Evidence that blood–CSF barrier transport, but not inflammatory biomarkers, change in migraine, while CSF sVCAM1 associates with migraine frequency and CSF fibrinogen

OBJECTIVE: Our objective is to explore whether blood–cerebrospinal fluid (CSF) barrier biomarkers differ in episodic migraine (EM) or chronic migraine (CM) from controls. BACKGROUND: Reports of blood–brain barrier and blood–cerebrospinal fluid barrier (BCSFB) disruption in migraine vary. Our hypothesis is that investigation of biomarkers associated with blood, CSF, brain, cell adhesion, and inflammation will help elucidate migraine pathophysiology. METHODS: We recruited 14 control volunteers without headache disorders and 42 individuals with EM or CM as classified using the International Classification of Headache Disorders, 3rd edition, criteria in a cross‐sectional study located at our Pasadena and Stanford headache research centers in California. Blood and lumbar CSF samples were collected once from those diagnosed with CM or those with EM during two states: during a typical migraine, before rescue therapy, with at least 6/10 level of pain (ictal); and when migraine free for at least 48 h (interictal). The average number of headaches per month over the previous year was estimated by those with EM; this enabled comparison of biomarker changes between controls and three headache frequency groups: <2 per month, 2–14 per month, and CM. Blood and CSF biomarkers were determined using antibody‐based methods. RESULTS: Antimigraine medication was only taken by the EM and CM groups. Compared to controls, the migraine group had significantly higher mean CSF–blood quotients of albumin (Q (alb): mean ± standard deviation (SD): 5.6 ± 2.3 vs. 4.1 ± 1.9) and fibrinogen (Q (fib) mean ± SD: 1615 ± 99.0 vs. 86.1 ± 55.0). Mean CSF but not plasma soluble vascular cell adhesion molecule‐1 (sVCAM‐1) levels were significantly higher in those with more frequent migraine: (4.5 ng/mL ± 1.1 in those with <2 headache days a month; 5.5 ± 1.9 with 2–14 days a month; and 7.1 ± 2.9 in CM), while the Q (fib) ratio was inversely related to headache frequency. We did not find any difference in individuals with EM or CM from controls for CSF cell count, total protein, matrix metalloproteinase‐9, soluble platelet‐derived growth factor receptor β, tumor necrosis factor‐alpha, interferon‐gamma, interleukin (IL)‐6, IL‐8, IL‐10, or C‐reactive protein. CONCLUSIONS: The higher Q (alb) and Q (fib) ratios may indicate that the transport of these blood‐derived proteins is disturbed at the BCSFB in persons with migraine. These changes most likely occur at the choroid plexus epithelium, as there are no signs of typical endothelial barrier disruption. The most striking finding in this hypothesis‐generating study of migraine pathophysiology is that sVCAM‐1 levels in CSF may be a biomarker of higher frequency of migraine and CM. An effect from migraine medications cannot be excluded, but there is no known mechanism to suggest they have a role in altering the CSF biomarkers.