Serum lipid abnormalities in migraine: A meta-analysis of observational studies

The association of migraine with vascular comorbidities is long-established. The contribution of the "traditional" cardiovascular risk factors to this connection remains unclear.

To determine-quantify the differences in the serum lipid concentrations between lipid-lowering agents-naïve individuals with migraine and healthy controls (HC).

The study protocol was not preregistered with an online systematic review-protocol registry. A literature search involving MEDLINE, EMBASE, CENTRAL, Google Scholar, and the OpenGrey database was performed. Case-control, cross-sectional, or cohort studies involving HC and participants with migraine (with and without aura regardless of the use of prophylactic treatment) that quantitatively assessed serum low-density lipoprotein cholesterol (LDL-C) (primary index) and/or total cholesterol (TC) and/or high-density lipoprotein cholesterol (HDL-C) and/or triglycerides (TG) (secondary indices) were retrieved. Articles including participants with known dyslipidemia (or under lipid-lowering medications) or with secondary causes of dyslipidemia (aside from the subjectively assessed lifestyle parameters) were excluded. Studies with abstracts and full texts not published in English and articles reporting the implementation of other study designs (reviews, meta-analyses, commentaries, case reports, etc.) were excluded as well. Conference abstracts and English abstracts from studies with full texts not published in English were evaluated as part of the gray literature. Each step of the review process was performed by two investigators independently, and relevant data were abstracted based on standardized extraction forms. Any discrepancies were resolved by a third investigator.

Seventeen studies (16 case-control and 1 cross-sectional) fulfilled the eligibility criteria. Retrieved articles involved adult participants, principally during the fourth decade of life. Results were compatible with higher LDL-C levels in migraine individuals (1370) than in HC (1215) [12 studies, mean difference (MD) = 10.4 mg/dl, 95% confidence interval (CI) = (1.6, 19.2)]. Similarly, higher TC levels were determined in migraine patients [14 studies, migraine = 1325, HC = 1213, MD = 10.6 mg/dl, 95% CI = (1.8, 19.3)], as were TG levels [15 studies, migraine = 1526, HC = 1262, MD = 11.8 mg/dl, 95% CI = (3.6, 20.0)]. HDL-C concentrations were not different between the two groups [14 studies, migraine = 1488, HC = 1328, MD = -0.4 mg/dl, 95% CI = (-2.2, 1.5)]. Prespecified sensitivity analysis following the exclusion of studies not presenting comparable body mass index values between the groups nullified the significant difference regarding LDL-C levels [MD = 5.3 mg/dl, 95% CI = (-0.1, 10.8)]. Subgroup analyses as well as the direct comparison of migraine with aura and migraine without aura individuals were compatible with no difference regarding lipid concentrations, but only a small fraction of the retrieved studies presented relevant figures.

Although our results are of limited generalizability, since most retrieved studies were performed in Turkey (nine studies), TC abnormalities may provide part of the explanation for the unfavorable cardiovascular profile of migraine patients. Lifestyle may be partly or entirely accountable for the determined increased serum TC. Additional studies that will completely address the effect that lifestyle parameters exert on lipid concentrations are required to better capture existing abnormalities.