Within the past decade, scientific and public interest in deep brain stimulation (DBS) as a promising investigational therapy for treatment-resistant depression (TRD) has dramatically increased.
1 DBS is a well-established first-line neurosurgical treatment for movement disorders and involves MRI-guided stereotaxic placement of electrodes within selected brain regions. An implantable pulse generator, typically placed in the chest, delivers focal stimulation to the electrodes according to clinician-programmed parameters. Relative to other types of neurosurgery, the main advantages of DBS are its reversibility and its adjustability in terms of stimulation settings.
The subgenual cingulate gyrus (SCG), the ventral capsule/ventral striatum, and the nucleus accumbens (NAcc) have been the most studied neurosurgical target areas for DBS in TRD to date, although single-case studies have identified other targets with positive symptomatic outcomes.
2–6 Large-scale controlled trials are already underway for some targets, with follow-up data suggesting that DBS for TRD is generally safe and moderately efficacious over the short term. However, continued long-term monitoring and data collection are needed to determine whether DBS represents a viable option for individuals who have not responded to standard treatments for depression and who continue to experience severe distress and functional impairments.
Concerns regarding the impact of DBS on cognition in those with TRD have been raised, given that the implant procedure and active stimulation may create functional lesions at the selected targets, and given that depression, per se, is already associated with cognitive dysfunction.
7–11 Clinical evaluation using neuropsychological measures appears clearly warranted in light of the following: 1) scarce data on long-term neuropsychological function following DBS for TRD; 2) previous findings that implicate the SCG in cognitive and memory function
12; 3) inconsistent findings that other treatments for depression (such as ECT or certain classes of antidepressant medications) are associated with cognitive impairment
13,14; and 4) previous findings suggesting that DBS for Parkinson’s disease may result in mild cognitive deficits depending on surgical trajectory and electrode location.
15,16Unfortunately, because of the relative novelty of this intervention for TRD, there have been no reliable data that describe long-term cognitive outcomes beyond the typical 12-month endpoint in studies of DBS for depression.
17,18 This article presents the first report of multi-year follow-up neuropsychological data for individuals receiving DBS for TRD.
Methods
Four consecutive subjects (one female, three male) from the Canadian multi-center DBS pilot study
19 targeting the SCG (Brodmann Area 25, Cg25) provided written informed consent to participate in a long-term follow-up (LTFU) protocol. A fifth subject who participated in the pilot study at this site was explanted due to nonresponse and was, thus, not enrolled in the LTFU protocol. However, analysis of this individual’s neuropsychological test scores from baseline through the 12-month period did not reveal any pattern of statistically reliable decline in performance. The research protocol was approved by the Clinical Research Ethics Board at the University of British Columbia and by the Vancouver Coastal Health Research Institute/Vancouver Coastal Health Authority.
All subjects underwent DBS implant surgeries with the Libra constant-current internal pulse generator device (St. Jude Medical) between January and December of 2006, as previously described,
19 and entered the LTFU portion of the study between September 2007 and August 2008. All subjects met DSM-IV-TR
20 criteria for Major Depressive Disorder (MDD) and had documented histories of resistance to standard treatments for MDD, including adequate trials with a minimum of four different classes of antidepressant medications and psychotherapy. Three of the four subjects were also unresponsive to ECT (the procedure contraindicated in the fourth because of the presence of a cerebral venous angioma). Patients did not have significant psychiatric or medical comorbidities and were deemed suitable neurosurgical candidates.
Subjects had received active DBS for approximately 42 months at the time of long-term neuropsychological follow-up. Following implantation, three of the four subjects had one contact activated per hemisphere, and the fourth had two contacts per hemisphere. DBS treatment parameters were initially programmed using continuous monopolar cathode stimulation, 130 Hertz frequency, 91 microseconds pulse widths, and 2.0‒4.5 milliamps amplitude. Over time, settings were adjusted, as dictated by patients’ responses and clinical need. All patients continued to take psychiatric medications after receiving DBS implants and into the LTFU period, with infrequent, standard-care medication adjustments as warranted. Clinical reassessments occurred every 6 months in the LTFU period. Demographic information for all four subjects, and as a group, is presented in
Table 1.
A comprehensive neuropsychological test battery was administered prior to surgery and again at 3, 6, 12 months, and an average of 42 months (LTFU) post-operatively. The test battery encompassed a number of domains of cognitive functioning including general intellectual ability, language ability, attention and working memory, verbal and visual memory, visuoperceptual and visuoconstructional ability, speed of information processing, and executive functioning. Specifically, the test battery was composed of the following measures: North American Adult Reading Test (NAART)
21; Wechsler Adult Intelligence Scale (WAIS‒III)
22; selected subtests (Logical Memory I and II, Family Pictures I and II) from the Wechsler Memory Scale‒3rd Edition (WMS‒III)
23; selected subtests (Trail Making, Fluency, Color-Word) from the Delis-Kaplan Executive Function System (DKEFS)
24,25; California Verbal Learning Test–II (CVLT‒II)
26; Benton Visual Form Discrimination Test (VFDT)
27; and the Boston Naming Test (BNT).
28 Further description of these standard neuropsychological measures is not included here but can be found in Spreen and Strauss.
29 Where available (DKEFS Verbal Fluency, CVLT‒II), alternate versions of these measures were used across testing sessions to mitigate practice effects. Depression was assessed via the Montgomery-Åsberg Depression Scale (MADRS),
30 the Hamilton Rating Scale for Depression (HRSD),
31 and the Inventory of Depressive Symptomatology (IDS).
32,33 Total scores for clinical scales are reported in
Table 2. In almost all cases, neuropsychological evaluations occurred within 7 days of the clinical assessments.
Neuropsychological test results were analyzed using the most appropriate normative sample with correction for age, gender, and education, where possible, derived from their respective test manuals. Z-score conversions were used for consistency across all data. Reliable change methodology
34 was used to evaluate for change in cognitive functioning, as this takes into account the test-retest reliability of the neuropsychological instruments in question. (RCI confidence interval=Initial test score +/− (Z*SEdiff) where SEdiff=Sqrt [(SEM1^2) + (SEM2^2)]. SEM1=Standard error of measurement Time1, SEM2=SEM at time 2, and SEM=SD*Sqrt[1-r12]. A 90% confidence interval was used, so 5% at each tail (i.e., Z score of 1.645).
Discussion
Despite fairly consistent reports by our subjects of post-implant word-finding difficulties, paraphasic errors of speech, and short-term memory issues, the neuropsychological data collected at various post-surgical timepoints, including at 42 months, did not indicate a systematic decline in any area of assessed cognitive functioning compared with pre-surgical performance. The results suggest that DBS is not only effective in reducing depressive symptomatology in the context of treatment-resistant depression, but that it also does not result in systematic detectable cognitive declines. However, it must be acknowledged that both pro-cognitive effects of reduced depressive symptomatology and/or practice effects from repeated test administration (particularly during the first 12 months) could be offsetting or masking some degree of cognitive decline.
With respect to the discrepancy between self-report and objectively measured cognitive function, there are several possible explanations. First, it is possible that having undergone brain surgery resulted in a heightened level of self-monitoring of internal cognitive performance. Another possibility is that self reports of cognitive impairment correlate poorly with objective indices of neuropsychological performance, particularly in patients with affective disorders.
35,36 Alternately, subjects’ self-reports may reflect subtle changes in cognitive efficiency that fell below the sensitivity of the cognitive measures. Finally, it is also possible that the surgical procedure/stimulation, itself, altered
perception of cognitive performance without altering actual performance. However, the fact that cognitive changes were objectively confirmed by relatives suggests that self-perception or self-monitoring was not the only responsible factor.
This report represents the first published long-term follow-up data on DBS in treatment-resistant depression. Further, complete neuropsychological data on all four participants across time points were available, the data collection procedure remained consistent across time points, and reliable change methodology was utilized. Although these data are not yet available, a 6-year post-operative follow-up is planned, and will provide further information with respect to the long-term stability of cognitive function following DBS for TRD.
As is the case with all research, this study had some specific limitations that may have influenced the results, as well as the conclusions that could be drawn. First, this study involved a small sample size of only four participants without a control group. Second, the high baseline intellectual ability of the sample is not characteristic of the general population and may limit the generalizability of the results. Third, despite efforts to control for reliability of neuropsychological instruments using reliable change methodology, the influence of repeated exposure to the test materials over multiple time points, particularly over the first 12 months of the study, was apparent. For example, despite already performing at a high-average to superior IQ level at the pre-surgical baseline, all four subjects displayed statistically reliable improvements in IQ at a number of time points, including at LTFU. Given that IQ is generally thought of as relatively stable, it is likely that the improved scores reflect some element of practice and increasing familiarity with the test materials and testing procedures.
37 (The exception to this is subject 2, who displayed a number of impaired scores at baseline and some statistically improved scores at LTFU [e.g., visual scanning speed, letter sequencing speed] that may reflect reversal of depression-related psychomotor slowing at baseline). It is also possible that these findings reflect pro-cognitive effects of reduced depression symptoms. However, 12-month follow-up data in a similar sample (Subgenual Cingulate Gyrus DBS for refractory depression) did not find an association between changes in cognitive function and improvement in mood.
18 In addition, 12-month follow-up data in a sample that received nucleus accumbens DBS for refractory depression revealed pro-cognitive effects that were independent of the antidepressant effects of DBS or changes in DBS parameters.
17 Further, a recent review of cognition and depression found little empirical support for pervasive depression-related deficits in general cognition, identifying only deficits in control of attention when not well-controlled by task demands.
8 Further, it has been identified that cognitive deficits persist in euthymic chronic unipolar depression suggesting an independent substrate from the depressed state contributing to cognitive deficits in this population.
39,40 Fourth, it is possible that medication or parameter changes influenced cognitive performance. However, no systematic relationship was noted between medication changes or parameter changes and subjective or objective measures of cognitive functioning in the follow-up period.
Although further research is needed in a larger sample to replicate the current LTFU findings and hopefully address some of the limitations of this study, this study provides preliminary reassurance of stability of cognitive functioning up to 42 months post-surgery and no evidence to suggest systematic observable declines in any area of cognitive ability.