmental||13.15 ± 1.97||14.10 ± 1.60||−1.789||0.080||0.516|
|Overall quality||3.20 ± 0.76||3.40 ± 0.59||−0.988||0.328||0.285|
|Overall health||3.17 ± 0.83||3.40 ± 0.75||−1.006||0.319||0.290|
|GAF|| || || || || |
|GAF score||71.53 ± 10.42||89.10 ± 3.68||Z = −5.454||<0.001**||0.771|
The correlation of cognitive variables of patients to the continuous clinical variables (total duration of illness and number of episodes) was examined using Spearman's correlation, the results of which are shown in Table 5. The correlation among the categorical clinical variables (family history, type of medication, age of onset, duration of euthymic period) was also examined using the Mann–Whitney U-test, none of which were found to be significant (not shown in table). The correlation among various cognitive variables, QoL domains and global functioning was examined and, as shown in Table 5, a moderate correlation was found with psychological and social domains as well as GAF score.
Table 5. Neuropsychological performance: Association among clinical variables, quality of life and global functioning in the patients (n = 30)
|rs (p)||Illness duration||No. of episodes||Physical domain||Psychological domain||Social domain||Environmental domain||GAF|
|Word card, time taken (s)||0.11||0.11||−0.21||0.01||−0.04||−0.30*||−0.25|
|Color card, time taken (s)||−0.02||0.21||−0.23||0.00||−0.35||−0.51**||−0.23|
|TMT-A, time taken (s)||−0.01||0.07||−0.22||−0.22||−0.39*||−0.49**||−0.21|
|TMT-B, time taken (s)||0.43*||0.41*||−0.36*||−0.46*||−0.63**||−0.57**||−0.14|
|Attention and concentration||0.02||0.05||0.09||−0.03||−0.08||−0.05||0.33|
|Verbal retention for similar pairs||0.07||0.09||0.10||−0.01||−0.07||0.27||0.14|
|Verbal retention for dissimilar pairs||0.05||0.15||0.25||0.08||0.03||0.23||0.23|
The stepwise multiple regression analysis was used to test for significant predictors for QoL domains. Only the cognitive variables with significant correlation were taken as independent variables. There were two independent variables (Visual Retention and TMT-B) each for physical and psychological domain, three (TMT-A, TMT-B, SCWT-CI) for social domain and six (TMT-A, TMT-B, SCWT-CI, word card, color card, 2-Back errors) for environmental domain. The detailed results for each domain are shown in Table 6. Visual retention accounted for 23% of variance in physical domain. TMT-B, which is indicative of cognitive flexibility and set shifting (executive functions), explained 17% of the variance in psychological domain (R2 = 0.17; F [1, 48] = 9.83, P = 0.003) and 32% of variance in social domain (R2 = 0.32; F [1, 48] = 22.45, P < 0.001). TMT-A, which is indicative of attention and psychomotor processing speed, accounted for nearly 12% of variance in environmental domain.
Table 6. Neuropsychological performance as a predictor for quality of life domains in patients
|R2 = 0.12; F (1, 48) = 6.22, P = 0.016; Physical domain as DV|
|R2 = 0.17; F (1, 48) = 9.83, P = 0.003; Psychological domain as DV|
|R2 = 0.32; F (1, 48) = 22.45, P < 0.001; Social domain as DV|
|R2 = 0.23; F (1, 48) = 14.58, P < 0.001; Environmental domain as DV|
Regression was also repeated after the addition of the significant clinical variables for each QoL domain (namely, age of patient for psychological domain, number of depressive episodes for social domain, education for environmental domain) in the first block and significant cognitive variables in the second block, however the results were not very different (not tabulated). The significant predictive model remained the same in the case of social domain, while in the case of psychological and environmental domain, the age of the patient and number of years of education were respectively added to the model, slightly increasing variance further to 26% and 32%.
GAF score showed significant correlation to Recognition (P = 0.001) subtest of memory scale, which explained nearly 22% of the variance (rs = 0.47; rs2 = 0.22).
- Top of page
To the best of our knowledge, this is the first report from the Indian subcontinent to comprehensively investigate the correlation between cognitive impairments and QoL. To date, very few studies[27-29] in the Western literature have focused on the influence of cognitive impairments on QoL of BD patients.
The findings from this study reveal that euthymic patients with BD have significant impairments in selected cognitive functions compared to healthy controls. These include poor performance in tests for attention, information processing speed and executive function in addition to impairments in verbal memory. Several similarities, as well as contrasts, emerged on comparison with previous literature.[1, 2, 30] The executive functions and verbal memory were found to be the most marked deficits in a recent meta-analysis for the euthymic period. In addition, response inhibition, sustained attention, psychomotor speed and set-shifting were also found to be impaired, though to a lesser degree. These findings are broadly similar to the findings from the current study.
A contrasting feature in our study is the stark preservation of response inhibition (SCWT-CI). It is noteworthy that the results for executive functions show a divide across the two tests used, which reflects in a poor cognitive flexibility and set-shifting ability in the TMT-B and a preserved selective attention and response inhibition in the color interference card of SCWT. A previous study had also found a divide in the results of executive functions, however the findings were the reverse of ours, with preserved cognitive flexibility and impaired response inhibition. In contrast to previous studies,[31, 32] visual memory was preserved in the present study. The observation that memory deficits in euthymic patients are mediated by deficient organizational strategies at encoding or poor retrieval processes[33, 34] rather than deficient retention was confirmed by a largely preserved memory retention or storage in the memory tests.
Most previous studies have reported that patients with a greater number of episodes and longer duration of illness suffer greater cognitive decline.[35-37] In this study, time taken for completion of TMT-B correlated significantly with the duration of illness and number of episodes, while number of errors in the 2-Back test showed significant correlation with duration of illness. These associations may be indicative of a progressive disease process, thereby increasing the degree of cognitive impairments with time.
The patients in the present study had a significantly poorer QoL in the psychological and social domains. The global functioning also differed markedly among the two groups, with patients having an overall poorer social, psychological and occupational functioning. In combination, these measures provide self-rating as well as clinician's judgment regarding patient functioning. This finding is consistent with most previous studies.[27, 38-40]
More importantly, what has emerged from this study is confirmation of the hypothesis that cognitive dysfunction is a predictor of poorer QoL in the stable phase of BD. So far, only the mood symptoms were thought to be related to a poor QoL. In fact, it has been mentioned that sub-depressive symptoms may be the most potent predictors of low QoL, even in remitted patients. We have made a strong effort to rule out the possibility of residual affective symptoms by incorporating the criteria of clinical remission for a minimum of 2 months as per the psychiatrist's assessment, as well as application of standard rating scales to screen for depression or mania just prior to assessment. In this regard, our study is more rigorous than some previous ones.
The cognitive impairments explained 12–32% of the variance for various QoL domains. Given the fact that psychological and social domains differed significantly among patients and controls, there may be a possible role of cognitive impairments in explaining these group differences. Perhaps a better cognitive functioning: (i) makes one more equipped to handle one's feelings and thoughts; (ii) increases psychological adjustment and adaptation to social situations; and (iii) enhances social relationships, resulting in better QoL. This study, however, does not investigate a causal correlation and indicates only association and predictive ability.
Of all the tests, the TMT-B, an indicator of cognitive flexibility and set shifting, was found to be a significant predictor of psychological and social domains of QoL, explaining 17% and 32% of the variance. Previous studies have also indicated that perhaps executive dysfunction represents the central bipolar trait deficit, and this impairment underlies not only the broad cognitive deficit pattern that is observed, but also the psychosocial and functional deficits exhibited by patients.[41-43] Recognition was found to be significantly predictive of GAF score, which is in contrast to a previous study where processing speed was robustly associated with social and global functioning in BD.
The influence of cognitive impairments on QoL has been demonstrated in only a few recent studies.[27-29] Brissos et al. studied the clinical and cognitive variables of QoL of BD patients, schizophrenia patients and controls. To their surprise, they found a negative impact of neurocognitive deficits on QoL, while in the same study there was no impact of cognitive deficits on the QoL of schizophrenia patients. A year later, in a different study, the same researchers reported that a poorer self-reported QoL correlated significantly with worse cognitive performance on tests of executive functioning and verbal abstraction. No significant correlation was found, however, between cognitive dysfunction and QoL in a separate study. The existing evidence base is limited and the area merits further research attention.
The present study has several important clinical implications. Cognitive deficits could be an important marker for future neurobiological and pharmacological research. Neurocognitive rehabilitation strategies and appropriate pharmacological strategies need to be developed with the goal of improving cognition and QoL of patients during the stable phase. Early diagnosis and active treatment potentially could reduce the cognitive morbidity associated with BD. Adequate cognitive functioning is desirable for the achievement of better QoL and special attention needs to be devoted to patients who remain functionally impaired despite the resolution of major affective symptoms.
The study was, however, limited by a small sample size and hospital setting. The findings cannot be generalized to individuals with BD from the community. Moreover, almost all patients were on medication, which could affect the cognitive functions and QoL due to their side-effect profile. The present cross-sectional nature of the assessment does not allow for causal inferences among these variables. Future research needs to better estimate the longitudinal correlation between cognitive function and psychosocial outcome across illness phases in the same patient cohort.
To conclude, it is important to understand the functional significance of cognitive impairments in the everyday lives of the patients. The cognitive assessment of patients with BD cannot, or rather should not, be done in isolation from assessment of its impact on psychosocial functioning and QoL of patients.