Association of Migraine and Blood Pressure—Does Obesity Severity Have a Moderating Role?

Abstract

Background

The relationship between migraine and blood pressure (BP) is equivocal, warranting exploration of potential moderators. Obesity associates with both migraine and BP in a dose-dependent fashion, although its role as a moderator has not been evaluated. We examined the relation between migraine and BP in women with comorbid migraine and obesity, and whether this relation was influenced by obesity severity.

Methods

Women with migraine and obesity (n = 134) completed a 28-day headache diary before randomization to lifestyle intervention or migraine education. BP (systolic (SBP)/diastolic (DBP)), body mass index (BMI), and waist circumference (WC) were measured before diary completion. Hierarchical linear regression assessed associations between BP and migraine characteristics (headache frequency, duration, and pain intensity), and obesity severity (both total (BMI) and abdominal (WC)) as moderators of these associations.

Results

Participants (BMI = 35.4 ± 6.5 kg/m²; WC = 105.4 ± 15.6 cm, SBP = 113.1 ± 12.1/DPB = 68.1 ± 8.0 mmHg) reported 8.4 ± 4.5 migraine days that lasted 20.2 ± 15.9 h with mean pain intensity of 5.9 ± 1.6 on a 10-point scale. DBP inversely related to migraine days in both total (β = −0.226, p = .010) and abdominal (β = 0.214, p = .015) obesity severity models. SBP and obesity severity did not relate to migraine characteristics. Obesity severity did not moderate relations between migraine characteristics and BP (p’s > .05).

Conclusion

Among women with comorbid migraine and obesity, DBP inversely related to migraine frequency; however, obesity severity did not affect the strength of this or other examined associations. Future studies including healthy weight controls and men and women with continuous BP measures are needed to confirm these findings and identify mechanisms and moderators.

Introduction

Migraine is a neurological condition characterized by recurrent headaches often accompanied by nausea, vomiting, and sensitivity to sensory stimuli (e.g., light and sound) [1]. Migraine is highly prevalent, affecting an estimated 15% of the global population [2]. Migraine impacts numerous aspects of patients’ lives and is especially common among women [3]. Hypertension, or high blood pressure, is also prevalent among US adults and is often comorbid with migraine [4]. Due in part to this comorbidity, there has been longstanding interest in the relationship between migraine and vascular diseases like hypertension.

Studies evaluating the relationship between blood pressure (BP) and migraine have yielded mixed findings, as highlighted in several reviews [5–7]. However, most of this work has been limited to comparisons of BP, and especially hypertension, in people with and without migraine in epidemiological studies [8, 9]. Fewer studies have focused on trying to understand how BP and migraine severity are related among individuals with diagnosed migraine, despite this information having potential to inform clinical care among migraine sufferers [10]. Results from these studies are also conflicting, with some studies showing that higher BP is associated with greater migraine prevalence and symptomatology (e.g., [11]), others suggesting that higher BP is associated with reduced migraine frequency and severity (e.g., [12, 13]), and still others finding no relationship (e.g., [10]). Multiple factors may contribute to these disparate findings, including differences in study population (e.g., age) [5] and a potentially complex relationship between migraine and BP (e.g., differential associations with diastolic vs. systolic BP) [14].

It is also possible that other variables impact the migraine-BP association and have contributed to equivocal findings to date. One underlooked factor that could improve understanding of the relationship between BP and migraine is weight status, and specifically degree of overweight. Obesity affects both BP and migraine symptoms in a dose-dependent manner: obesity severity relates to risk for elevated BP levels [15] and more severe migraine symptoms [16]. Thus, it could be useful to directly evaluate the role, if any, obesity severity plays in the relationship between migraine and BP. For example, if migraine severity and BP are positively related [11], obesity severity may strengthen this association given its positive association with both migraine severity and BP. In contrast, if migraine severity and BP are inversely related [12], obesity severity may attenuate this association. While some prior work on the migraine-BP link has controlled for weight status [11], research evaluating the potential moderating role of weight status is lacking.

The Women’s Health and Migraine (WHAM) study provides a unique opportunity to evaluate interrelationships of migraine severity, BP, and obesity severity among individuals with migraine. The WHAM study was a randomized trial that compared behavioral and educational interventions to reduce weight and headache symptoms among women with comorbid migraine and overweight/obesity. As part of the baseline assessment, patients had their BP and weight objectively measured, completed a 28-day electronic headache diary, and completed a brown bag medication review. Importantly, daily headache diaries overcome many retrospective reporting biases associated with one-time, self-report measures [17, 18]. Yet, these diaries have rarely been used in studies investigating the association between migraine and BP. In addition, not all studies have accounted for medication use [19, 20], despite several medications that are frequently used for migraine (e.g., beta-blockers) having an antihypertensive effect [21]. The WHAM data can evaluate whether obesity severity influences the association between BP and migraine severity in patients with migraine, while also addressing several methodological limitations of prior research.

This study aims to assess the associations between systolic and diastolic BP and migraine severity (as indicated by headache frequency (days/month and episodes/month), headache duration, and pain intensity) among a sample of women with migraine and varying degrees of obesity severity, and to evaluate whether obesity severity moderated these associations. Given inconsistent findings to date, we did not specify directional hypotheses.

Method

Details of the full study protocol are available in Bond et al. [22]. All aspects of the procedure relevant to this secondary analysis are described below. All study procedures took place at a research office building associated with an academic medical center.

Participants

Eligibility requirements for the WHAM trial included being assigned female at birth, being 18–50 years of age, having a BMI between 25.0 and 49.9 kg/m², and having migraine, as specified by International Classification of Headache Disorders 3 criteria and as confirmed by the study neurologist, with ≥ 3 migraine attacks and 4–20 migraine headache days/month over the past 90 days. Participants were excluded if they had a headache disorder other than migraine or tension-type migraine; were participating in another weight loss program; had lost ≥ 5% body weight within the past 6 months; were using weight loss medication; were diagnosed with cancer within the past year; had undergone bariatric surgery; were pregnant, breastfeeding, or expecting pregnancy during the trial period; were unable to read or understand study materials; or had another condition that would interfere with the study protocol (e.g., plans to move).

Procedure

Between November 2012 and June 2016, participants were recruited from community and clinical settings through direct mailing, website and social media postings, and newspaper advertisements. Interested individuals contacted the research center by phone and completed a phone screening to determine initial eligibility. Eligible individuals were then invited to an in-person orientation during which they received study details; provided written informed consent; had their migraine diagnosis verified by the study neurologist; had their height, weight, and blood pressure measured; completed questionnaires; and completed a brown bag medication review. Participants were also given a smartphone with a custom headache diary app for daily monitoring of headaches over the following 28 days (see details below). After this 28-day pre-treatment headache monitoring period, participants returned the smartphone and were randomized to a 16-week behavioral weight loss intervention for migraine or migraine education control. Participants for this study completed all measures of interest prior to randomization.

Measures

Headache Characteristics

Migraine headache characteristics were recorded for 28 days through a custom smartphone-based headache diary. At the end of each day, participants reported if they experienced a migraine headache that day (Y/N), which was used to determine the frequency of migraine headache days per month and number of headache episodes. Participants also reported peak severity of headache pain (0 = no pain, 10 = worst possible pain) on days headaches were experienced, which was averaged to determine a mean pain intensity rating. Furthermore, when headaches were experienced, participants were asked to record when the headache began and ended, which was used to determine headache duration. Diary completion was monitored daily by study staff and participants were contacted by phone to collect any missing data.

Blood Pressure (BP)

Blood pressure was measured at the baseline visit by trained research staff. Specifically, participants were seated in a private clinic room and made to feel comfortable, sitting in a chair with feet flat on the floor and back supported. After being fitted with the proper cuff, participants sat quietly for 5 min. Participants then underwent three automated BP measurements separated by 1-min intervals using the GE Dinamap Monitor Pro 100 oscillometric device. Participants’ baseline appointments, and thus their BP measurements, were scheduled during morning hours when participants were in a fasted state. Participants confirmed they had not smoked, used caffeine, or exercised after waking and prior to BP measurement. The multiple systolic and diastolic readings were averaged to create a mean systolic and diastolic BP value for analyses.

Anthropometric and Sociodemographic Characteristics

At the baseline visit, height was measured via a wall-mounted Harpenden stadiometer and weight was measured using a calibrated digital scale to determine BMI. Waist circumference was found at the midpoint between the highest point of the iliac crest and the lowest part of the costal margin in the mid-axillary line. Two measurements were taken for each anthropometric characteristic and were averaged for analyses. Questionnaires assessed age, race, ethnicity, and level of education.

Medication Usage

At baseline, a brown bag review was conducted, in which all participants were asked to bring in all of their medicines (prescription and over-the-counter) as well as vitamins, supplements, and herbal medicines, and a comprehensive medication list was recorded by research staff. These data were used to create a BP medication usage (Y/N) variable.

Analytic Approach

Data were analyzed in SPSS version 25. Critical ∝ was set at 0.05. Variables were characterized with mean and standard deviation or counts and percentages, as appropriate. Data were screened for missingness. Complete data were available for all variables of interest, including headache diary variables given that staff monitored diary completion daily and contacted participants by phone to obtain any missing data. Hierarchical linear regression was used to assess the relationship between BP and migraine characteristics, as well as the moderating effect of obesity severity. Each of the four migraine variables of interest (days, episodes, duration, pain intensity) was regressed on diastolic and systolic BP separately; all models also included BMI or waist circumference and covariates. The first step of each hierarchical regression included the following variables: BP (diastolic or systolic), obesity severity (BMI or waist circumference), and use of BP medication, and sociodemographic variables (age, race, ethnicity). The second step included the interaction term between the respective BP measure and obesity severity measure. All variables were centered prior to creating interaction terms. The sample size was determined through a power analysis for the parent trial [22], which was designed to detect between-group differences in migraine days/month at posttreatment with 0.80 power.

Results

One hundred thirty-four participants were included in analyses (see Supplemental Materials for a flow diagram). Mean BMI was 35.4 (SD = 6.5) kg/m², mean waist circumference was 105.4 (15.6) cm, mean diastolic and systolic BP were 68.1 (8.0) and 113.1 (12.1) mmHg, respectively, and mean migraine headache characteristics were as follows: 8.4 (4.5) headache days/month, 5.4 (2.8) headache episodes/month, 20.2 (15.9) hours duration, and 5.9 (1.6) pain intensity. Mean participant age was 38.4 (SD = 8.3) years. Twenty-five (18.7%) participants identified as Hispanic and 109 (81.3%) identified as non-Hispanic; 15 (11.2%) identified as Black, 102 (76.1%) as White, 4 (3.0%) as mixed race, and 13 (9.7%) as “other.” Fifteen (11.2%) participants were taking BP medication.

Table 1 shows a summary of findings across models; full model results, which include values for model covariates, are available in Supplemental Materials. Of note, BP medication use related to more headache days/month and longer headache duration in all models (p’s < 0.05). No other sociodemographic variables nor obesity severity (as measured by both BMI and waist circumference) related to any migraine severity variables of interest (p’s > 0.05).

Table 1.

Summary of findings from hierarchical regression models evaluating diastolic BP, systolic BP, and their interaction with obesity severity (as measured by both BMI and waist circumference) as predictors of migraine severity

	Related to headache days?	Related to headache episodes?	Related to headache duration?	Related to pain intensity?
BMI models
Diastolic BP	Yes	No	No	No
	− .13 (.06) *	− .06 (.03)	− .30 (.18)	− .01 (.02)
Diastolic BP × BMI	No	No	No	No
	− .002 (.01)	< .001 (.004)	− .003 (.03)	− .001 (.003)
Systolic BP	No	No	No	No
	− .07 (.04)	− .02 (.02)	− .29 (.13)	− .02 (.01)
Systolic BP × BMI	No	No	No	No
	− .002 (.004)	− .001 (.003)	− .01 (.02)
WC models
Diastolic BP	Yes	No	No	No
	− .12 (.05) *	− .06 (.03)	− .29 (.18)	− .01 (.02)
Diastolic BP × WC	No	No	No	No
	.001 (.003)	< .001 (.002)	.01 (.01)	< .001 (.001)
Systolic BP	No	No	No	No
	− .05 (.04)	− .02 (.02)	− .22 (.13)	− .01 (.01)
Systolic BP × WC	No	No	No	No
	− .001 (.002)	− .001 (.001)	− .003 (.01)	− .001 (.001)

BMI body mass index, WC waist circumference, BP blood pressure

^*p < .05. All values shown are b (SE) and are from the final step of the hierarchical regression model. Statistically significant findings are bolded

As shown in Table 1, diastolic BP was inversely related to headache days/month in both the BMI and waist circumference models, such that every 1 mmHg increase in diastolic BP corresponded to a ~ 0.22 day decrease in headache days/month (p = 0.010 and beta = −0.226 in BMI model; p = 0.015 and beta = 0.214 in waist circumference model). Obesity severity did not moderate the associations of diastolic BP with headache days (p’s > 0.05).

Diastolic BP was not related to migraine episodes per month, migraine duration, or pain intensity in either the BMI or waist circumference models (p’s > 0.05). The interactions between diastolic BP and obesity severity (both BMI and waist circumference) in these models were also non-significant (p’s > 0.05) Additionally, systolic BP was not related to any headache characteristics of interest as an independent predictor or when considered with obesity severity (both BMI and waist circumference) as a moderator (p’s > 0.05).

Discussion

Despite longstanding interest in the association between BP and migraine, research in this area remains largely inconclusive [5–7]. In addition, most prior research has focused on risk of hypertension among individuals with vs. without migraine, rather than trying to understand how BP and migraine severity are related among those with diagnosed migraine, despite the clinical utility of this information for migraine sufferers. This study assessed associations among BP and daily diary-assessed migraine severity among women with comorbid obesity and migraine and was the first to our knowledge to evaluate whether obesity severity moderates these associations. While diastolic BP was inversely related to migraine frequency, BP was otherwise unrelated to migraine severity in this sample. Further, obesity severity did not impact these associations. Although replication with a sample that contains individuals across the full weight spectrum is needed, these findings contribute to the mixed literature on the BP-migraine link and provide preliminary indication that obesity severity may have a limited role in explaining past disparate findings.

Greater diastolic BP was related to fewer headache days per month. In terms of the clinical significance of this finding, each 1 mmHg increase in diastolic BP corresponded to approximately 1/5 fewer headache days per month. This finding aligns with some prior work showing that BP is lower among individuals with migraine [19] and is inversely related to headache frequency [12]. Notably, however, some studies have only observed an inverse relationship between systolic BP and migraine [23]. Inconsistencies in findings could owe in part to differences in study methods and populations (e.g., assessment with daily diaries vs. retrospective self-reports; adult vs. adolescent samples) [24]. As to why an inverse relationship may exist, individuals with higher BP may have an elevated response threshold to unpleasant stimuli, causing diminished perception of headaches (a phenomenon termed hypertension-associated hypalgesia) and lower reported symptomatology. Given that BP was assessed continually rather than with clinical hypertension thresholds in this study and the inconsistent findings on these associations to date, more research is needed to understand whether and why BP relates to headache characteristics in individuals with migraine. Research that assesses daily associations between BP and headache characteristics over time is also needed to determine the causal nature of this relationship, if any.

Although diastolic BP related to headache days, BP was not related to the other headache outcomes evaluated in this study. Given the dearth of research on the association between BP and headache characteristics among patients with migraine, additional research on this topic is needed that uses larger, more representative samples (e.g., with men and normal weight controls). However, the current study’s pattern of findings raises the possibility that BP is not meaningfully related to migraine severity, even if migraine and BP are linked from an epidemiological perspective as multiple past studies would suggest [8, 11, 12].

Analyses examining whether weight status moderated the association between BP and migraine severity—which could offer a potential explanation for some of the variability in past research on this topic—were non-significant. In contrast to prior work [16], obesity severity was also unrelated to migraine severity in this sample. Although the BMI in this sample ranged from overweight to class III obesity, there was not a normal weight control group. It is possible that this somewhat limited range in weight status hindered the ability to detect a potential association between obesity severity and migraine severity, as well as potential interactions among obesity severity and BP. Alternatively, obesity severity may serve as an independent risk factor for migraine but not combine with BP in a predictable fashion to exert influence on migraine. Future research that includes a healthy weight control group is needed to draw firmer conclusions about the interrelationships among migraine severity, BP, and obesity severity.

Strengths of the current study include use of daily diary-assessed headache data in conjunction with objectively measured BP data (the first study to our knowledge to do so), neurologist-confirmed migraine diagnosis and objective measurement of indices of obesity severity, statistical adjustment for use of BP medication and sociodemographic characteristics, and examination of the research questions in a sample at elevated risk for both hypertension and more severe migraine due to the presence of comorbid migraine and obesity. Limitations include a somewhat restricted range in BMI, which may have influenced our ability to detect associations; measurement of BP on only one occasion, which precluded more time-sensitive analysis of the association between BP and migraine severity, and in a research center, which may have influenced BP readings (although steps were taken to mitigate any such effects); a somewhat limited sample size, which may not have provided adequate power to detect smaller effects; use of an all-female and relatively racially and ethnically homogenous sample, which may limit generalizability; and assessment of the relationships of interest at a single time point, which precludes causal conclusions. Future studies are thus needed that continue to explore the potential interrelationships among migraine severity, hypertension, and obesity among patients with migraine in larger and more diverse samples with repeated measurements of both BP and migraine characteristics [10] over more extended time periods.

In conclusion, this study adds to current literature on migraine and BP by demonstrating presence of an association between diastolic BP and headache frequency, and by suggesting that obesity severity may have limited bearing on the BP-migraine link, although replication is needed. Given the significant public health burden of migraine, hypertension, and obesity, as well as projected increases in obesity prevalence [25] that are likely to increase the number of migraine sufferers with both elevated BP and obesity, further research is warranted to clarify the nature of the relationships among these three conditions.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s12529-023-10241-7.

Data Availability

Data are available from the authors upon reasonable request and with the permission of the supporting Institutional Review Board