Changes of migraine aura with advancing age of patients

Aim Given the similar presentation of migraine aura and acute ischemic stroke, advancing patient age might change the characteristics of migraine with aura (MA) and be clinically important. Clinical data, however, are limited. Experimental studies indicate a decrease in the magnitude of cortical spreading depression (CSD), the pathophysiological correlate of migraine aura, with advancing age. Our study aimed to assess the influence of age on the clinical features of MA. Methods Three hundred and forty-three patients were interviewed using a structured questionnaire. The questions covered the headache characteristics and symptom types including the characteristics of the C-criterion, as defined by the International Classification of Headache Disorders 3rd Edition. The association of age with MA characteristics was assessed. Results The median age was 29 (IQR 28–52) and 235 of the 343 patients were women (69%). Individual symptoms of the C-criterion such as gradual aura spreading over longer than 5 min (P < 0.001), two or more aura symptoms occurring in succession (P = 0.005), duration of at least one MA symptom for longer than 60 min (P = 0.004), and associated headache (P = 0.01) were more frequent in younger patients. The number of symptoms (P = 0.003) including the C-characteristics decreased with increasing age (P < 0.027). Patients with sensory (P < 0.001), motor (P = 0.04) and speech disturbance (P = 0.02) were younger, and older patients with headache had less photophobia (P = 0.04) and phonophobia (P = 0.03). Sensitivity analyses yielded similar results. Conclusion The frequency of typical characteristics of migraine aura and migraine headache including photophobia and phonophobia decreases with advancing patient age. This might have potentially difficult implications for the diagnosis of MA in the elderly. Supplementary Information The online version contains supplementary material available at 10.1186/s10194-023-01642-w.


Introduction
Migraine with aura (MA) symptoms are defined by the current diagnostic criteria [1] and include, for the aura, slow development, duration of 5-60 min, complete reversibility, positive irritative focal neurological symptoms, and accompaniment by headache within one hour of symptom onset.However, clinical observations showed that some patients with migraine aura do not have headache [2][3][4] and that an individual aura symptom might last longer than the 60 min allowed by the current criteria [5,6].The canonical cortical spreading depression (CSD) represents a slowly occurring and spreading depolarization across the cortex that is assumed to be the pathophysiological explanation for migraine aura [7].Although CSD fits many of the clinical features of MA, it does not explain the full picture, especially not the clinical variability.
Experimental studies in rats showed that older age correlated negatively with CSD propagation [8].However, there are limited clinical data on the influence of age on MA symptoms in humans.Aiba et al. [2] report a biphasic age-distribution of MA without headache.Donnet et al. [9] found that new onset MA in patients over 50 was not associated with headache or was accompanied more often by headache with non-migrainous features, and Hansen et al. describe less nausea with advancing age [10].By contrast, Viana et al. did not find age to be a significant factor associated with prolonged aura in univariate analysis [11].There is limited information on the influence of age on the individual symptoms of migraine aura.
The aim of this study was to assess the impact of age on MA symptoms.We hypothesized that the number of typical migraine features in MA decreases with advancing age.

Methods
This monocentric cross-sectional study was approved by the local ethics committee (2018-02258).The methods were previously reported [12].In short, patients with MA diagnosed according to the International Classification of Headache Disorders 3 rd edition (ICHD-3) [1] were interviewed using a structured questionnaire.The questions covered the type of symptoms, temporal aspects such as symptom onset and duration, the succession of symptoms, headache characteristics of a usual MA attack, as well as co-morbidities.Inclusion criteria for the study were verified diagnosis of MA, age more than 18 years and last migraine attack less than 12 months prior to the interview.

Statistical analyses
Frequencies are presented as counts, crude odds ratios and confidence intervals, calculated using Wilson's method.Continuous variables are presented as medians with interquartile range (IQR).For univariate comparisons of categorical variables x 2 or Fisher's exact were used, as appropriate.For comparisons of continuous variables, the Mann-Whitney or Kruskal-Wallis test was used, as appropriate.For multiple comparisons, Bonferroni correction has been performed.A P value of < 0.05 was considered significant.Further, the association between the number of symptoms and age was calculated.Only visual, sensory, motor and speech symptoms were considered, given their well-defined cortical distribution.Given that the influence of acute or preventative therapy on headache was not assessed, headache frequency, intensity and duration were not included in the analysis.Specifically, we looked at the temporal succession of headache and aura, headache location, headache character, aggravation of headache by routine physical activity, nausea, photo-and phonophobia.

Sensitivity analyses
As sensitivity analyses univariate comparisons between patients younger and older than 60 years were performed.Furthermore, the interaction between C-characteristics, age per 10 years and sex was assessed using multivariable logistic regression.To assess the goodness of fit of the prediction model, the Pearson x 2 test was performed.Additionally, we performed post-hoc power analyses for all outcomes with a significance level of alpha of 0.05.For outcomes consisting of comparisons between two groups, we used the mean differences, for three or more groups the one-way ANOVA test.A power of at least 0.8 was considered sufficient.The statistical analyses were performed in STATA/MP 16.0, StataCorp LLC.

Results
Three hundred and forty-three patients were interviewed.The median age was 29 (IQR 28-52) and 235 of the 343 patients were women (69%).The interviews took place after a median of 28 days (IQR 7-118) after the last MA attack.For an overview of the association of each C-characteristic and age see Fig. 1.

C2-characteristic of the C-criterion-two or more aura symptoms occur in succession
The C2-characteristic of the C-criterion was fulfilled in 201/343 (59%).Those with fulfilled C2-characteristic were

Number of symptoms
There were 134 patients (39%) with one, 87 (25%) with two, 84 (24%) with three and 38 (11%) with four symptoms.Patients with one (median age 42 years [IQR ).The four groups differed significantly from each other (P Bonferroni = 0.003, Fig. 3).When comparing the subgroups for post-hoc analysis, there was a significant difference in age between patients with one and three symptoms (P = 0.0007), one and four (P = 0.007) and two and three symptoms (P = 0.02) but no difference between those with two and four (P = 0.07) and three and four symptoms (P = 0.89).Of the patients without visual symptoms (n = 17), all (100%) had sensory, three (18%) had motor, and 7 (41%) had speech disturbance.

Sensitivity analyses
The differences in C-characteristics were confirmed in our sensitivity analysis comparing people younger to older than 60 years old (Table 2), even after adjustment for sex (Tables 2 and 3).
Looking at the differences in headache characteristics between patients younger and older than 60 years (Table 4), the significances for photophobia and phonophobia remained.However, a higher proportion of younger patients reported pulsating headache (45% vs 25%, P = 0.01) and aggravation by routine physical activity (77% vs 56%, P = 0.002).There was no difference for nausea (66% vs 54%, P = 0.09).
With regard to the analyzed outcomes, the post-hoc analyses showed power levels of 1 for all C-characteristics, 0.95 for number of C-characteristics and 0.96 for number of symptoms (Supplemental Tables S1, S2 and  S3).
The goodness of fit analyses with one degree of freedom showed P values higher than the predefined significance level of 0.05 and therefore the adjustment model for the logistic regression could not be rejected (Supplemental Table S4).

Discussion
We analyzed the influence of age on the MA characteristics.The main findings of our study are: a) patients with gradually spreading symptoms over 5 min, symptoms occurring in succession, and symptoms lasting longer than 60 min were younger, b) patients with a higher number of aura symptoms were younger, c) patients with lack of headache were older and the headache, when present, was less often accompanied by  The gradual onset of symptoms and symptoms occurring in succession are thought to be clinical expressions of the canonical CSD [1].In this context, our results in a large group of patients fit the experimental findings in animals of a negative impact of older age on CSD propagation [8].Similarly, the longer duration of aura symptoms in younger patients may correspond to a higher susceptibility to, or a longer duration of, CSD.It is plausible that the lower number of symptoms reported by older patients might reflect a smaller area of the cortex affected by CSD.
The association of older age with a lack of headache has been reported in previous studies [3,4,9,13].However, the precise reason for this association is not yet fully understood.The less intense CSD with advancing age might, at least in part, explain this finding.Howit is still debated which role CSD plays in the generation of migraine headache in patients with MA [7].Curiously, with advancing age, less photo-and phonophobia were reported by those who did have headache with their aura, with patients younger than 60 years having more often pulsating headache and aggravation by routine physical activity.In our study, there was no significant difference in the prevalence of nausea based on age.However, it is worth noting that a nonsignificantly higher proportion of patients younger than 60 years experienced nausea with their episodes, in-line with previous reports [9,10].
Overall, with advancing age, the MA tends to exhibit fewer typical features, and this has significant clinical implications.Considering that older patients are more prone to having vascular risk factors, there is a risk of misdiagnosing a migraine aura as an ischemic event, particularly if it occurs for the first time.The consequence may be unnecessary ancillary examinations and secondary prevention treatments.Conversely, an actual ischemic event might present with migraine-like symptoms, making it challenging or even impossible to differentiate between the two without further ancillary investigations [14][15][16][17].
In our study, we observed that patients with visual symptoms tended to be older.Therefore, even with advancing age, their presence should argue in favor of a migraine aura diagnosis.As previously discussed by Russel & Olesen, aura with sensory, motor and/or speech symptoms without visual symptoms should prompt suspicion of alternative etiologies [13].Nevertheless, younger patients reported non-visual symptoms more frequently, which might make them less useful in the clinical differential diagnosis of MA in general.This is of particular importance given that younger patients are more likely to present with migraine-like symptoms during an ischemic stroke as well [18].
In our study, we included patients who met the ICHD-3 criteria for definite MA.Given the tendency of MA to become less typical for migraine aura with advancing age, it is possible that some cases of MA in older patients might have been overlooked.
Our study has limitations.First, we did not assess the intra-individual variability of the MA attacks.Second, the interview took place after a median of 28 days after the last attack with a large interquartile range (7-118), making our results susceptible to recall bias.However, patients with MA experience many attacks during their lifetime, which might reduce this problem.We did not assess whether the patients took treatment for their headache and therefore, we did not analyze the headache intensity and duration.Furthermore, we did not assess the use of migraine preventive medication, which could potentially influence the presentation and characteristics of MA.Also, we did not assess the socio-economic status, which may have an influence on the diagnosis and symptoms of MA [19,20].Lastly, the study design was cross-sectional, not allowing firm conclusions and further longitudinal studies are needed.The strengths are the large number of patients included and the assessment of symptoms using a personal interview and a structured questionnaire.
In conclusion, with advancing age, the characteristics of MA have fewer migraine-like features and the headache, when present, is accompanied by less migrainous features such as photo-and phonophobia.Our findings highlight the difficulties and challenges in the clinical diagnosis of MA with advancing age.

Table 2 C
-characteristics stratified by age

Table 3
Multivariable logistic regression for the C-characteristics and age (per 10 years) adjusted for sex

Table 4
Characteristics of headache stratified by age