In the current study, we evaluated neurocognitive performance in a cohort of bipolar I patients using the MCCB, a battery initially devised for use in clinical trials targeting cognition in patients with schizophrenia. This study is the first to report on the MCCB in a clinical sample other than schizophrenia and the results indicate that the MCCB adequately captures the most common deficits reported in BPD. Specifically, statistically significant impairments were present in the BPD patients in the domains of processing speed, attention, working memory, verbal learning, and visual learning; whereas certain executive functions (Reasoning and Problem-solving) and social cognition were not statistically significantly impaired. The severity of the impairments noted in our euthymic group were comparable to those previously reported by meta-analyses with effect sizes ranging from −0.6 to 1.2 SDs below average (Arts et al, 2008
; Bora et al, 2009a
; Robinson et al, 2006
; Torres et al, 2007
Of note, the symptomatic status of the BPD groups only modestly influenced the degree of impairment detected for most of the MCCB domains. There were no significant differences between euthymic BPD patients and symptomatic BPD patients on any of the MCCB domain scores; however, although symptomatic patients showed significant impairment on Reasoning and Problem-solving and Social Cognition, euthymic BPD patients did not significantly differ from the healthy controls in these domains. Future studies with larger samples using longitudinal designs will be necessary to better understand the effects of acute and subclinical affective symptoms on cognition within subjects, as the degree of cognitive recovery during affective remission is likely to be somewhat heterogeneous in BD (Bora et al, 2010
In a recent review, Yatham et al (2010)
rated the potential utility of different MCCB subtests in patients with BPD. Many of the MCCB subtests were considered to be highly applicable to bipolar samples based on previous research; however, several received reduced ratings based on limited available data. One such task, the HVLT has not been widely used in previous BPD studies, most of which have utilized a more challenging memory test (California Verbal Learning Test (CVLT)). The effect size reported in this paper for the euthymic BPD subjects in the verbal learning domain (Cohen's D
=−0.64) is somewhat smaller than those reported in meta-analytic studies using the CVLT (Cohen's D
=−0.83) (Bora et al, 2009a
). This smaller effect size is consistent with a previous report (Schretlen et al, 2007
), suggesting that this is not specific to our sample but rather reflects a slightly reduced sensitivity of the HVLT as compared with CVLT (Lacritz and Cullum, 1998
) to detect the full degree of impairment in bipolar patients. Still, for use in clinical trials in which repeated testing is inherent, the HVLT is likely to maintain an advantage over the CVLT because of the availability of many alternate equivalent forms. Moreover, Yatham et al, (2010)
rated the NAB Mazes and MSCEIT subtests as having ‘unclear' utility because of lack of published data on these measures in BPD. It is interesting that these are the two tasks/domains that we did not detect significant deficits in the euthymic bipolar sample. Executive functioning deficits in BPD patients are reported on some (Trailmaking Test Part B (D
=0.86); Wisconsin Card Sorting Test (WCST; D
=0.70)) (Bora et al, 2009a
), but not all, executive tasks. Thus, it is likely that NAB Mazes does not adequately assay the domain of Reasoning and Problem-solving in BPD, whereas a task such as the WCST might be more sensitive. Again, in the context of clinical trials for cognition, other test criteria must be considered including practice effects and alternate forms; therefore, although the WCST may tap into a different executive process, it still may not represent an optimal outcome measure for cognitive enhancement trials in BPD. Finally, research on social cognition in BPD is limited. Several studies have demonstrated impaired performance on some theory of mind (ToM) tasks in BPD that might assay a different construct than the MSCEIT; however it is unclear that to which extent ToM deficits are a byproduct of impaired executive functions (Bora et al, 2009b
Although BPD patients have a similar cognitive profile when compared qualitatively with schizophrenia patients, the deficits are generally less severe (Tabares-Seisdedos et al, 2008
) and there are several features of illness that influence cognitive functioning that differentiate BPD from schizophrenia. The episodic nature of BPD requires that studies account for current mood symptoms and clinical course factors, including number of previous episodes, age at onset, and duration of illness, bears directly on cognitive functioning (Martinez-Aran et al, 2004
; Robinson and Ferrier, 2006
). In addition, psychosis is present in a subgroup of approximately 50–60% of BPD I patients during acute affective episodes, whereas nearly as many BPD I patients never experience psychosis at all (APA (DSM-IV-TR), 2000). Data indicate that BPD I patients with a history of psychosis consistently perform worse than BPD patients who never experience psychosis, even during affective remission (Bora et al, 2009a
; Martinez-Aran et al, 2008
Given these BPD-specific features of cognition, it is likely that the optimal neurocognitive battery for many studies of BPD, including genetic and neuroimaging paradigms aimed at understanding its pathophysiology, would not be restricted to the MCCB. It is also likely that clinical trials will be carried forward from schizophrenia directly to a BPD sample in the case that an intervention is deemed successful for cognitive enhancement in schizophrenia. Thus, it is important to address whether the MCCB is sensitive to the more subtle impairments in BPD before it is applied in the context of a clinical cognitive enhancement trial. Our data suggest that, at least upon initial exposure to the MCCB, the deficits detected in bipolar patients are significant enough to suggest that there is room for improvement if a successful cognitive enhancement agent was to be administered.
Although the current study has taken the first step in evaluating the sensitivity of the MCCB in BPD, its appropriateness in a clinical trials design, which will necessitate repeated exposure to the battery, has not been addressed here. Additional data will be needed to evaluate the MCCB's test–retest reliability (repeatability), practice effects, and its potential ceiling effects in BPD before its use in clinical trials is justified. In addition, the relationship between the MCCB outcome measures and everyday functioning will be an area of interest for future trials (Harvey et al, 2010
), given the stipulation that the US FDA will require cognitive enhancement trials to not only show evidence of cognitive benefit on the MCCB but also an improvement on co-primary measures of functional capacity.
Our study has a number of limitations. Although the cohort was relatively large, we were only able to differentiate groups based on those currently experiencing clinically significant affective symptoms vs
patients who were euthymic at the time of testing. We were able to obtain an estimate of the influence of these symptoms on cognitive performance; however, we could neither differentiate mania from depression nor acute from subsyndromal level pathology. Moreover, our sample had a higher rate of psychosis history (72.5%) when compared with population estimates in bipolar I disorder (50–60%). This composition may have resulted in an overestimation of the degree of impairment common to a more typical cohort of bipolar I subjects, as it is known that psychosis history has a deleterious influence on cognition in BPD (Bora et al, 2009a
; Martinez-Aran et al, 2008
). The limited number of euthymic patients with no psychosis history in our sample (n
=9) did not allow for a meaningful comparison in this study. The effects of the course of illness on cognitive performance, including number of previous episodes and age at onset, should be a focus of future studies with larger samples. Finally, all but two, of the patients were taking psychotropic medications at the time of evaluation. The effects of medication were assessed based on dichotomous grouping of patients either taking or not taking a specific drug class. We did not detect significant medication effects using this approach; however, it remains possible that medications commonly used to treat BPD may contribute to cognitive impairment.
We present the first report of the MCCB in a well-characterized sample of patients with bipolar I disorder. Our findings provide evidence supporting the potential use of this battery as a platform in future clinical trials targeting cognition in BPD. The MCCB appears to adequately assay cognitive functioning in patients with BPD, regardless of clinical state at the time of testing; however, several additional psychometric properties should be addressed before its acceptance as the primary outcome measure in such trials. Finally, extension of this work into a broader cohort of bipolar patients including non-psychotic bipolar I and bipolar II patients will be useful in understanding differences in cognitive profiles among these subtypes.