The present study compared treatment-naïve patients with migraine to age/sex matched controls and found that patients with migraines had a pro-inflammatory state (higher CRP levels), an altered endothelial profile with lower levels of vasodilator factors (nitrate, nitrite, and isoprostanes), and worse vascular and pro-coagulant profiles (HDL cholesterol, vWF antigen and activity, and CIMT) compared with controls.
An association between CRP and migraine has been previously demonstrated in four case-control studies [25–28]. Elevated CRP levels in migraine have also been previously demonstrated during ictal or interictal periods [5,22,25–30], suggesting that inflammation is present in both scenarios and might be a key factor in migraine pathophysiology, although no conclusive evidence of its relationship with aura status exists [16,23]. This pro-inflammatory state has also been associated with higher levels of vascular disease biomarkers insinuating a relationship between migraine and endothelial dysfunction [5,25,26,28,30]. Alternatively, some studies have found no differences in CRP levels between migraine and control groups [31–33].
There is no consistent evidence on whether vascular biomarkers such as cholesterol levels differ between migraine patients and healthy individuals [30]. Considering that migraines have been associated with an increased risk of stroke, mainly ischemic stroke [7,9–11,13,16], and that ischemic stroke is related to cardiovascular risk, our findings of lower HDL cholesterol levels in patients with migraine compared with the control group, suggests an increased risk of vascular diseases in migraine. A recent metabolomics study proposed that patients with migraine have altered HDL cholesterol metabolism [34]. Another study found a positive association between metabolic syndrome and low HDL cholesterol in patients with MA [14], but a more recent study found no association between HDL cholesterol levels and cardioprotection [35].
In the literature, the migraine-biomarker association appears to be influenced by migraine frequency and subtype, with MA demonstrating a stronger association than M0 [6,36,37]. In the present study, significant differences in some biomarkers were observed between patients with MA and those with M0. A worse vascular profile was seen in patients with MA, with higher levels of triglycerides, platelets and fibrinogen, and lower levels of HDL cholesterol, compared with those with M0. No differences in CRP, vWF activity, or CIMT were seen across the two subtypes. On the other hand, some vasodilator factors (nitrite) were increased in patients with MA. This may have occurred due to a compensatory effect. We found a correlation between migraine severity and biomarker levels, with the most important biomarker alterations observed in patients with high frequency and chronic migraine (independent of aura status), unlike findings from previous studies [36].
We found vWF activity and CIMT significantly increased in the chronic migraine group compared with controls at T0. vWF is associated with cardiovascular risk factors as well as ischemic stroke [38] and is a well-established marker of endothelial activation [30]. Xiang and colleagues suggest that vWF regulation may serve as a powerful therapeutic target in treating thrombotic diseases such as stroke or myocardial infarction [39]. However, differences in CIMT measurements in patients with migraine are conflicting. Wang and colleagues found that CIMT was increased in patients with migraines compared with controls, independent of aura status [40]. Other studies have found that patients with migraine [26,32,33,36,37], chronic migraine [29], women [33], and those with MA [36] have elevated CIMT. Our results support the idea that patients with migraine are likely to have vascular dysfunction, atherosclerosis and higher cardiovascular risk. In contrast, some studies did not find any difference in CIMT levels between patients with migraine and controls [41–43].
Regarding the influence of sex on biomarkers, we found a pro-inflammatory and pro-oxidative profile in women with migraine, with higher levels of CRP, fibrinogen, platelets and TOS in women compared with men, whereas HDL cholesterol levels were higher in women than in men, suggesting an atheroprotective effect and reduced vascular risk. However, comparisons between women and men in our study should be interpreted with caution, given the small sample size of men (n = 11).
Increasing evidence points towards the role of oxidative stress in migraine [44,45]. Previous studies provide evidence for the role of oxidative stress and altered metabolism in migraine pathophysiology, and highlight it as a suitable therapeutic target [13], although discrepancies do exist [33,46]. According to our results the only statistically significant difference in oxidative stress biomarkers between the migraine and control groups at T0 were for nitrate and isoprostane levels, possibly due to the small sample size. Additionally, we observed higher levels of TOS and TBARS, and lower levels of thiol in patients with migraines, however, not significantly so. Previous studies have found decreased interictal concentrations of nitric oxide metabolites in patients with migraines [32], especially in those with MA [5]. Endothelial dysfunction is characterized by a reduction in the bioavailability of vasodilators and the consequent impairment in the reactivity of vasculature. It is characterized by a decrease in bradykinin-mediated endothelial nitric oxide synthesis [5]. Isoprostanes play an important role in transducing the effects of metabolic and hemodynamic abnormalities on platelet and vascular smooth muscle cell activation [47]. Increasing evidence suggest isoprostanes are involved in the complex interplay underlying cardiovascular homeostasis between low-grade inflammation, lipid peroxidation, and platelet activation [48]. To date there are few clinical studies aimed at evaluating isoprostanes or other arachidonic acid derivatives in chronic pain [49] and specifically in migraine [50]. Helmersson and colleagues found no differences in isoprostane urine rates between ictal and interictal periods in 21 patients with migraine [51]. Another study found prostaglandin E2 (PGE2) serum levels were significantly lower in patients with migraines compared with controls. Moreover, low serum levels of PGE2 have been positively correlated with headache frequency [52]. With the present study, we wish to highlight that nitrates and isoprostanes may be useful biomarkers of migraine severity and treatment response.
Gross and colleagues observed that one-third of patients with migraines had abnormally low thiol values [35]. Other studies have found serum TOS values to be significantly higher in patients with M0 than in controls [44,53], whilst Eren and colleagues only observed low thiol levels in patients with migraine compared with controls [54]. Togha and colleagues observed that patients with chronic migraines had lower total non-enzymatic antioxidant capacity (catalase, superoxide dismutase) and higher oxidative stress (nitric oxide, malondialdehyde) compared with patients with episodic migraines or healthy controls [55]. In the same way, Lucchesi and colleagues found lower thiol levels in patients with chronic migraine and medication-overuse headache compared with healthy controls [56]. In the present study, the differences between CRP, CIMT, vWF antigen, vWF activity, and nitrate levels had a positive correlation with migraine severity. On the other hand, TBAR, TOS and thiol peaked in participants with low migraine frequency (4–9 headache days per month) and normalized as migraine severity increased. This may be due to a compensatory mechanism between pro- and antioxidant factors.
After 3 months of prophylactic treatment all endothelial and oxidative parameters had improved, except for HDL cholesterol. HDL cholesterol levels decreased in metoprolol recipients (and were unchanged in topiramate recipients); in both instances levels were < 60 mg/dL. Although HDL cholesterol levels ≥ 60 mg/dL are desirable in clinical practice, to date, no clinical trials have determined specific target levels. Of note, low HDL cholesterol levels are associated with residual risk in atherosclerotic cardiovascular disease patients [57]. Metoprolol has been associated with a decrease in HDL cholesterol in previous short and long-term studies [58]. One hypothesis for this effect is that its indirect inhibition of lipoprotein lipase may lower HDL cholesterol plasma levels [58].
Differences in nitrate and nitrite levels were statistically significant in patients with migraines that responded to prophylactic treatment versus non-responders to metoprolol or topiramate.
CRP and prothrombin activity (in the low frequency subgroup) were the only endothelial parameters that showed statistically significant improvements in this study; there were no differences in PI between T0 and T1. A previous study assessing endothelial dysfunction by measuring arterial stiffness by pulse wave velocity, found an improvement in endothelial dysfunction after 1 month of prophylactic treatment (propranolol, topiramate or flunarizine), suggesting that a reduction in the frequency of migraines facilitates improvement in endothelial function [20].
We observed topiramate/metoprolol treatment-related reductions in oxidative damage factors such as TOS and TBARS, and an increase in antioxidant activity (thiol) and vasodilator bioavailability (isoprostanes and nitrates). Other migraine treatments have also been found to have antioxidant properties or limit oxidative stress, including flunarizine [17], BoNT/A [19], and anti-calcitonin gene-related peptide treatment [18].
To the best of our knowledge, this is the first study that assesses both endothelial dysfunction and oxidative stress biomarkers after 3 months of prophylactic treatment in patients with migraines. However, our study also has certain limitations. Our sample size was relatively small. Moreover, despite the cases being consecutively enrolled, selection bias was possible. Factors such as diet and physical activity, which can alter oxidative stress levels, were uncontrolled and blood samples were not taken after night fasting. Additionally, the assessed biomarkers may be surrogate markers of changes in the brain. Another limitation may have been our short duration of follow-up, which may not have been long enough to detect other differences. It would be beneficial to increase the sample size, and the duration of follow-up (to 6 or 12 months), in future studies.