Abstract
In a recent paper (1) on ‘a prospective cohort study among 36,016 women without a history of depression enrolled in the Women's Health Study who provided information about migraine and headache at baseline’ the age- and multivariable-adjusted relative risk of depression according to headache status were: history of non-migraine headache 1.43 (95% CI 1.31–1.56), migraine with aura 1.51 (95% CI 1.33–1.71), migraine without aura 1.38(95% CI 1.24–1.54), and past history of migraine 1.53 (95% CI 1.35–1.74) (1).
In an older smaller study (n = 1.284) from 2000 the lifetime prevalence of major depression was higher in patients with aura (MA) than in patients with migraine without aura (MO). Odds ratios were 4.90 (95% CI 3.31–7.19) and 3.03 (95% CI 2.23–4.11), respectively, p = 0.013 (2). In the current study (1) the risk for depression was only slightly, non-significantly, higher for MA than for MO, see above. The authors mention that there are a few possible explanations for the difference in effect size between the two studies (1): (a) the age ranges of participants are different (1,2); (b) the current study relied on self-report of physician diagnosis of depression (1), whereas the other study (2) used a structured interview for the diagnosis of depression.
In addition, the diagnostic process classifying patients as suffering from MA was quite different in the two studies (1,2). The first study (2) used ‘detailed questions about the International Headache Society (IHS) defining features of migraine’, whereas in the second study (1) the response options for the participants included: ‘Aura or any other indications a migraine is coming’, which was used to classify women who experience MA (1). ‘Or any other indications a migraine is coming’ is most likely not aura but more likely to be premonitory symptoms, which are symptoms that precede the migraine attack by 2–48 hours (3). Premonitory symptoms occur in 33% to 87% migraine patients (3). By asking about ‘or any indication that a migraine was coming’ (1), the investigators most likely included some patients without aura but with premonitory symptoms in the MA group.
One could argue that an imprecise diagnosis of MA does not matter as long as an increased risk for depression is observed, ‘the signal is coming through’ as in this large study (1). However, there was no difference in risk for depression between MA and MO (1) when an imprecise diagnosis of MA was used. If MO patients in reality have less risk of depression than MA patients, inclusion of MO patients with premonitory symptoms in the MA group would decrease the chance of finding a difference between the two forms of migraine.
In conclusion, the current study (1) did not demonstrate the previously shown higher risk for MA than MO patients for depression (2), but as mentioned above this could be because of imprecise classification of MA. If possible, the data should be re-evaluated after more precise classification of MA and MO patients.