Dear Editor,
Re: A response from Noble et al. to e-letter by Psychology Task Force of the International League Against Epilepsy
Cognitive behavioural therapy (CBT) has been recommended for treating depression in people with epilepsy (PWE).[1, 2] The clinical significance of the effects of CBT for PWE has though, not been considered. We therefore systematically searched the literature for randomised controlled trials of CBT for PWE [3] and used Jacobson’s criteria [4] to empirically determine whether PWE made clinically reliable improvement. We compared this to that seen in the control arms of these trials.
Our main findings were that the likelihood of statistically reliable improvement in symptoms of depression was significantly higher for those PWE randomised to CBT compared to control conditions. The overall proportion of PWE achieving reliable improvement was low – 30% compared to 10% in the control arms. For most PWE, symptoms were unchanged.
The proportion of PWE who improve following CBT is limited. It should serve as a clarion call for the development of more effective treatments. Indeed, our review may have inflated CBT’s benefit since some trials included PWE without clinical distress at baseline and so it was not possible to apply Jacobson’s second, more stringent criterion and calculate for what proportion CBT also resulted in recovery.
The Psychology Task Force of the International League Against Epilepsy submitted a response to our article. [5] We are pleased that our article has drawn attention to the issue of what currently available treatments are able to offer PWE and that is has stimulated a needed discussion on the topic. Our paper agreed with many of the points raised by the Task Force:
1. More well conducted trials of psychological interventions are needed. Methodologically weak trials are known to overestimate treatment effects;
2. CBT can alleviate depression for some PWE in clinical practice, but substantial room for improvement exists;
3. Participant samples need to be better described and standardised criteria should be used to do this. Evaluations of sample representativeness can then be made, and potential treatment moderators identified;
4. Trials need to test the benefit of treatments for the full range of clinical depression. Trials have so far tended to exclude those with severe levels of depression, and some have included PWE without depression at baseline;
5. Trials should use measures of depression that are valid and reliable for the epilepsy population and for the context in which they are used.
However, there were several comments in the Task Force’s response which misrepresented our position. The Task Force also presented some reasons that they felt might explain the low rate of improvement in depression detected by our review. Below we clarify our position and address the explanations forwarded by the Task Force.
MISREPRESENTATIONS
Comment 1 by Task Force: “...we hope to encourage health professionals to continue to refer patients for psychotherapy... We argue that if just under 1 in 3 PWE reliably improve with CBT, which has no known side-effects, this is better than a possible alternative of unmanaged depression... the ILAE Psychology Task Force believes that it is inaccurate to label CBT as ‘ineffective’.”
Nowhere in our article did we recommend not referring PWE who are distressed for CBT or psychotherapy. Our review explicitly states that CBT does lead to statistically reliable improvement for some PWE and is preferable to offering no treatment. Depression has many deleterious effects and should not be left unmanaged.
Our review indicates that substantial room for treatment improvement exists. Consequently, we said “that it seems reasonable to conclude that when only 30% of treated patients make an improvement that highly effective CBT treatments for distress in epilepsy do not exist.” We did not, as asserted, label CBT as ineffective. Rather, we said the approach was “largely ineffective”.
Comment 2 by Task Force: “the authors suggest CBT... could even lead to lower ‘self–esteem’ and ‘helplessness’... the latter conclusions were based on hypothetical reactions to treatment, rather than empirically supported outcomes.”
Here a discussion point is mistaken for a conclusion. In our discussion we appropriately considered what implications our results have for what therapists should tell PWE considering CBT and related this to ethical and professional guidelines. We questioned how patients might feel if they were not informed about likely treatment response and went on to experience no benefit. Based on clinical experience amongst our team, and evidence from the literature,[6] we suggested possible reactions might include increased feelings of helplessness and hopelessness and lower expectations of therapy being successful in the future.
REASONS PRESENTED BY TASK FORCE TO ACCOUNT FOR LOW RATE OF IMPROVEMENT:
Comment 3 by Task Force: “we offer alternative interpretations of the findings...we would like to highlight the heterogeneity of the studies pooled and how this impacts the findings... First, the interventions were very diverse... Interestingly, one trial that utilised a standardised CBT protocol, resulted in 50% reliable change reductions in depressive symptoms... Second, over 10% of patients in the analyses had depressive symptoms within the non-clinical ranges.”
The pooled risk difference (RD) across the trials in our review was 0.20 – meaning 20% more PWE randomised to the CBT treatment arms reliably improved compared to the control arms. The trials in our review used different modes of treatment delivery and had samples with different levels of pre-treatment distress. We acknowledged this in our article and to explore whether this heterogeneity obscured higher rates of improvement in some trials, we presented RDs for the individual trials and also pooled RDs for trials using similar treatments and/or samples.
Unfortunately, regardless of how one approached the dataset (i.e., just looking at change in individuals who received face to face treatments or focusing on change in just participants who were distressed at baseline) only around 20% more patients in the CBT groups reliably improved compared to the control arms (see supplementary Figures 1-2 of our review [3]).
The Task Force highlighted the rate of reliable change reported by one trial in our review and implied that it’s results better reflect the potential of CBT for PWE. The trial in question was by Ciechanowski et al.[7]. Unfortunately, the benefit of CBT in that trial was actually not much better than indicated by the pooled RD. The Task Force correctly point out that the proportion of reliable change in the treatment arm of Ciechanowski et al.’s trial was 50%. They neglected to consider though that the proportion who reliably improved in that trial’s (treatment as usual) control arm was 24%. Thus, the RD for Ciechanowski et al.’s trial is 0.26 – not that much different to the pooled RD of 0.20.
Comment 4 by Task Force: “One important limitation of previous trials is the relatively short duration of psychotherapy offered to PWE, a factor that Noble et al. [1] acknowledge... it is very likely that participants did not receive a sufficient dosage of CBT... many PWE experience cognitive difficulties, including memory impairment, which may require more intensive and tailored CBT [5].”
Our review could only comment on the extent of change elicited by CBT treatments so far evaluated for PWE by trials. On average, the treatments offered by the trials in our review lasted 8.8 sessions. The Task Force suggest more treatment sessions may be required for PWE due to the presence of cognitive impairment. Whilst some PWE do experience such difficulties, this explanation does not appear to be able to account for the low rate of improvement seen in the trials in our review since most of them excluded patients with such difficulties.
Comment 5 by Task Force: “One advantage of CBT is that many of the behavioural skills; such as problem solving, sleep hygiene and controlled relaxation can also be tailored to assist with the self-management of epilepsy... which Noble et al. did not consider.”
We focused on trials for which the primary outcome measure was depression and examined the degree of change achieved

We thank you for the opportunity to respond to the Task Force’s letter. We trust that we have clarified our position on a range of points, including the need at present to continue to refer PWE who are distressed for CBT. We stand by our position though that there remains an urgent need for more effective treatments to be developed.

Yours sincerely,
Adam J. Noble
James Temple
James Reilly
Peter L. Fisher.

REFERENCES
[1] Kerr MP, Mensah S, Besag F, de Toffol B, Ettinger A, Kanemoto K, Kanner A, Kemp S, Krishnamoorthy E, LaFrance WJ, Mula M, Schmitz B, van Elst LT, Trollor J, Wilson SJ. International consensus clinical practice statements for the treatment of neuropsychiatric conditions associated with epilepsy. Epilepsia 2011;52: 2133-2138.
[2] Barry JJ, Ettinger AB, Friel P, Gilliam FG, Harden CL, Hermann B, Kanner AM, Caplan R, Plioplys S, Salpekar J, Dunn D, Austin J, Jones J, Advisory Group of the Epilepsy Foundation as part of its Mood Disorder. Consensus statement: the evaluation and treatment of people with epilepsy and affective disorders. Epilepsy & Behavior 2008 13: S1-29.
[3] Noble AJ, Reilly J, Temple J, Fisher PL. Cognitive-behavioural therapy does not meaningfully reduce depression in most people with epilepsy: a systematic review of clinically reliable improvement. Journal of Neurology, Neurosurgery & Psychiatry 2018; doi: 10.1136/jnnp-2018-317997.
[4] Jacobson NS, Truax P. Clinical significance: a statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology 1991;59: 12-19.
[5] Gandy M, Reuber M, LaFrance Jr WC, Modi A, Wagner JL, Goldstein LH, Tang V, Donald KA, Valente KD, Michaelis R. Why it’s still important to consider referring patients with epilepsy (PWE) with depression for psychotherapy – including Cognitive Behaviour Therapy. Response from the Psychology Task Force of the International League Against Epilepsy. Journal of Neurology, Neurosurgery & Psychiatry 2018; Published on: 11 July 2018.
[6] Crawford MJ, Thana L, Farquharson L, Palmer L, Hancock E, Bassett P, Clarke J, Parry GD. Patient experience of negative effects of psychological treatment: results of a national survey. The British Journal of Psychiatry 2016;208: 260-265.
[7] Ciechanowski P, Chaytor N, Miller J, Fraser R, Russo J, Unutzer J, Gilliam F. PEARLS depression treatment for individuals with epilepsy: a randomized controlled trial. Epilepsy & Behavior 2010;19: 225-231.

Dear Editor,

We read with great interest the recent paper of Sechi et al. that describes 13 patients with anti-GFAP related myelitis and compares them with 41 patients with anti-AQP4 related myelitis.1 To date, very little data is available about anti-GFAP related disorders.2-3 Sechi et al. highlight some differences between the two entities to help clinicians differentiate them.1 One of these clinical differences relates to area postrema syndrome (APS). Indeed, it is well known that APS is a classical feature of neuromyelitis optica spectrum disorders, particularly among anti-AQP4 positive patients.4 Sechi et al. report this syndrome as a prodromal event in 20% of anti-AQP4 related myelitis. Conversely, the authors do not report any case of APS preceding or accompanying myelitis related to anti-GFAP.1 Given these data, APS could be an indicator for ruling out anti-GFAP encephalomyelitis, particularly useful for centers that do not yet have access to biological testing for anti-GFAP Abs.
However, we report the case of a 41-year-old woman who in April 2016 developed intractable nausea and vomiting lasting for five weeks and leading to 35 kilograms in weight loss. An extensive search for a digestive disease was negative, and no neurological explorations were performed. One month following the resolution of digestive symptoms, she developed mental confusion, diplopia, dysarthria, dizziness, bilateral blurred vision (with optic disc edema) and paraparesis. Brain MRI was highly suggestive of anti-GFAP encephalomyelitis, notably with the presence of diffuse T2 abnormalities in periventricular white matter, with linear perivascular enhancement oriented radially to the ventricles.Spinal cord MRI revealed longitudinally extensive T2 hyperintensity (extended to the brainstem, with an upper limit involving the area postrema), with ill-defined margins, as described by Sechi et al.1. Anti-AQP4 and anti-MOG IgG were negative in the serum. Anti-GFAP IgG were detected in the serum (by cell-based assay and tissue-based assay). Lumbar puncture was performed before the patient was referred to our department, and was not repeated given the dramatic clinical improvement rapidly achieved with intravenous steroids. Among the 13 patients reported by Sechi et al., anti-GFAP IgG were detected only in the serum for two patients for which no CSF was available, like for our patient.
The patient gave written informed consent for information and images to be published
Our case underlines that the clinical spectrum of anti-GFAP encephalomyelitis is not fully described and understood yet and that clinical features of this syndrome may overlap with those of neuromyelitis optica spectrum disorders (related to anti-AQP4 IgG, to anti-MOG IgG or negative for both antibodies). As such, combining clinical data with radiological and biological data remains essential in order to properly classify patients with such rare central nervous system inflammatory disorders.

References
1 Sechi E, Morris PP, McKeon A et al. Glial fibrillary acidic protein IgG related myelitis: characterisation and comparison with aquaporin-4-IgG myelitis. J Neurol Neurosurg Neuropsychiatry 2018 (in press).
2 Fang B, McKeon A, Hinson SR et al. Autoimmune glial fibrillary acidic protein astrocytopathy: a novel meningoencephalomyelitis. JAMA Neurol 2016;73(11):1297-1307.
3 Flanagan EP, Hinson SR, Lennon VA et al. Glial fibrillary acidic protein immunoglobulin G as biomarker of autoimmune astrocytopathy: analysis of 102 patients. Ann Neurol 2017;81:298-309.
4 Wingerchuk DM, Banwell B, Bennett JL et al. International consensus diagnostic criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177-189.

Imagine you are an epilepsy health professional seeing a patient with clinical symptoms of depression. What should you do? If you have read Noble et al.’s [1] recent JNNP review, entitled ‘Cognitive-behavioural therapy does not meaningfully reduce depression in most people with epilepsy…’ you may have become sceptical about the potential of CBT, or psychotherapy in general, to alleviate depression in people with epilepsy (PWE). This recent systematic review pooled data from five small randomised controlled trials (RCTs), with some elements of CBT for PWE, and performed an analysis of reliable change. ‘Pooled risk difference indicated likelihood of reliable improvement in depression symptoms was significantly higher for those randomised to CBT’, but the authors focused on the finding that ‘only’ 30% of patients receiving interventions, compared to 10% of controls, could be considered ‘reliably improved’. Emphasising the fact that over 2/3 of patients did not meet this criterion for improvement, the authors suggest CBT is ‘ineffective’, has ‘limited benefit’ and could even lead to lower ‘self–esteem’ and ‘helplessness’. Notably, the latter conclusions were based on hypothetical reactions to treatment, rather than empirically supported outcomes.

Therefore, the purpose of this letter, written by the Psychology Task Force of the International League Against Epilepsy (ILAE), is to argue that caution is needed when considering the authors’ conclusions and potential implications for the psychological care of PWE. Moreover, we offer alternative interpretations of the findings and raise the pertinent issue of how psychotherapy for PWE may continue to improve. Importantly, we hope to encourage health professionals to continue to refer patients for psychotherapy, which is an effective intervention for a substantial subgroup of PWE.
Consistent with Reuber’s [2] editorial commentary, Cognitive-behavioural therapy does meaningfully reduce depression in people with epilepsy, we would like to highlight the heterogeneity of the studies pooled and how this impacts the findings. First, the interventions were very diverse, and most would not be considered standardised CBT protocols for depression. Interestingly, one trial that utilised a standardised CBT protocol, resulted in 50% reliable change reductions in depressive symptoms, equivalent to CBT in the general population [3]. Second, over 10% of patients in the analyses had depressive symptoms within the non-clinical ranges. Further, unlike previous analysis of reliable change in depression [3], this review failed to control for baseline levels of depression severity. Third, Noble et al. collapsed data from four different self-report depression measures, only one designed for PWE. Depression in PWE can have distinct symptomatology, given the presence of seizures (peri-ictal depression) and anti-seizure medication effects, both of which can limit the validity of generic depression measures [4].
Noble et al. [1] describe their conclusion that 30% reliable improvements in depressive symptoms across trials is ‘ineffective’ as a ‘value judgement’, illustrating subjectivity, which Reuber’s [2] editorial commentary argued could be interpreted completely in reverse. That is, one could conclude from the data that the treatments are ‘effective’ for PWE. There is little consensus about what we expect a ‘reliable improvement’ to be in psychological distress for PWE, or for patients with a disabling neurological disorder in general. A stated goal of epilepsy treatment is “no seizures, no side-effects.” However, many PWE continue to have seizures, and all biological treatments have potential side-effects. We argue that if just under 1 in 3 PWE reliably improve with CBT, which has no known side-effects, this is better than a possible alternative of unmanaged depression. Arguments regarding quantifying ‘reliable improvement’ aside, we do agree with Noble et al.’s [1] conclusions that there is ‘substantial room for improvement’ in the treatment of depression for PWE.
One important limitation of previous trials is the relatively short duration of psychotherapy offered to PWE, a factor that Noble et al. [1] acknowledge. Across the five RCTs, there was only an average of 7 hours (8 sessions) of psychotherapy, the adherence of which is unclear [5]. Thus, it is very likely that participants did not receive a sufficient dosage of CBT, especially given a minimum of 12 sessions is indicated for depression in the general population [3]. In addition, many PWE experience cognitive difficulties, including memory impairment, which may require more intensive and tailored CBT [5]. These limitations need to be addressed and psychotherapy should be tailored to the unique needs of PWE. One advantage of CBT is that many of the behavioural skills; such as problem solving, sleep hygiene and controlled relaxation can also be tailored to assist with the self-management of epilepsy (e.g. avoidance of seizure triggers), which Noble et al. did not consider.
Another critical area for improvement is the treatment of comorbid anxiety symptoms within psychotherapy for depression. Anxiety and depression are highly comorbid, and in clinical practice it is difficult to evaluate them separately [4]. As such, transdiagnostic treatments, which treat depression and anxiety in one protocol, are increasingly being adopted and proven to be effective in the general population. We disagree with Noble et al.’s [1] comments regarding the ‘disappointing’ evidence for the treatment of anxiety, as only one small trial is cited assessing the impact of CBT for depression (not anxiety), on a secondary anxiety measure. A conclusion of insufficient evidence would have been more accurate, given the state of the anxiety literature.
Depression in PWE remains underdiagnosed and treated, perhaps partially due to uncertainty about effective treatments [4]. At worse this results in poorer quality of life and higher suicide rates in PWE. Thus, the development of more effective psychotherapies, including alternatives to CBT, is warranted. However, the ILAE Psychology Task Force believes that it is inaccurate to label CBT as ‘ineffective’ based on the findings of Noble et al.’s [1] review. Instead, we encourage health professionals to interpret the Noble et al. [1] conclusions with caution, given the concerns raised with respect to depression outcome measures, dosage and quality of the psychotherapies, and interpretation of results. Further, even with a conservative estimate of 30% responders to the psychotherapies, we posit that CBT shows promise for treating depression in PWE and should remain a strong treatment consideration for the referring clinician.

This document was written by experts selected by the International League Against Epilepsy (ILAE) and was approved for publication by the ILAE. Opinions expressed by the authors, however, do not necessarily represent the policy or position of the ILAE.

References:
1. Noble AJ, Reilly J, Temple J, et al. Cognitive-behavioural therapy does not meaningfully reduce depression in most people with epilepsy: a systematic review of clinically reliable improvement. Journal of Neurology, Neurosurgery & Psychiatry 2018 doi: doi: 10.1136/jnnp-2018-317997
2. Reuber M. Cognitive-behavioural therapy does meaningfully reduce depression in people with epilepsy. Journal of Neurology, Neurosurgery and Psychiatry 2018 doi: 10.1136/jnnp-2018-318743
3. Ogles BM, Lambert MJ, Sawyer JD. Clinical Significance of the National Institute of Mental Health Treatment of Depression Collaborative Research Program Data. Journal of Consulting and Clinical Psychology 1995;63(2):321-26. doi: 10.1037/0022-006X.63.2.321
4. Kwon O-Y, Park S-P. Depression and Anxiety in People with Epilepsy. Journal of Clinical Neurology (Seoul, Korea) 2014;10(3):175-88. doi: 10.3988/jcn.2014.10.3.175
5. Modi AC, Wagner J, Smith AW, et al. Implementation of psychological clinical trials in epilepsy: Review and guide. Epilepsy and Behavior 2017;74:104-13. doi: 10.1016/j.yebeh.2017.06.016

We appreciate Dr. Laura S. Boylan’s interest in our article. However, her viewpoint rather strikingly exemplifies the behaviors that she mistakenly believes we were guilty of in our report, beginning with the “eminence based” statements “in my view” occurring twice in the first paragraph and her conclusion with a personal “old saw I use in teaching”. We strongly disagree with her misinterpretation about our application of “cognitive bias” in the selection of our patients for this case-control study. In fact, most patients with Parkinson’s disease (PD) had functional complications ascertained after several visits –requiring a diagnostic revision once they fulfilled the appropriate positive criteria.1 The diagnostic “delay” in part may have highlighted the absent recognition of functional comorbidities in PD prior to our study, forcing a conservative approach before ascertaining what may be considered a “second” diagnosis in these patients. Furthermore, in contrast to Dr. Boylan’s suggestion, we did not select patients on the basis of comorbid depression, anxiety, cognitive symptoms, pain, nausea, or fatigue. Instead these features segregated more commonly among cases than controls after the patient selection had been completed. She argues that we considered them “supportive” for a diagnosis of functional movement disorder, but we did not. We have instead emphasized the potentially misleading influence of both history and psychiatric features and argued in favor of a diagnosis of a functional movement disorder exclusively reliant on neurological examination (i.e., using positive criteria rather than applying a diagnosis of exclusion).2 Importantly, we did not review colleagues' charts; these were our own patients, who had been followed by movement disorders experts over several years. Dr. Boylan exercises far more of her own cognitive bias in highlighting how the diagnosis of functional disorders has been overly considered in females (despite considerable epidemiological evidence that these disorders are more common in women) and in her old teaching saw that “proper diagnosis” of movement disorders (including functional movement disorders) "is the one offered the most senior professor in the vicinity". Her viewpoint ignores modern literature that has shown a very low incidence of misdiagnosis using current evaluations and the application of positive criteria for the diagnosis of functional disorders.3 Thus, in citing old, historical literature and ignoring a growing body of work in the field,4 Dr. Boylan’s letter propagates old thinking and misinformation.

References
1. Gupta A, Lang AE. Psychogenic movement disorders. Curr Opin Neurol. 2009;22(4):430-436.
2. Espay AJ, Lang AE. Phenotype-specific diagnosis of functional (psychogenic) movement disorders. Curr Neurol Neurosci Rep. 2015;15(6):556.
3. Stone J, Carson A, Duncan R, et al. Symptoms 'unexplained by organic disease' in 1144 new neurology out-patients: how often does the diagnosis change at follow-up? Brain. 2009;132(Pt 10):2878-2888.
4. Hallett M, Stone J, Carson A. Functional Neurologic Disorders. Vol 139. Amsterdam, Netherlands. : Handbook of Clinical Neurology, Elsevier. ; 2016.