Migraine and obesity: epidemiology, possible mechanisms and the potential role of weight loss treatment
DS Bond, Department of Psychiatry and Human Behavior, The Warren Alpert Medical School of Brown University/The Miriam Hospital, Weight Control and Diabetes Research Center, 196 Richmond Street, Providence, RI 02903, USA. E-mail: firstname.lastname@example.org
Migraine and obesity are two public health problems of enormous scope that are responsible for significant quality of life impairment and financial cost. Recent research suggests that these disorders may be directly related with obesity exacerbating migraine in the form of greater headache frequency and severity, or possibly increasing the risk for having migraine. The relationship between migraine and obesity may be explained through a variety of physiological, psychological and behavioural mechanisms, many of which are affected by weight loss. Given that weight loss might be a viable approach for alleviating migraine in obese individuals, randomized controlled trials are needed to test the effect of weight loss interventions in obese migraineurs. Large-scale weight loss trials have shown that behavioural interventions, in particular, can produce sustained weight losses and related cardiovascular improvements in patients who are diverse in body weight, age and ethnicity. Consequently, these interventions may provide a useful treatment model for showing whether weight loss reduces headache frequency and severity in obese migraineurs, and offering further insight into pathways through which weight loss might exert an effect.
Migraine and obesity represent two major public health problems. Both disorders are prevalent, particularly in the USA where 12% of individuals are migraineurs and one-third of persons are obese (1,2). Migraine and obesity also exact a considerable toll on health. Both conditions are independent risk factors for cardiovascular disease (3,4), comorbid with pain-related and psychiatric conditions (5–7), and associated with quality of life impairment (8,9). Finally, migraine and obesity both impose a large economic burden on society. Healthcare costs attributable to migraine are estimated to exceed $11 billion annually (10), whereas productivity losses caused by migraine cost American employers $13 billion per year (11). Obesity is projected to account for $147 billion in healthcare costs (12), with indirect costs contributing an additional $65 billion to the total economic burden (13).
Recent research suggests that migraine and obesity may be directly linked. Obesity is related to higher frequency and severity of headache attacks among individuals who have migraine (14–16). There is also some evidence to suggest that obesity is associated with migraine prevalence (17,18). These findings recently gave rise to the notion that weight management strategies should be incorporated within a migraine treatment plan for patients who are obese (19). In this article, we describe the symptoms, epidemiology and pathophysiology of migraine. We then review studies that have examined the epidemiological association between migraine and obesity followed by a discussion of possible mechanisms that may link the two disorders. Finally, we explore the potential role of weight loss in treatment of migraine and potential mechanisms through which weight loss may alleviate headache attacks.
Migraine: symptomatology, epidemiology and disability, and pathophysiology
Migraine is a neurological disorder involving episodes of head pain that are frequently throbbing, unilateral and severe. According to the International Headache Society (20), attacks typically last 4–72 h and are often accompanied by nausea, vomiting, or sensitivity to light, sound, or movement. Approximately one-third of migraineurs have attacks that are preceded or accompanied by an aura, characterized by transient focal neurological symptoms that are most often visual but may also involve disturbances in sensory, speech and motor functioning.
Epidemiology and disability
Migraine prevalence is highest in individuals aged 25–55 years, typically the most economically productive period in people's lives (1). In the USA, 1-year prevalence of migraine is 18% in women, a rate that is three times that of men (9). Globally, migraine is most prevalent in the Americas and Europe and least prevalent in Africa and Asia (1,21). A similar race pattern is observed in the USA where Caucasians have higher rates than African Americans and Asian Americans, suggesting race-related differences in genetic vulnerability to migraine (22). Population-based data also suggest that low socioeconomic status contributes to elevated risk for migraine (9).
Migraine is not only common, but disabling. In the USA, an estimated 28 million adults have severe migraine headaches. Thirty-one per cent of migraineurs experience ≥3 severe headaches per month and over half (54%) report that their headaches cause severe impairment or necessitate bed rest (23). Migraine is estimated to cause 112 million bedridden days per year, a figure that corresponds to 300 000 people staying in bed every 24 h because of headaches (11,21). Several studies also suggest that migraine-related disability causes substantial disruption to family life in terms of postponement of household work, withdrawal from social and family activities, and adverse effects on relationships at home and at work (24).
While genetic factors are likely involved in susceptibility to migraine (25), the events that initiate migraine attacks are not well understood. Cortical spreading depression, described as a self-propagating wave of neuronal depolarization that spreads over the cortex (particularly the occipital region), is believed to underlie the migraine aura and trigger headaches via stimulation of the trigeminal nerve (26–29).
Specifically, cortical spreading depression is proposed to stimulate release of substances from the brain and blood vessels including ions and neurotransmitters such as glutamate and nitrous oxide. These substances activate the trigeminal innervations of meningeal nociceptors to release several neuropeptides, including calcitonin gene-related peptide (CGRP), substance P and neurokinin A, all of which induce vasodilation and leakage of plasma proteins, leading to neurogenic inflammation and employment of macrophages and mast cells (26–29). Simultaneously, modulation of both central pain processing pathways, in the thalamus, gray matter of the midbrain and hypothalamus, as well as vascular and autonomic function occurs via serotonin and catecholamines (29). The throbbing quality of the pain of migraine is associated with activation of pain-sensitive structures along the cerebral vasculature, augmented by the effect of vasoactive intestinal peptide on the extracerebral vasculature. Nausea and other autonomic symptoms are associated with central and peripheral autonomic activation, while photophobia and phonophobia, and perhaps allodynia are attributed to sensitizing of central pain processing pathways (28). The clinical picture of migraine is thus represented as a culmination of autonomic and nociceptive peripheral and central nervous system dysfunction, coupled with sterile ‘neurogenic’ inflammation and neuropeptide dys-regulation.
Epidemiological relationship between migraine and obesity
Research is more consistent in supporting the link between obesity and migraine frequency and severity than it is between obesity and migraine prevalence. The relationship between migraine and obesity was first evaluated in a clinic-based study that showed obese patients were three times as likely as age-matched normal-weight controls to have migraine (30). Several large cross-sectional studies have since investigated this relationship in population-based samples (see Table 1). Bigal et al. (14) evaluated 30 215 participants with headache history, 3719 of whom had migraine. Obesity was not associated with increased prevalence of migraine, but was related to headache attack frequency. Reporting of frequent headaches (10–14 headache d/mo) increased from 4.4% in the normal-weight group to 13.4% in the obese group and 20.7% in the severely obese group. Additionally, migraineurs in the severely obese group were nearly twice as likely to report experiencing severe headache pain, compared with those in the normal-weight group. Finally, headache disability and clinical features of migraine (photophobia, phonophobia) were also shown to increase with body mass index (BMI) category.
Table 1. Migraine and obesity: findings from population-based studies
|Bigal and Lipton(15)||30 215||38.7 (18–89)||65||Underweight (<18.5)||15.8||• Migraine prevalence was not related to obesity.||• Migraineurs who were overweight (OR = 1.3), obese (OR = 2.9) and morbidly obese (OR = 5.7) were at increased risk for having high headache frequency (10–15 headache d/mo) compared with those who were normal weight.|
• Migraineurs who were overweight (1.25), obese (1.31) and morbidly obese (1.9) were at increased risk for having severe headaches compared with those who were normal weight.
|Normal weight (18.5–24.9)||13.2|
|Morbidly obese (≥35.0) ||14.0|
|Bigal et al. (16)||162 576||≥12||NA||NA||NA||NA||• Migraineurs who were obese (1.3) and morbidly obese (1.7) were at increased risk for experiencing frequent migraine headaches (10–15 headache d/mo) compared with those who were normal weight.|
|Ford et al. (17)||7 601||≥20||48||Underweight (<18.5)||34.0||• Participants who were underweight (OR = 2.1) and obese (OR = 1.4) were at increased risk for having severe headaches or migraine compared with normal-weight participants.||NA|
|Normal weight (18.5–24.9)||18.9|
|Keith et al. (31)||220 370||16–94||100||NA||NA||• Migraine prevalence was not related to obesity.||NA|
|Mattsson et al.(32)||684||54 (40–74)||100||Non-obese (<30)||19.9||• Migraine prevalence was not related to obesity.||• Migraine frequency and severity was not related to obesity.|
|Peterlin et al. (18)||21 783||≥20||51||≤55 years old|| ||• In men and women ≤ 55 years of age, total obesity (based on BMI) and abdominal obesity (based on waist circumference) were associated with higher migraine prevalence.|
• In women > 55 years of age, abdominal obesity was associated with lower prevalence of migraine.
|>55 years old|| |
|Obese (≥30) ||NA|
|Winter et al. (33)||63 467||54 (≥45)||100||Underweight/normal weight (<23.0)||13.8||• Obesity was not associated with active migraine or prior history of migraine.||• Migraineurs who were underweight/normal weight (2.1) and severely obese (3.1) were at increased risk for having daily migraine compared with those with the lowest associated risk (overweight II, migraine frequency < 6 year−1).|
|Normal weight (23.0–24.9)||14.4|
|Overweight I (25.0–26.9)||15.2|
|Overweight II (27.0–29.9)||14.3|
|Severely obese (≥35.0)||16.4|
Bigal et al. (16) subsequently reported findings that further supported the relationship between obesity and migraine frequency. In this study, 18 968 of 162 576 (11.7%) individuals aged 12 years and older screened positive for migraine. Among these individuals, the occurrence of very frequent headaches was significantly higher in the obese (8.2%) and morbidly obese (10.4%) groups, compared with the normal-weight group (6.5%). Additionally, the percentage of participants with some headache disability was higher in the obese (38.4%) and morbidly obese (40.9%) groups, relative to the normal-weight group (32.0%).
Consistent with the findings of Bigal et al. (14), several recent studies have found no relationship between obesity and increased prevalence of migraine. Keith et al. employed 11 different datasets to examine the association between BMI and headache or migraine in more than 200 000 women (31). Findings showed that obese women had increased risk for headache, but not specifically migraine. Mattsson(32) evaluated the relationship between BMI and migraine in 684 women aged 40–74 years. Results showed that neither migraine prevalence nor frequency, severity and duration of headache attacks varied as a function of obesity status. Finally, Winter et al. (33) assessed the relationship between BMI and migraine in 63 467 Women's Health Study participants aged ≥ 45 years. Findings showed that migraine prevalence was higher in those with a BMI ≥ 35, although this association disappeared after adjustment for cardiovascular factors and postmenopausal status. However, increasing BMI among migraineurs was associated with more frequent headaches and clinical features.
By contrast, two studies conducted in National Health and Nutrition Examination Survey participants have shown an association between obesity and migraine prevalence. Participants were considered to have migraine if they answered ‘yes’ to having severe headaches or migraines during the past 3 months. In the first study, Ford et al. (17) evaluated 7601 participants and found that those who were underweight (<18.5) or obese (BMI ≥ 30) were at higher risk for having severe headaches or migraine compared with those who were normal weight. In the second study, Peterlin et al. (18) examined the relationship between severe headaches or migraine and both total (based on BMI) and abdominal (based on waist circumference) obesity in nearly 22 000 participants. In men and women aged 20–55 years, higher prevalence was associated with both total and abdominal obesity. However, in women older than 55 years, migraine prevalence was reduced in those with abdominal obesity.
The reasons for inconsistency in findings regarding the relationship between BMI and migraine prevalence are not entirely clear, but several methodological differences may underlie this variation. Whereas some of the studies used measured height and weight (17,18,32), others relied on self-reported height and weight (14,16,31,33), which may be problematic given that obese migraineurs have been shown to underreport their weight (34). There have also been a variety of methods used to define and diagnose migraine, thus creating difficulty in interpreting and comparing findings across studies. Finally, some studies were conducted exclusively in women of perimenopausal and postmenopausal age who have lower migraine rates than women in their reproductive years (35).
Based on the studies above, there appears to be a relationship between obesity and migraine, particularly in women during their reproductive years. While there is stronger evidence to suggest that obesity exacerbates migraine, it is less clear whether obesity increases risk for having migraine. Additional population studies that use direct measures of obesity and standardized International Headache Society criteria to diagnose migraine are needed to better address this important clinical and research question.
Possible mechanisms underlying the relationship between migraine and obesity
Several potential mechanisms may help to explain the linkage between migraine and obesity. These mechanisms can be classified into three groups (i) physiological; (ii) psychological and (iii) behavioural.
The relationship between migraine and obesity might be explained in part by common inflammatory mediators. Neurogenic inflammation resulting from activation of the trigeminal vascular system figures predominantly in the pain of migraine. Stimulation of trigeminal ganglion nociceptors induces the release of proinflammatory substances, most notably CGRP and substance P (26,28,29). In obese individuals, CGRP levels are elevated and show further increases after fat intake (36). Similarly, substance P has been identified in adipose tissue and may contribute to enlargement of fat depots and thus the proinflammatory milieu that occurs with obesity (37). Furthermore, levels of inflammatory adipocytokines, including tumour necrosis factor-α and interleukin-6, that increase with higher levels of adiposity (38) are elevated at the onset of migraine attacks (39). Finally, levels of C-reactive protein (CRP) that increase during systemic inflammation are elevated in both migraineurs and obese individuals (40,41). Thus, the inflammatory state that exists with obesity may exacerbate the inflammatory response in migraine, possibly contributing to headaches that are more frequent or severe (42).
Various neurotransmitters and peptides directed by the hypothalamus in regulation of eating behaviour also play a role in migraine pathophysiology. For example, the neurotransmitter serotonin binds to receptors (5-HT1A, 5-HT1B, 5-HT2A and 5-HT2C) involved in control of energy intake with the 5-HT1B and 5-HT2C receptors signalling satiety (43). Except for transient rises in serotonin levels during headache attacks, migraine is postulated to involve chronically low serotoninergic activity (43), which could contribute to increased caloric intake and weight gain (44,45). Serotonin further affects eating behaviour via various neuropeptides, including orexin-A. Along with its roles in appetite and feeding, orexin-A is also associated with reward, arousal, stimulation of spontaneous physical activity and modulation of glucose levels (46–48). Moreover, obese women have been shown to have lower plasma levels of orexin-A compared with controls (49), which holds particular relevance for migraine as orexin-A deficiency is postulated to promote inflammation in the trigeminal system (42). This notion is further supported by research using rat models that shows injection of orexin-A reduces perception of painful stimuli (50) and inhibits neurogenic vasodilation and release of CGRP from trigeminal neurons (51). Interestingly, another orexigenic connection between migraine and obesity may be sleep dys-regulation. Both disorders are common comorbid conditions in patients with narcolepsy (52,53), a sleep disorder characterized by orexin-A deficiency (54).
Adipocytokines such as adiponectin and leptin that regulate body weight through effects on metabolism and appetite may promote inflammatory processes underlying migraine and obesity. For example, low adiponectin levels are shown to coincide with a proinflammatory environment marked by elevated levels of tumour necrosis factor-α, interleukin-6 and CRP, and have been implicated in both obesity and migraine (55,56). Similarly, leptin may induce release of cytokines (57) and contribute to heightened pain sensitivity (58). Elevated leptin concentrations occurring with obesity are also associated with increased susceptibility to chronic inflammatory/autoimmune diseases (59). However, less is known about the role of leptin in migraine. A recent study found lower leptin levels in episodic migraineurs compared with controls, although this difference did not persist after adjusting for the migraineurs' lower fat mass (60). Consequently, further studies are needed to directly examine whether increased leptin levels in obese migraineurs contribute to alterations in cytokines and exacerbation of headache activity.
Finally, sympathetic dys-regulation plays a role in both migraine and obesity (61,62). Migraineurs are shown to have sympathetic hypofunction during asymptomatic periods and sympathetic hyperfunction during attack periods (62). Similarly, obesity is related to sympathetic hyperstimulation as well as reduced serotonin tone (42,61). Consequently, it has been postulated that as a result of sympathetic hypofunction, obese migraineurs may have difficulty in adapting to the elevated sympathetic tone associated with obesity, thus making them susceptible to an increased number of attacks (42).
Certain psychological factors and conditions may provide insight into the migraine–obesity relationship. For example, psychological stress affects both migraine and obesity. Stress can promote migraine onset in predisposed individuals (63), precipitate or aggravate headache attacks in those who already have migraine (64), and increase risk for transformation of episodic to chronic migraine (65). These effects are presumed to be the product of biochemical changes inherent to the physiological stress response that can increase sensitivity of the trigeminal system, promote neurogenic inflammation and reduce pain threshold (66). Additionally, the experience of migraine itself can be a stressor and through maladaptive coping strategies such as medication overuse may increase headache frequency (67). Similarly, studies have shown associations between psychological stress and increased risk of obesity and weight gain (68–70), which may be mediated by dys-regulation of the hypothalamic-pituitary-adrenal axis (71), greater preference for high-calorie, palatable foods (72), and reduced control over food intake (72). As previously described, such disturbances in hypothalamic function and eating regulation are thought to modulate the migraine–obesity relationship.
Certain psychiatric disorders related to obesity, including major depression and anxiety disorders, are also risk factors for developing migraine and experiencing frequent, disabling attacks (6,7). Tietjen et al. (73) recently examined the effects of anxiety and depression on the relationship between migraine and obesity in 721 migraineurs (88% female, 30% obese) from eight different headache treatment centres. Findings showed that depression and anxiety were most common in the obese migraineurs. Additionally, both depression and anxiety in this group were associated with higher headache frequency and disability, suggesting that these disorders may modify the relationship between obesity and migraine.
Behavioural risk factors represent another area of overlap between migraine and obesity. For example, disturbances in sleep behaviour, particularly short sleep duration, are more common in migraineurs than in non-migraineurs (74), are a common trigger factor for migraine attacks (75) and are associated with increased headache frequency and severity (76). Short sleep duration also appears to play a role in weight gain and obesity, particularly in children and young adults (77). Obstructive sleep apnoea, a common complication of obesity, is also associated with transformation of episodic to chronic migraine (78). Thus, it is possible that sleep disturbances in migraineurs may be compounded by obesity, thereby increasing risk for more frequent and/or severe attacks.
Certain dietary habits are important in both migraine and obesity. For example, skipping breakfast has shown to be associated with increased risk of migraine (79) as well as weight gain (80) and obesity (81). Similarly, irregular meal frequency is associated with migraine (82) and is shown to increase energy intake (83), possibly contributing to weight gain and obesity. Consumption of a high-fat diet, which is a major contributor to weight gain and obesity (84), has also been shown in migraineurs (85). Furthermore, particular consequences of a high-fat diet including elevated blood pressure and cholesterol levels may increase risk for having migraine and migraine progression (86).
Low physical activity may be another factor contributing to the co-occurrence of migraine and obesity. Research has shown that low physical activity levels are associated with obesity status (87), as well as higher migraine prevalence and more frequent migraine attacks (88,89).
Why weight loss might be effective for alleviating headaches in obese migraineurs
Given that weight loss affects many of the physiological, psychological and behavioural mechanisms proposed to link migraine and obesity, it may well prove to be a means of reducing headache frequency and severity in obese migraineurs. With respect to physiological mechanisms, weight loss could prevent or decrease the severity of migraine attacks by altering central nervous signalling pathways via increases in orexin-A (90), and through inflammatory pathways mediated in part by increases in adiponectin and decreases in leptin (91,92). Associated reductions in inflammatory cytokines (93), CRP (94) and sympathetic tone (95) may also play a role in altering the central and peripheral inflammatory cascades that lead to the pain of migraines.
Additionally, weight loss-related improvements in psychological factors including mood (96) and stress-coping ability (97) could help to alleviate migraine (42,67,73). Similarly, enhancement of sleep quality and disturbances that may occur with weight loss could reduce headache frequency and severity (98,99). Decreases in fat consumption and increases in physical activity are strongly correlated with weight loss (100,101), and both behaviours have shown to be associated with reduction in headache frequency and severity (85,102). Finally, weight loss may impact migraine in obese individuals via improvements in conditions that are comorbid to both disorders, such as hypertension and hyperlipidaemia (86,103), depression (73,96), diabetes (104,105) and sleep apnoea (78,106).
Weight loss and headache
Despite these plausible mechanisms linking obesity and migraine, few studies have examined the impact of weight loss on headache, and none has studied whether weight loss improves migraine in obese adults. Sugerman et al. (107) examined the effects of bariatric surgery on idiopathic intracranial hypertension, an obesity-related disease of which severe, persistent headaches is a symptom. Weight loss at 4 months post surgery was associated with resolution of headaches in all but 1 of 19 female patients; in addition, headaches recurred in two patients who had subsequent weight regain.
More recently, Hershey et al. (108) examined the effect of weight and weight change on various headache parameters in 913 paediatric headache sufferers (71% of whom who had migraine). BMI percentile was positively associated with headache frequency at the initial visit. Greater decreases in BMI were associated with greater reductions in headache frequency at 3- and 6-month follow-up for children who were initially overweight or obese, but not for those who were normal weight.
While it is difficult to draw any substantive conclusions from the studies above, the epidemiological and mechanistic links between migraine and obesity suggest that weight loss may be a viable treatment approach for alleviating headaches. In order to properly test the effects of weight loss on migraine, randomized controlled trials involving interventions that can effectively and reliably produce weight loss in obese individuals are needed.
Rationale for conducting randomized trials of behavioural weight loss interventions in obese migraineurs
There is strong justification for conducting randomized trials of behavioural weight loss interventions in obese migraineurs. Behavioural interventions have been demonstrated to be the most effective treatment for mild to moderate obesity, resulting in an average weight loss of 10 kg or approximately 10% of starting weight after 6 months of treatment (109). Moreover, large randomized controlled trials such as the Diabetes Prevention Program and Look AHEAD (Action for Health in Diabetes) have demonstrated that significant weight losses achieved through behavioural intervention can be maintained to produce long-term reductions in cardiovascular disease risk and diabetes incidence in patients of different ages, weights and ethnicities (103,110). Diabetes Prevention Program and Look AHEAD also provide strong examples of how behavioural treatment length, format and components can be standardized and administered across multiple sites. Finally, changes in eating and activity habits, body weight and disease risk factors that occur during behavioural interventions affect many of the previously discussed mechanisms proposed to link migraine and obesity.
Migraine is a prevalent and debilitating disorder. Evidence indicates a possible link between migraine and obesity, with the latter either exacerbating headache activity in current migraineurs or possibly increasing the risk for having migraine. A wide array of physiological, psychological and behavioural mechanisms may contribute to the co-occurrence of these disorders, the majority of which can be effectively targeted with weight loss. Therefore, conduct of randomized controlled trials to evaluate the effects of weight loss on migraine is warranted. Behavioural weight loss interventions, in particular, may provide a useful treatment model for showing whether modest weight loss reduces headache frequency and severity in obese migraineurs as well as enhancing understanding of different mechanisms through which weight loss might impact migraine.
Conflict of Interest Statement
Dr Bond is supported by National Institutes of Health Grant DK083438.