Cognitive functions during migraine attacks

Tese de doutoramento, Medicina (Neurologia), Universidade de Lisboa, Faculdade de Medicina, 2015Background: Attack-related cognitive symptoms in migraine are frequent yet scarcely characterized and undervalued as contributors of disability. Conflicting evidence arose about an increased risk of cognitive decline in older migraine patients. Objectives: (1) to study the occurrence of cognitive symptoms in migraine attacks; (2) to evaluate objective evidence of cognitive dysfunction in migraine attacks and its neuronal correlates and (3) to study the effect of persisting migraine in cognitive function or cognitive decline in older adults. Methods: Occurrence of attack-related cognitive symptoms was detailed by systematic literature review and a cross-sectional clinical-based systematic survey; their relevance to disability was studied prospectively using headache diaries. An instrument (Mig-SCog) was developed, validated and tested to identify and quantify attack-related subjective cognitive symptoms. Cognitive function during attacks was evaluated by a systematic literature review and a clinical-based prospective two-period randomized cross-over study using an extensive neuropsychological battery. A briefer battery was tested in repeated applications in interictal patients and controls. Brain perfusion during attacks was studied with arterial spin labeling magnetic resonance imaging (ASL-MRI) and cortical response to a working memory task with blood-oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI). A prospective controlled cross-sectional population-based study of neuropsychological performance of older adults with persisting migraine and non-migraine headache was followed by a 5 years re-evaluation of the same sample, to screen for cognitive decline. Results: Cognitive symptoms were the most frequent non-migraine defining symptoms reported in the prodromic(37%) and headache(38%) phases of migraine attacks in a systematic review of 28 series, with a total sample of 8392 patients. Cognitive symptoms are also present in the postdromic or resolution phase, although fatigue (71%) is reported more often. Of 165 patients prospectively surveyed, 87% reported an average of 2.5 attack-related symptoms, over two-thirds executive (attention, processing efficiency and speed). Cognitive symptoms were ranked prospectively by 34 migraine patients recording 229 attacks, being second only to pain in terms of intensity and attack-related disability. An instrument to quantify migraine attack-related symptoms was constructed from a set of 43 candidate items, using factor analysis. The reduced 9 item Mig-SCog is fast to apply covering executive functions and language, having good internal consistency (Cronbachs’ alpha 0.82) and reliability (Cohen’s kappa 0.55) and high correlation with external validity measures such as the 43-candidate item list (rho=0.69) and the Cognitive Failures Questionnaires(rho=0.61). The Mig-SCog presents negligible recall bias (no difference in scores obtained during an attack or while headache free) and Migraine patients score it higher for migraine higher for migraine (7.9±4.6) than for non-headache pain (2.3±2.9, p<0.0006) or pain free (1.6±2.4, p<0.0006). Comparing Mig-SCog scores in migraine and tension-type headache patients, those were higher for migraine in all scale items (p<0.0001) except those related to naming (8 and 9). The AUC of Mig-SCog score for the diagnosis of Migraine was 0.835 (95% CI of 0.763-0.906, p< 0.0001) reinforcing specificity for migraine. Ten studies of neuropsychological evaluation during migraine attacks are available in the literature, only half had data allowing comparison of cognitive performance within and outside attacks (encompassing 163 migraine patients). All these were able to demonstrate some type of impairment (most often executive) although some bias could not be excluded from their study design. In our sample of 24 patients which completed an extensive neuropsychological evaluation in these two conditions (attack and headache-free) controlling for the majority of relevant bias (in particular the practice effect), performance was worse during the attack in the majority of cognitive tests, in particular in word reading speed (p=0.013), verbal learning (p=0.01), short term verbal recall with (p=0.01) and without (p=0.013) semantic cueing and delayed recall with (p=0.003) and without (p=0.05) semantic cues. Another sample of 24 interictal migraine patients and 24 matched controls performed equally in a shorter battery focused on executive functions that was applied twice with a short interval (average 45 days) to test the practice effect of repeated evaluations that was demonstrated in all tests, being significant in Stroop Interference test (p=0.002, multiplicity corrected); a meaningful score change was determined for each raw test scores. We were unable to find any relevant brain perfusion nor brain activation differences evoked by a working memory task during a spontaneous migraine without aura attack of an average intensity of 6.8 on a 0-10 VAS scale and an average duration of 16 hours in a sample of 13 women, compared to being headache-free. Persistent migraine or headache after the age of 50 related to worse performance in some neuropsychological tests (attention and processing speed in migraine patients, n=61; sematic memory and memory retrieval in non-migraine headache, n=50) in a population sample of 478 individuals tested extensively. After 5 years, 275 (57.5%) of the same sample were screened for cognitive decline, that occurred in 14.9% of the sample. Neither migraine nor non-migraine headache influenced the odds of decline. Discussion: Attack-related cognitive symptoms are very frequent, mostly executive and contribute to disability, supporting that they should be addressed as endpoint in clinical trials of acute migraine treatments and included in disability assessments. An efficient way to assess attack-related subjective cognitive symptoms in clinical practice or research is now available – the Mig-SCog. Although migraine-related reversible cognitive dysfunction was demonstrated during attacks, no advances on potential brain mechanisms underlying these findings were made. Interest is focused to obtain more functional data, with studies of evoked activation paradigms, functional connectivity and combined imaging and neurophysiological studies. Although persisting headache in older adults seems to influence executive performance, these changes are most likely adaptive and do not seem to influence the process of brain degeneration and associated cognitive decline