Background There is very limited evidence for the efficacy of any specific therapeutic intervention in chronic, treatment refractory major depression. Thermal anterior capsulotomy (ACAPS) is a rarely performed but established therapeutic procedure for this patient group. While benefit has been claimed, previous ACAPS reports have provided limited information. Detailed prospective reporting of therapeutic effects and side effects is required.
Objective To report a prospective study of therapeutic effect, mental status, quality of life, social functioning and neurocognitive functioning in individuals with chronic treatment refractory major depression, treated with ACAPS.
Method A prospective case series of 20 patients treated with ACAPS between 1992 and 1999 were reassessed at a mean follow-up of 7.0±3.4 years. Data were collected preoperatively and at long term follow-up. Structural MRI was performed in 14 participants.
Results According to a priori criteria, at long term follow-up, 50% were classified as ‘responders’ and 40% as ‘remitters’. Fifty-five per cent were classified as ‘improved’; 35% were ‘unchanged’; and 10% had ‘deteriorated’. Neurocognitive and personality testing were not significantly different at follow-up. A trend towards improvement in some aspects of executive neuropsychological functioning was observed. Significant adverse effects were infrequent and there were no deaths.
Conclusions ACAPS may represent an effective intervention for some patients with chronic, disabling, treatment refractory major depression that has failed to respond to other therapeutic approaches. The adverse effect burden within this population was modest, with no evidence of generalised impairment of neurocognitive functioning.
Statistics from Altmetric.com
Chronic depression, defined by DSM-IV as an episode persisting for over 2 years in duration, occurs in approximately 20% of patients with major depression.1 2 Remission rates decline substantially with increasing numbers of failed antidepressant trials, with one-third of individuals not remitting after four antidepressant trials.3 There is limited evidence to support specific interventions for chronic treatment refractory depression.4 In recent years, vagus nerve stimulation, deep brain stimulation (DBS) and the cognitive behavioural analysis system of psychotherapy5 6 have been reported to confer benefit. However, numbers of patients treated by DBS remain low with limited long term data. Also, the efficacy of vagus nerve stimulation and the cognitive behavioural analysis system of psychotherapy appears modest compared with treatment as usual,7 and there is no evidence for either of these treatments being effective in chronic, highly treatment refractory populations such as reported here.
First described by Talairach in 1949, anterior capsulotomy (ACAPS) places lesions in the anterior one-third of the anterior limb of the internal capsule, a large white matter bundle connecting prefrontal and anterior cingulate cortices with the thalamus, hippocampus and amygdala. ACAPS lesions are generated by the thermal effects of radiofrequency current or focal irradiation. While several published studies report the clinical outcomes of ACAPS for obsessive–compulsive disorder (OCD)8 and anxiety disorders,9 there is only one report on the outcome of ACAPS as a treatment for depression. Herner (1961) described outcomes for 19 patients with a diagnosis of ‘a depressive state’ (not otherwise defined clinically), treated by thermal ACAPS with a maximum follow-up of 2 years.10 Fourteen (74%) were rated as showing ‘permanent improvement’; three (16%) showed ‘temporary improvement’; one (5%) was ‘unchanged’; and one (5%) was ‘worse’. Fifteen (75%) experienced permanent side effects which were mild in 47%. Additionally, Simpson et al reported ‘marked’ improvements in almost 60% of 24 patients undergoing ACAPS for depression.11 However, details of these outcomes have not been published.
Although ACAPS has now been superseded in our unit (Dundee) by anterior cingulotomy, ACAPS remains the only neurosurgical procedure performed in England and Wales for intractable psychiatric illness. Between 2005 and 2009, the Mental Health Commission approved seven procedures which were performed at the University Hospital of Wales in Cardiff.12 13 ACAPS is also offered as a treatment for refractory OCD and major depression in several other European countries. The anterior limb of the internal capsule is one of the preferred targets for studies of DBS for OCD14 and remains a target for ablative neurosurgery for OCD. It is also emerging as a DBS target for major depression.15 This anatomical region, therefore, remains very important for neurosurgical interventions for psychiatric illness.
In Dundee, anterior cingulotomy became the procedure of choice due to early reports of adverse effects from ACAPS for OCD and anxiety disorders, many of which came from Sweden. We have previously highlighted the possibility of differences in adverse effects depending on the indication,16 but given the similar clinical outcomes reported for anterior cingulotomy,17 a decision was made to change procedures.
All patients who underwent ACAPS in Dundee from 1992 were considered. The only a priori exclusion criteria were inability to trace a participant or inability of the participant to provide written informed consent for the study. Ethics approval was provided by the Tayside Medical Research Ethics Committee. Written informed consent was obtained from all patients and, as a requirement of the relevant Scottish Mental Health Acts (1984 and 2003), all patients undergoing neurosurgery were independently assessed preoperatively by three panel members from the Mental Welfare Commission for Scotland with regard to the suitability of the proposed intervention and capacity to provide sustained informed consent. Diagnosis before surgery was determined according to ICD-10 by the lead psychiatrist for the neurosurgical programme. This diagnosis was subsequently corroborated retrospectively by the current study authors using OPCRIT.18 Information regarding the number of previous illness episodes and the duration of inter-illness episode recovery was extracted from detailed case note review to identify the course of individual episodes according to the ACNP Task Force definitions.19
The following rating scales were completed at the preoperative baseline, 12 month follow-up and long term follow-up: Hamilton Rating Scale for Depression, 17 item version (HRSD17);20 Montgomery–Åsberg Depression Rating Scale (MADRS);21 Clinical Global Impression (CGI);22 Pippard Postoperative Scale;23 Lancashire Quality of Life Profile;24 and the Personality Assessment Schedule.25 In addition, a semi-structured interview for adverse effects was conducted at long term follow-up. Case notes were reviewed for the occurrence of adverse effects occurring in the first postoperative year that were not recalled at later interview.
Assessment of treatment resistance
Ratings of the adequacy of previous antidepressant trials were made according to the Antidepressant Treatment History Form.26 Each treatment had to score ≥3 on the Antidepressant Resistance Rating to be defined as ‘adequate’. Treatment resistance was also rated according to the Massachusetts' General Hospital Scoring Method,27 and classified using the Thase and Rush Treatment Resistant Depression Staging (TR-S) method.28
Definitions of response
‘Response’ and ‘remission’ were defined a priori. Using the HRSD-17, ‘response’ was ≥50% reduction in baseline score, with ‘remission’ being a score of ≤7. For the MADRS, ‘response’ was ≥50% reduction in baseline score and ‘remission’ was a score of ≤10. On the clinician rated CGI-I, ‘response’ was a score of 1 or 2. Participants were deemed to meet criteria for clinical response if they met ‘response’ criteria on the HRSD-17, MADRS or the CGI ratings. Similarly, ‘remission’ was deemed present if criteria were met on either the HRSD-17 or MADRS scales. Categorical outcomes were also defined by CGI-I and the Pippard Postoperative Scale. ‘Improved’ was defined as a score of 1 or 2 on the CGI-I or I or II on the Pippard Scale. ‘Unchanged’ was defined as a score of 3 or 4 on the CGI-I or III or IV on the Pippard Scale. ‘Worse’ was a score of 5, 6 or 7 on the CGI-I or V on the Pippard Scale. As statistically significant changes may not necessarily reflect meaningful benefit to a particular individual, ‘clinically significant change’ (CSC) was calculated for each participant. These methods are described in detail by Jacobson and Truax.29
Clinical neuropsychological assessment
The following tests were performed at baseline and long term follow-up: National Adult Reading Test30; Information/Orientation, Word Lists, Paired Associates Learning, Logical Memory, Visual Reproduction and Digit Span tasks taken from the Wechsler Memory Scale–Revised (WMS-R)31 and subsequently the WMS-III32; Arithmetic, Comprehension, Block Design, Digit Symbol and Similarities tasks from the Wechsler Adult Intelligence Scale–Revised (WAIS-R)33 and more latterly the WAIS-III34; ‘6 elements’ test from the Behavioural Assessment of the Dysexecutive Syndrome35; the Verbal Fluency Test/Controlled Oral Word Association Test36; Stroop Neuropsychological Screening Test37; and Trail Making Test.38 Individuals who had been tested on a version of the WMS or WAIS prior to revision were tested again on the same version, with each new patient being tested on the most current version. Data at 12 months postoperatively were available for some patients.
Cambridge Neuropsychological Test Automated Battery (CANTAB)
CANTAB was used to assess a variety of cognitive domains (attention; memory; non-strategic learning; and ‘executive’ functions). These tests are sensitive to the effects of a range of neuropsychiatric disorders, including major depression.39–41
Surgical procedures were conducted under general anaesthesia. Volumetric CT and MRI images were obtained parallel to the intercommissural plane for each patient and merged using the FrameLink Stereotactic Linking System (Medtronic Inc, Minneapolis, Minnesota, USA). The bilateral target volume was 19–21 mm anterior to the anterior commissure. Two small scalp incisions and two burr holes were made at the entry points. A 3 mm wide, 6 mm long radiofrequency probe was sited within the target such that its tip was at the inferior aspect of the target volume. Single lesions were made in each hemisphere by heating the probe to 70°C for 90 s using a Radionics radiofrequency lesion generator. Additional bilateral lesions were similarly made such that the initial lesion was extended caudally to produce a total lesion within each capsule of approximately 18–20 mm in height and 8 mm in width. Individual lesion volumes were therefore predicted to range from 670 mm3 (ellipsoid) to 1000 mm3 (cylinder) depending on shape.
Where data were normally distributed, summary data were reported using the mean±SD. Where data were not normally distributed, the median and range were chosen as measures of central tendency. Differences between two normally distributed variables were assessed using t tests. Comparison between groups was performed with independent sample t tests and comparison between two time points was performed with paired sample t tests. Corresponding non-parametric tests were the Mann–Whitney U test.
Incomplete HRSD-17 scores
Preoperative HRSD-17 scores were unavailable for 30% of participants but preoperative MADRS scores were available. To test the validity of inferring HRSD-17 scores from MADRS scores, MADRS imputed HRSD-17 scores for all participants at long term follow-up were compared with actual clinician rated scores (completed for all participants) completed at the same time. The correlation between the two scores was highly significant (Pearson correlation coefficient, r=0.963, p<0.001). Consequently, preoperative MADRS scores were used to infer preoperative HRSD-17 scores.
Incomplete MADRS scores
Baseline MADRS scores were not available for one participant and missing at 12 months for eight (40%). To determine if there were differences between those with complete data and those with missing data at 12 months, groups were compared on a number of demographic variables (gender, age at surgery) and indices of severity (duration of presenting episode, number of previous episodes, number of adequate treatment trials and baseline MADRS scores). The only significant difference was a greater number of preoperative hospital admissions in those with missing data.
MR imaging was obtained for 14 patients, 6.5±3.6 years after ACAPS using a Siemens 1.5 T scanner (Magnetom Avanto and Symphony models). T2 weighted images were obtained with a matrix of 512×512 and a slice thickness of 4.5–5 mm. Using SPM2 (http://www.fil.ion.ucl.ac.uk/spm), images were spatially normalised to the SPM2 template which conforms to Montreal Neurological Institute anatomical space, then re-sliced to form 1 mm3 isotropic voxels. Each spatially normalised image was then segmented to form separate grey matter, white matter and CSF images. ACAPS lesions appear as CSF density structures. For each scan, the outline of each lesion was highlighted as a ‘region of interest’ (ROI) on each slice using MRIcro (http://www.sph.sc.edu/comd/rorden/) independently by two raters (DC and JDS). The former based the estimation on the spatially normalised T2 image; the latter used the CSF images. The ratings made by both assessors were averaged. The centres of mass of the lesions were calculated by averaging the voxel coordinates in the ROIs in each dimension and the volumes were calculated from the sum of all voxels within the ROIs (each being 1 mm3).
At long term follow-up, seven participants had undergone subsequent procedures: as data were available to determine ‘response’ at 12 months, lesion coordinates for ACAPS were compared between responders (n=4) and non-responders (n=6) at 12 months postoperatively. Differences between groups were examined using Mann–Whitney U tests (recognising the low power of this comparison to reject the null hypothesis of no difference). In order to visualise possible differences in lesions between responders and non-responders, the ROI for each participant was plotted on the same spatially normalised MRI.
Twenty-three patients were identified as suitable for inclusion. Three subjects were excluded (one had died; one was untraceable; and one had a previous subcaudate tractotomy), leaving a study sample of 20 patients. Long term follow-up assessments were completed 7.0±3.4 years (range 47.4–662.2 weeks) after ACAPS. These data represent outcomes from the initial ACAPS only. Seven participants went on to have a second procedure (ACAPS=1; anterior cingulotomy=6) due to inadequate clinical response. One individual subsequently had a vagus nerve stimulator implanted (outwith the Dundee service). The second procedure was performed after a mean interval of 154.5±55.2 weeks. For two participants, this interval was less than 120 weeks and the longest follow-up reported below represents the assessment prior to the second procedure. Long term outcome data, therefore, represent outcomes from the first ACAPS only. Demographic details for participants (n=20) are shown in table 1 and diagnoses are given in table 2. All were Caucasian and six (30%) were subject to compulsory detention at the time of neurosurgery.
Indices of treatment resistance are shown in table 3. All participants had at least two lifetime courses of electroconvulsive therapy (ECT). Fifty per cent of participants had four or more courses of ECT during the presenting major depressive episode and only one participant (5%) had not had ECT in the presenting major depressive episode due to previous non-response.
At 12 months, five (25%) participants met criteria for ‘response’ and two (10%) also met criteria for ‘remission’. At long term follow-up, 10 (50%) met criteria for ‘response’ with eight (40%) also meeting criteria for ‘remission’. These categories are not mutually exclusive and remitters also met criteria for response. According to categorical rating scales, at long term follow-up, 11 (55%) were ‘improved’, seven (35%) were ‘unchanged’ and two (10%) were ‘worse’. Categorical scores were also obtained from participants and next of kin. Correlations between raters were highly significant and ranged from p=0.025 to p<0.001. Changes in scores are shown in table 4. At 12 months, 4/20 (20%) participants had achieved CSC and at long term follow-up nine (45%) met criteria for CSC. Correlations between rates of CSC and clinical definitions of ‘response’ at 12 months were highly significant using Spearman's rank correlation (r=0.909, p<0.001) and remained significant for ‘remission’ at 12 months (r=0.592, p=0.002), and for both ‘response’ (r=0.761, p<0.001) and ‘remission’ (r=0.757, p<0.001) at long term follow-up.
A summary of changes in score on a variety of neuropsychological measures is presented in table 5. The median duration of follow-up for clinical neuropsychological assessment was 396.0 (range 117.8–625.8) weeks, and for CANTAB testing the median was 401.1 (48.4–662.2) weeks. There were non-significant improvements in the majority of tests of general intelligence, memory and executive function.
Scores on the Personality Assessment Schedule were available at baseline for 11 (55%) participants, long term follow-up for 11 (55%) and at both time points for six (30%) participants. At baseline, 2/11 had either personality difficulties or disorder, and at long term follow-up, 1/6 of those with scores at both time points had moved from the ‘personality difficulties’ to ‘no personality disorder’ category. Clinical interview at long term follow-up did not identify features that were suggestive of significant personality dysfunction. The IIP-PD, a self-report scale reflecting ‘personality dysfunction’,43 was completed by 14 participants at long term follow-up, seven (50%) of whom had scores suggestive of personality disorder (cluster C=6, cluster B=1). Four of the seven had undergone a subsequent anterior cingulotomy and only one had a score suggestive of personality disorder (cluster C) at long term follow-up.
The study team did consider that some subtle deficits in reciprocal social communication may have been evident for a small number of participants (n=4) during informal interactions but it is not known if these were present prior to surgery. Despite these subjective perceptions, it is worth noting these patients seemed to function with considerable social competence and did not show significant deteriorations in neurocognitive functioning on the clinical or computerised batteries. Gross frontal lobe symptoms or signs were not detected in any of the participants at clinical interview.
There were no deaths by suicide. One patient died prior to study completion at the age of 70 years. This occurred 12 years after ACAPS and was considered attributable to causes unrelated to neurosurgery. One individual developed a left-sided temporal haematoma approximately 6 years after ACAPS. Smoking and hypertension were considered key aetiological factors. The most common adverse effects reported are shown in table 6.
Adverse effects were typically transient and there were no adverse effects present at long term follow-up that could be reliably attributed to ACAPS, with the exception of disturbance of bladder control (n=2) which was mild in both cases. All but one participant were taking prophylactic antidepressant medication at long term follow-up. Vulnerability to addiction was difficult to assess. One participant had given up smoking 4 years after ACAPS while another participant developed alcohol dependence following a period of psychosocial stress. A possible relationship with surgery was unclear in both cases. There were self-reported weight gains made by many of the study participants over the follow-up period. Although this could not be attributed definitively to the procedure, such findings are consistent with other reports of ACAPS in the literature. Similarly, reduced motivation was reported by a number of participants and it was not possible to fully distinguish between amotivation consistent with frontal lobe impairment and lack of motivation caused by depression and/or medication. A classic frontal lobe syndrome was not observed in any participants.
Lesion coordinates and volumes are shown in table 7. The measured lesion volumes for both hemispheres lay within the surgically expected range. The results are depicted in figure 1. Although the differences are not statistically significant, which is unsurprising given the small numbers in each group, there is a suggestion of a trend for ‘responder’ lesions to be placed more inferiorly within the internal capsule. There were no significant differences between groups in x, y or z planes.
This is the largest follow-up study of ACAPS for chronic refractory major depression and the first since 1961. These data represent the most complete and most detailed assessment of the clinical outcomes of ACAPS for treatment refractory major depression. In our study, we found that 10/20 (50%) patients met the criteria to be described as responders approximately 7 years after surgery, with 8/20 (40%) meeting criteria for clinical remission (categories not mutually exclusive). The population described in our study is drawn from the extreme end of the treatment refractory depressive spectrum. Recent reports of the outcomes for vagus nerve stimulation and electrical DBS for chronic refractory depression describe populations that are less chronic and refractory than our cohort.7 44 45
Detailed prospective testing of neuropsychological function failed to demonstrate consistent deteriorations over time in any cognitive domain. Indeed, there was a general trend towards improvement on many measures of executive function. We believe this may be mediated indirectly by reduced global symptom burden. However, it could be argued that the neuropsychological battery used in this study was insufficient to detect all potential sequelae of ACAPS. Previously, we have reported that patients treated with anterior cingulotomy and ACAPS exhibited a general impairment of accuracy of recognition of dynamic emotional displays compared with healthy controls.46
This study differs from recent reports from ACAPS for OCD and anxiety disorders9 47 in a number of important respects: data on both clinical and neuropsychological measures for our cohort were collected prospectively and we used modern quantitative voxel based methods (VBM) to examine lesion characteristics. Importantly, VBM overcomes the problems inherent with traditional radiological subjective descriptive methods which do not correct for individual differences in brain size and differences in brain orientation at the time of scanning. While attempts have been made to describe the optimal target for OCD or anxiety disorders,48 49 almost nothing is known about optimal ACAPS lesion characteristics for treating depressive illness. Based on our current work,50 we suggest that the reporting of lesion coordinates in Montreal Neurological Institute space and the use of VBM should be generally adopted.
There are several limitations to our study. Firstly, the dataset is incomplete. However, full data on symptom severity were available for 95% of participants at long term follow-up. Secondly, the timing of the assessments (12 months and long term follow-up) can provide only a snapshot of an individual's level of functioning over time and the results should be viewed within the context of a chronic and relapsing–remitting illness. Thirdly, this study, like many others, suffers from a lack of validated psychometric tools for assessing personality before and after intervention. There are, to our knowledge, no validated scales that are designed for repeated testing, that are insensitive to changes in depressive symptoms and which have sufficient sensitivity to detect potentially subtle alterations in personality characteristics. We currently use the Iowa Scales of Personality Change,51 an informant rated tool, in addition to other assessments to rate personality postoperatively. Finally, the absence of a control group is common to all reports of interventions of this type. Available data would suggest that response rates to ‘treatment as usual’ in much less refractory populations are low,7 and in a more refractory cohort ‘spontaneous’ remission is unlikely to account for the improvements seen.
The target of ACAPS is the horizontally oriented fibres in the inferior part of the capsule which connect the frontal lobes with the thalamus.52 The orbitofrontal cortex has rich connections with limbic structures, including the subgenual cingulate cortex, insular cortex and the hippocampus.53 Fibres within the anterior limb of the internal capsule constitute three distinct fibre tracts54: (1) fibres originating in the anterior thalamic peduncle, which connect to the cingulum bundle,55 the target of anterior cingulotomy; (2) a deeper system of fibres made up of frontopontine fibres; and (3) a third set which does not run in bundles and connects the caudate nucleus and the lentiform nucleus. A recent neuroimaging study demonstrated neuronal degeneration within these three pathways in five patients after ACAPS.56 A mechanism of action for ACAPS is, as yet, unknown. Dalgleish et al suggested that the therapeutic effects of a different ablative neurosurgical procedure, subcaudate tractotomy, might be affected by a relative insensitivity to negative feedback information.57 However, prospective data were not available to confirm whether such changes were, indeed, acquired.
One persisting criticism of psychiatric neurosurgery is that the benefits occur in a minority and that the risk of neuropsychological and/or personality dysfunction is high. This ignores varying surgical techniques and patient selection criteria at different centres. We have not been able to find empirical support for either of these statements, for the technique and criteria used in Dundee. More broadly, ACAPS remains a treatment option for patients with chronic treatment refractory major depression and is offered by a number of centres performing psychiatric neurosurgery. The ACAPS lesion target is of active interest in trials of DBS for OCD and depression, and patients are often offered ablative ACAPS should their trial of DBS prove unsuccessful. Therefore, the clinical outcomes of ACAPS remain of importance for contemporary clinical practice. In our series, the ACAPS procedure was associated with significant long term clinical improvements for almost half of all patients. In addition, the adverse effect burden does not appear to be as great as described in the literature, at least with our selection criteria and neurosurgical procedure. ACAPS, as defined here, remains a viable treatment option for highly selected patients with chronic refractory depression.
The Dundee psychiatric neurosurgical programme was conceived and established by George Fenton, Professor of Psychiatry, and TK Varma, Consultant Neurosurgeon. The authors thank Craig Adam, Alex Bell and Dr June Gilchrist for assistance with the conduct of neuropsychological testing and video recording of assessments, and Eleanor Sorrell for the conduct of Personality Assessments.
Funding The study was funded by the Chief Scientist Office of the Scottish Executive Health Department, grant No CZG/2/223. The funding body had no influence on the design of the study, collection or interpretation of the data.
Competing interests KM, DC, RM and MSE provide clinical management for the Dundee Advanced Interventions Service. KM has received research funding from GlaxoSmithKline and Cyberonics Inc, honoraria associated with lecturing and/or travel from Eli Lilly, Wyeth, Medtronic and Bristol Myers Squibb, and has accepted consultancy fees and/or hospitality from Lundbeck and St Jude Medical. MSE has received research funding from Cyberonics Inc and honoraria or travel support from Medtronic and St Jude Medical. DC has received consultancy fees from Servier Laboratories and honoraria for lectures from Wyeth and Lilly. He has received travel from Medtronic and Cyberonics Inc. JDS has received research funding via an honorarium associated with a lecture from Wyeth.
Ethics approval Ethics approval for this study was provided by the Tayside Medical Research Ethics Committee. Written informed consent was obtained from all patients, and as a requirement of the relevant Scottish Mental Health Acts (1984 and 2003), all patients undergoing neurosurgery were independently assessed preoperatively by three panel members from the Mental Welfare Commission for Scotland.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.