Artificial intelligence and headache.

BACKGROUND AND METHODS: In this narrative review, we introduce key artificial intelligence (AI) and machine learning (ML) concepts, aimed at headache clinicians and researchers. Thereafter, we thoroughly review the use of AI in headache, based on a comprehensive literature search across PubMed, Embase and IEEExplore. Finally, we discuss limitations, as well as ethical and political perspectives. RESULTS: We identified six main research topics. First, natural language processing can be used to effectively extract and systematize unstructured headache research data, such as from electronic health records. Second, the most common application of ML is for classification of headache disorders, typically based on clinical record data, or neuroimaging data, with accuracies ranging from around 60% to well over 90%. Third, ML is used for prediction of headache disease trajectories. Fourth, ML shows promise in forecasting of headaches using self-reported data such as triggers and premonitory symptoms, data from wearable sensors and external data. Fifth and sixth, ML can be used for prediction of treatment responses and inference of treatment effects, respectively, aiming to optimize and individualize headache management. CONCLUSIONS: The potential uses of AI and ML in headache are broad, but, at present, many studies suffer from poor reporting and lack out-of-sample evaluation, and most models are not validated in a clinical setting.

Characterizing neck pain during headache among people with migraine: Multicountry results from the Chronic Migraine Epidemiology and Outcomes - International (CaMEO-I) cross-sectional study.

OBJECTIVE: To assess the prevalence and impact of neck pain during headache among respondents with migraine in the multicountry Chronic Migraine Epidemiology and Outcomes - International (CaMEO-I) Study. BACKGROUND: Neck pain among individuals with migraine is highly prevalent and contributes to disability. METHODS: The CaMEO-I was a prospective, cross-sectional, web-based study conducted in Canada, France, Germany, Japan, United Kingdom, and the United States. A demographically representative sample of participants from each country completed a screening survey to evaluate headache characteristics. Respondents with headache were identified as having migraine or non-migraine headache based on modified International Classification of Headache Disorders, third edition, criteria; those with migraine completed a detailed survey with migraine-specific assessments. Results were stratified by the presence or absence of neck pain with headache (NPWH). For these analyses, data were pooled across the six countries. RESULTS: Of 51,969 respondents who reported headache within the past 12 months, 14,492 (27.9%) were classified as having migraine; the remaining 37,477 (72.1%) had non-migraine headache. Overall, 9896/14,492 (68.3%) of respondents with migraine headache reported NPWH, which was significantly higher (p < 0.001) than the proportion of respondents with non-migraine headache who reported NPWH (13,536/37,477 [36.1%]). Among respondents with migraine, moderate-to-severe disability was significantly more prevalent for those with NPWH versus without (47.7% [4718/9896] vs. 28.9%, p < 0.001). Respondents with NPWH versus without also had significantly greater work productivity losses, at a median (interquartile range [IQR]) of 50.0 (20.0, 71.3) vs. 30.0 (0.0, 60.0) (p < 0.001), lower quality of life (Migraine-Specific Quality of Life questionnaire version 2.1, median [IQR] Role Function-Restrictive domain score 60.0 [42.9, 74.3] vs. 68.6 [54.3, 82.9], p < 0.001), higher prevalence of depression and anxiety symptoms (depression, 40.2% [3982/9896] vs. 28.2% [1296/4596], p < 0.001); anxiety, 41.2% [4082/9896] vs. 29.2% [1343/4596], p < 0.001), higher prevalence of cutaneous allodynia during headache (54.0% [5345/9896] vs. 36.6% [1681/4596], p < 0.001), and higher prevalence of poor acute treatment optimization (61.1% [5582/9129] vs. 53.3% [2197/4122], p < 0.001). CONCLUSIONS: Nearly 70% of respondents with migraine reported NPWH. Individuals with migraine with neck pain during their headaches had greater disability, depression, anxiety, and cutaneous allodynia (during headache) than those without neck pain during their headaches. They also had diminished quality of life and work productivity, and poorer response to acute treatment compared with those without neck pain.

Characterizing gaps in the preventive pharmacologic treatment of migraine: Multi-country results from the CaMEO-I study.

OBJECTIVE: To analyze data from the Chronic Migraine Epidemiology and Outcomes-International (CaMEO-I) Study in order to characterize preventive medication use and identify preventive usage gaps among people with migraine across multiple countries. BACKGROUND: Guidelines for the preventive treatment of migraine are available from scientific organizations in various countries. Although these guidelines differ among countries, eligibility for preventive treatment is generally based on monthly headache day (MHD) frequency and associated disability. The overwhelming majority of people with migraine who are eligible for preventive treatment do not receive it. METHODS: The CaMEO-I Study was a cross-sectional, observational, web-based panel survey study performed in six countries: Canada, France, Germany, Japan, the United Kingdom, and the United States. People were invited to complete an online survey in their national language(s) to identify those with migraine according to modified International Classification of Headache Disorders, 3rd edition, criteria. People classified with migraine answered questions about current and ever use of both acute and preventive treatments for migraine. Available preventive medications for migraine differed by country. MHD frequency and associated disability data were collected. The American Headache Society (AHS) 2021 Consensus Statement algorithm was used to determine candidacy for preventive treatment (i.e., ≥3 monthly MHDs with severe disability, ≥4 MHDs with some disability, or ≥6 MHDs regardless of level of disability). RESULTS: Among 90,613 valid completers of the screening survey, 14,492 met criteria for migraine and completed the full survey, with approximately 2400 respondents from each country. Based on the AHS consensus statement preventive treatment candidacy algorithm, averaging across countries, 36.2% (5246/14,492) of respondents with migraine qualified for preventive treatment. Most respondents (84.5% [4431/5246]) who met criteria for preventive treatment according to the AHS consensus statement were not using a preventive medication at the time of the survey. Moreover, 19.3% (2799/14,492) of respondents had ever used preventive medication (ever users); 58.1% (1625/2799) of respondents who reported ever using a preventive medication for migraine were still taking it. Of the respondents who were currently using a preventive medication, 50.2% (815/1625) still met the criteria for needing preventive treatment based on the AHS consensus statement. CONCLUSIONS: Most people with migraine who qualify for preventive treatment are not currently taking it. Additionally, many people currently taking preventive pharmacologic treatment still meet the algorithm criteria for needing preventive treatment, suggesting inadequate benefit from their current regimen.

Does pain self-efficacy predict, moderate or mediate outcomes in people with chronic headache; an exploratory analysis of the CHESS trial.

BACKGROUND: Chronic headache disorders are disabling. The CHESS trial studied the effects of a short non-pharmacological intervention of education with self-management support for people affected by migraine and/or tension type headache for at least 15 days per month for at least three months. There were no statistically significant effects on the Headache Impact Test-6 (HIT-6) at 12-months. However, we observed improvement in pain self-efficacy questionnaire (PSEQ) and short-term HIT-6. We explored the impact of the CHESS intervention on PSEQ, and subsequently, on the HIT-6 and chronic headache quality of life questionnaire (CH-QLQ) at four, eighth and 12 months. METHODS: We included all 736 participants from the CHESS trial. We used simple linear regression models to explore the change of HIT-6 and CH-QLQ with treatment and PSEQ at baseline (predictor analysis), and the interaction between treatment and baseline PSEQ (moderator analysis). We considered the change of PSEQ from baseline to four months as a mediator in the mediation analysis. RESULTS: Baseline PSEQ neither predicted nor moderated outcomes. The prediction effect on change of HIT-6 from baseline to 12 months was 0.01 (95% CI, -0.03 to 0.04) and the interaction (moderation) effect was -0.07 (95% CI, -0.15 to 0.002). However, the change of PSEQ from baseline to 4-month mediated the HIT-6 (baseline to 8-, and 12-month) and all components of CH-QLQ (baseline to 8-, and 12-month). The CHESS intervention improved the mediated variable, PSEQ, by 2.34 (95% CI, 0.484 to 4.187) units and this corresponds to an increase of 0.21 (95% CI, 0.03 to 0.45) units in HIT-6 at 12-months. The largest mediated effect was observed on the CH-QLQ Emotional Function, an increase of 1.12 (95% CI, 0.22 to 2.20). CONCLUSIONS: PSEQ was not an effective predictor of outcome. However, change of short-term PSEQ mediated all outcomes, albeit minimally. Future behavioural therapy for chronic headache may need to consider how to achieve larger, and more sustained increases level of self-efficacy than that achieved within the CHESS trial. TRIAL REGISTRATION: ISRCTN79708100.

Consensus recommendations on the role of nurses in headache care: A European e-Delphi study.

BACKGROUND: Nurses work at headache centres throughout Europe, and their care for migraine patients is acknowledged. However, the specific roles and tasks of nursing vary, and a unified understanding is lacking, posing challenges to knowledge sharing and research. OBJECTIVES: Using an e-Delphi study method, the objective is to obtain healthcare professional headache experts' opinions on nursing-specific roles and tasks and combine this into consensus statements for nurse recommendations for migraine treatment. METHODS: A three-round questionnaire study was conducted with nurses and neurologists from 18 specialised headache centres in 10 countries. In round 1, statements were compiled from a systematic examination of existing literature and expert opinions. In rounds 2 and 3, the experts rated the importance of statements (from round 1) on a 5-point Likert scale. Statements were analysed using a content analysis method, and the consensus of pre-defined statements was evaluated with gradually increased predetermined criteria using descriptive statistics. RESULTS: Twenty-one experts, representing all 10 countries, participated. The predetermined consensus of ≥70% agreement was reached for 42 out of the initial 63 statements. These statements formed the final recommendations within two themes: "The nurses' roles and tasks in the clinical setting" and "The nurses' roles and tasks in educating patients and colleagues." The consensus level of statements was strong, with 40% receiving unanimous agreement (100%) and 97% achieving relatively high agreement (>80%). CONCLUSION: Nursing plays a vital role with diverse tasks in migraine care. This study offers practical recommendations and a framework for nurses, equipping them with a clinical tool to enhance care and promote a coordinated approach to migraine treatment.

What predicts citation counts and translational impact in headache research? A machine learning analysis.

BACKGROUND: We aimed to develop the first machine learning models to predict citation counts and the translational impact, defined as inclusion in guidelines or policy documents, of headache research, and assess which factors are most predictive. METHODS: Bibliometric data and the titles, abstracts, and keywords from 8600 publications in three headache-oriented journals from their inception to 31 December 2017 were used. A series of machine learning models were implemented to predict three classes of 5-year citation count intervals (0-5, 6-14 and, >14 citations); and the translational impact of a publication. Models were evaluated out-of-sample with area under the receiver operating characteristics curve (AUC). RESULTS: The top performing gradient boosting model predicted correct citation count class with an out-of-sample AUC of 0.81. Bibliometric data such as page count, number of references, first and last author citation counts and h-index were among the most important predictors. Prediction of translational impact worked optimally when including both bibliometric data and information from the title, abstract and keywords, reaching an out-of-sample AUC of 0.71 for the top performing random forest model. CONCLUSION: Citation counts are best predicted by bibliometric data, while models incorporating both bibliometric data and publication content identifies the translational impact of headache research.

Climate change and disorders of the nervous system.

Anthropogenic climate change is affecting people's health, including those with neurological and psychiatric diseases. Currently, making inferences about the effect of climate change on neurological and psychiatric diseases is challenging because of an overall sparsity of data, differing study methods, paucity of detail regarding disease subtypes, little consideration of the effect of individual and population genetics, and widely differing geographical locations with the potential for regional influences. However, evidence suggests that the incidence, prevalence, and severity of many nervous system conditions (eg, stroke, neurological infections, and some mental health disorders) can be affected by climate change. The data show broad and complex adverse effects, especially of temperature extremes to which people are unaccustomed and wide diurnal temperature fluctuations. Protective measures might be possible through local forecasting. Few studies project the future effects of climate change on brain health, hindering policy developments. Robust studies on the threats from changing climate for people who have, or are at risk of developing, disorders of the nervous system are urgently needed.

Adverse and serious adverse events incidence of pharmacological interventions for managing chronic and episodic migraine in adults: a systematic review.

Background: Migraine is the second most common prevalent disorder worldwide and is a top cause of disability with a substantial economic burden. Many preventive migraine medications have notable side effects that affect different body organs. Method: We systematically searched for published randomised controlled trials (RCTs) using terms for migraine/headache and preventive medications. Using eligibility criteria, two reviewers independently assessed the articles. Cochrane risk-of-bias tool was applied to assess the quality of the studies. Data were classified by system organ class (SOC). Results: Thirty-two RCTs with 21 780 participants met the eligibility criteria for the incidence of adverse events (AEs). Additionally, 33 RCTs with 22 615 participants were included to synthesise the incidence of serious AEs (SAEs). The percentage of attributed AEs and SAEs to each SOC for 10 preventive drugs with different dosing regimens was calculated. Amitriptyline and topiramate had a higher incidence of nervous system disorders; Topiramate was also associated with a higher incidence of psychiatric disorders. All drugs showed a certain incidence of infections and infestations, with Onabotulinumtoxin A (BTA) having the lowest rate. BTA had a higher incidence of musculoskeletal disorders than the other drugs. Calcitonin gene-related peptide (CGRP) monoclonal antibodies (MAbs) such as fremanezumab and galcanezumab were linked to more general disorders and administration site conditions than other drugs. Conclusion: Notably, the observed harm to SOCs varies among these preventive drugs. We suggest conducting head-to-head RCTs to evaluate the safety profile of oral medications, BTA, and CGRP MAbs in episodic and/or chronic migraine populations. PROSPERO registration number: CRD42021265993.

Cognitive assessment during the phases of a spontaneous migraine: a prospective cohort study.

INTRODUCTION: Cognitive symptoms are reported commonly throughout all phases of a migraine; however, there is a paucity of objective cognitive profiling. Previous studies have been limited by practice effect, and variable populations. METHODS: Participants completed 1 month of daily testing with a computerised cognitive battery involving a simple reaction (SRT), choice reaction (CRT) and a working memory test (WM). Results were correlated with their diary to identify interictal scores, and scores during each phase of a migraine, and non-migraine headache days. RESULTS: A total of 16 patients with episodic migraine participated. During the headache phase of a migraine, responses to SRT, CRT and WM tasks were significantly slower and less accurate than interictally. During the postdrome, WM task performance was slower and less accurate. Non-migraine headache days were not associated with significant change. CONCLUSION: The headache and postdromal phase of a migraine day was associated with objective evidence of cognitive dysfunction in patients with episodic migraine.

Relationship of migraine and other forms of chronic pain.

The disability of migraine, a highly prevalent condition, is worsened by a second comorbid chronic pain condition. There is evidence of a relationship between migraine and several visceral pain conditions including irritable bowel syndrome, endometriosis, and dysmenorrhoea, as well as nonvisceral conditions including temporomandibular dysfunction, fibromyalgia, and lower back pain. While the mechanisms linking these conditions are inadequately surmised, a two-way relationship between migraine and these comorbidities likely exists. The progression and chronification of migraine is associated with peripheral and central sensitization, which may predispose to other conditions. Conversely, aspects of the mechanism of each comorbid condition may promote further migraine attacks. This chapter introduces each comorbidity, briefly summarizes the existing evidence, and discusses implications for treatment.

Twelve-month efficacy of CGRP monoclonal antibodies and predictive value of short-term response: results of an Australian multicentre study.

Introduction: Clinical trials show that calcitonin gene-related peptide monoclonal antibodies (CGRP mAbs) are effective preventative treatments for chronic migraine. Their efficacy over longer time periods and in cohorts originally excluded from trials remains uncertain. This study aims to explore the impact of CGRP mAbs in an Australian real-life setting. Methods: A multicentre cohort study was performed in the tertiary headache clinics of the Alfred and Austin Hospitals, Melbourne, Australia. Patients were commenced on a CGRP mAb for chronic migraine and asked to keep a headache diary, recorded at 3 monthly appointments for 12 months. Primary outcome was a ≥50% reduction in monthly headache days (MHD). Results: From a population of 105 patients, 90 patients commenced galcanezumab and 15 commenced fremanezumab. The ≥50% responder rate of the cohort was 52.4% after 3 months. Over 12 months follow-up, 25.7% of the cohort ceased due to a lack of efficacy and 16.2% ceased due to an adverse event. There was no difference in response or cessation between medications. There was poor agreement in 3-month and 12-month response rates (Cohen's κ=0.130; p=0.171). On subgroup analysis, continuous headache at baseline and number of trialled preventative treatments were the only factors associated with efficacy. Conclusion: CGRP mAbs were associated with sustained reductions in MHD over 12-month follow-up in patients with resistant migraine in Australia. Further studies are required to determine treatment options for patients with continuous headache. Poor agreement between outcomes at 3 and 12 months highlights the need to assess some patients at later timepoints.

Spontaneous intracranial hypotension.

Spontaneous intracranial hypotension (SIH) is a highly disabling but treatable secondary cause of headache. Recent progress in neuroradiological techniques has catalysed understanding of its pathophysiological basis and clinical diagnosis, and facilitated the development of more effective investigation and treatment methods. A UK-based specialist interest group recently produced the first multidisciplinary consensus guideline for the diagnosis and treatment of SIH. Here, we summarise a practical approach to its clinical and radiological diagnosis, symptomatic and non-targeted interventional treatment, radiological identification of leak site and targeted treatment of the leak once it has been localised.

Competing treatments for migraine: a headache for decision-makers.

BACKGROUND: Migraine is the world's second most common disabling disorder, affecting 15% of UK adults and costing the UK over £1.5 billion per year. Several costly new drugs have been approved by National Institute for Health and Care Excellence. AIM: To assess the cost-effectiveness of drugs used to treat adults with chronic migraine. METHODS: We did a systematic review of placebo-controlled trials of preventive drugs for chronic migraine. We then assessed the cost-effectiveness of the currently prescribable drugs included in the review: Onabotulinum toxin A (BTA), Eptinezumab (100mg or 300mg), Fremanezumab (monthly or quarterly dose), Galcanezumab or Topiramate, each compared to placebo, and we evaluated them jointly. We developed a Markov (state-transition) model with a three-month cycle length to estimate the costs and quality-adjusted life years (QALYs) for the different medications from a UK NHS and Personal Social Services perspective. We used a two-year time horizon with a starting age of 30 years for the patient cohort. We estimated transition probabilities based on monthly headache days using a network meta-analysis (NMA) developed by us, and from published literature. We obtained costs from published sources and applied discount rates of 3.5% to both costs and outcomes. RESULTS: Deterministic results suggest Topiramate was the least costly option and generated slightly more QALYs than the placebo, whereas Eptinezumab 300mg was the more costly option and generated the most QALYs. After excluding dominated options, the incremental cost-effectiveness ratio (ICER) between BTA and Topiramate was £68,000 per QALY gained and the ICER between Eptinezumab 300mg and BTA was not within plausible cost-effectiveness thresholds. The cost-effectiveness acceptability frontier showed that Topiramate is the most cost-effective medication for any amount the decision maker is willing-to-pay per QALY. CONCLUSIONS: Among the various prophylactic medications for managing chronic migraine, only Topiramate was within typical cost-effectiveness threshold ranges. Further research is needed, ideally an economic evaluation alongside a randomised trial, to compare these newer, expensive CGRP MAbs with the cheaper oral medications.

Abdominal migraine and cyclical vomiting syndrome.

Abdominal migraine and cyclical vomiting syndrome (CVS) are characteristic syndromes which have overlapping characteristics with migraine but lack the cardinal symptom of headache. Both abdominal migraine and CVS are characterized by recurrent attacks of nausea, vomiting, and/or abdominal pain lasting hours to a few days, with symptom freedom between attacks. Both abdominal migraine and CVS typically occur in children and adolescents, who often go on to develop more typical migraine headaches when older, but may also present for the first time in adults. Due to their shared characteristics and association with migraine headaches, abdominal migraine and CVS are sometimes called "migraine equivalents," and their pathophysiology is assumed to overlap with migraine headache. This chapter describes what is known about the clinical characteristics, epidemiology, pathophysiology, and prognosis of abdominal migraine and CVS, and explores their relationship to migraine. We also review the existing evidence for the nonpharmacological management, acute treatment of attacks, and preventive treatments for both abdominal migraine and CVS.

Epidemiology of migraine.

Migraine affects over a billion people worldwide and brings with it a huge burden of disability. It is a disease which disproportionally affects the working age population which heightens its economic impact, both at the individual family level and the societal level. Women are significantly more affected by migraine at every age and in all social and geographical groups. At the most severe end of the spectrum, chronic migraine is associated with poorer overall physical and mental health as well as increased risk of unemployment and lower household income. Estimates of the incidence and prevalence of migraine vary with sex, race, ethnicity, geography, socioeconomic, and educational status, suggesting there are many factors at play. In many cases, it is not clear whether these factors are causative of migraine, the effects of migraine, or (as is most likely) a combination of both. Future studies should aim to clarify these links, so that modifiable factors can be addressed where possible and those at risk of developing chronic migraine might receive targeted treatment at an early stage.

Clinical effectiveness of pharmacological interventions for managing chronic migraine in adults: a systematic review and network meta-analysis.

BACKGROUND: Chronic migraine can be a profoundly disabling disorder that may be treated with preventive medications. However, uncertainty remains as to which preventive medication is the most effective. We present a network meta-analysis to determine the effectiveness and rank of preventive drugs for chronic migraine in adults. METHODS: We identified, reviewed, and extracted data from randomised controlled trials (RCTs) of preventive drugs for chronic migraine with at least 200 participants. Data were analysed using network meta-analysis. FINDINGS: We included 12 RCTs of six medications (Eptinezumab, Erenumab, Fremanezumab, Galcanezumab, Onabotulinumtoxin A, and Topiramate) compared to placebo or each other. All drugs effectively reduced monthly headache and migraine days compared with placebo. The most effective drug for monthly headache days was Eptinezumab 300mg, with a mean difference of -2.46 days, 95% Credible Interval (CrI): -3.23 to -1.69. On the Surface Under the Cumulative Ranking Area (SUCRA) analysis, the probability that Eptinezumab 300mg was ranked highest was 0.82. For monthly migraine days, the most effective medication was Fremanezumab-monthly, with a mean difference: -2.77 days, 95% CrI: -3.36 to -2.17, and 0.98 probability of being ranked the highest. All included drugs, except Topiramate, improved headache-related quality of life. No eligible studies were identified for the other common preventive oral medications such as Amitriptyline, Candesartan, and Propranolol. The main reasons were that the studies did not define chronic migraine, were undertaken before the definition of chronic migraine, or were too small. INTERPRETATION: All six medications were more effective than the placebo on monthly headache and migraine days. The absolute differences in the number of headache/migraine days are, at best, modest. No evidence was found to determine the relative effectiveness of the six included drugs with other oral preventive medications. REGISTRATION: PROSPERO (number CRD42021265990).