The results of the present study showed that CM patients being NR or R to 12 months of anti-CGRP monoclonal antibody erenumab resulted bear a different psychological burden. In detail, the presence of personality disturbances, in particular those belonging to Cluster C (including obsessive-compulsive personality disorder) and anxiety disorders together with current stressful events of serious and very serious impact, and alexithymic traits resulted as substantial determinants of failure of treatment with erenumab in CM. By contrast, other kinds of personality or mood disorders, childhood traumas, even when associated to a very serious impact, were comparable between non-responders and responders to this preventive treatment. These findings deserve some specific considerations.
A personality trait can be considered as a pervasive and stable over time pattern of acting and interpreting one’s environment, as well as oneself (22). What characterizes a personality trait is that it affects individual’s life and it is not the result of medical conditions or mental disorders neither of substance abuse. The personality traits can be grouped into three clusters: A, B, and C. Cluster A comprises odd or eccentric traits, Cluster B dramatic or emotional disorders traits and Cluster C consists of anxious or fearful traits (23). Many studies explored the prevalence of personality disorders in CM/CM + MO (e.g., 5, 6, 22, 24, 25), recognizing them as factors able to complicate and interfere with headache treatment (26). What is interesting in our study is that the group of CM patients not responding to erenumab were characterized by a higher prevalence of personality disorders with respect to those responding (82% vs 40%). In detail, personality disorders included in Cluster C (77% vs 37%), and in particular the obsessive-compulsive personality disorder (68% vs 36%), were significantly more frequent among NR than R.
Cluster C reflects an “anxious-fearful” and stress reactive personality. It includes people viewing the world as hostile and potentially harmful to them (i.e. avoidant), those considering themselves helpless and believing they need to attach themselves to a strong caretaker in order to get through daily life (i.e., dependent), and those having strong traits and strategies of control and responsibility (i.e., obsessive-compulsive). Interestingly, even not particularly surprising being in line with previous literature on CM (e.g., 21, 23, 27), obsessive-compulsive personality disorder resulted as the most prevalent among NR. A recent study has shown that obsessive-compulsive personality disorder severity is strictly associated to the intolerance of uncertainty (28), that is defined as “a dispositional characteristic that results from a set of negative beliefs about uncertainty and its implications and involves the tendency to react negatively on an emotional, cognitive, and behavioral level to uncertain situations and events” (29). Accordingly, if we consider the prototypical description of “obsessive personality”, it refers to individuals tending to adhere rigidly to their daily routine, becoming uncomfortable and anxious when something goes wrong (30, 31). Hence, we believe that this finding of ours needs to be interpreted also in view of the high prevalence of anxiety disorders and the high number of ‘at least serious’ stressful events and alexithymic traits characterizing NR.
It is well known that CM is associated to psychiatric comorbidities, including anxiety and depression (32, 33), which may play a significant role in the way one perceives pain, copes with it and maintains a normal lifestyle. Furthermore, evidence from other fields of research showed that adults with anxiety disorders reportedly experienced significantly more total life events, perceived them as being more stressful, and adapted to them less well than normal controls (34, 35). Our group of research has also demonstrated that childhood traumas, together with stressful events and alexithymia, are more prevalent among CM + MO with respect to patients with an episodic pattern of migraine (8) and that these factors may contribute to the outcome of the detoxification treatment (2). In particular, stress is the result of the inability to deal with experienced life events (36). It is one of the most common migraine trigger factor and has been implicated in migraine chronification (37, 38). In addition, alexithymia seems to be influenced by the environmental influences, including stressful events (39). There are indeed some pieces of evidence showing that alexithymia might also bias the perception of stress and lead to a decoupling between subjective and physiological responses to it (40, 41). In this frame, our findings seem to support the hypothesis of a further psychological vulnerability in these difficult-to-treat patients, which derives from the complex interaction between all these psychological and psychosocial factors – specific personality disturbances, anxiety, stressful events, and alexithymia – having a bidirectional relationship with migraine complexity and treatment-refractoriness. The bidirectional association between anxiety disorders and migraine, where the presence of one disorder enhances the risk for and the severity of the other, is well known (42, 43). By contrast, personality disorders are usually considered maladaptive traits and behaviors that are stable over time. It has however been demonstrated that their stability is moderate and they can improve over time (44, 45). Hence, it is possible to hypothesize that under particular circumstances, psychological vulnerabilities and clinical conditions may reinforce each other and they may significantly interfere with the outcome of treatment.
Mood disorders and childhood traumatic experiences did not differentiate R and NR. This is an unexpected observation, when considering our previous findings where we showed that early life emotional traumas and major depression during life time were among prognostic factors for the outcome of detoxification therapy in CM + MO (2). However, it is important to note that mood disorders were quite common in NR and R and both groups experienced at least one trauma during their childhood on average. This finding suggests that mood disorders and childhood traumatic experiences do not influence the treatment outcome. Future studies are needed to further explore this topic.
A few other studies have sought to identify possible predictors of response to anti-CGRP monoclonal antibody in CM. Ornello and colleagues (46) compared CM patients converting or non-converting to episodic migraine after erenumab treatment and found that depressive symptomatology was not a significant predictor of conversion. It must be noted that they selected a different timing of efficacy (months 4–6) and the numerosity of the non-converter groups was quite limited (18 subjects), which may have affected the study power. In a population of refractory CM (47), depressive and anxiety symptomatology assessed via self-reported questionnaires did not show predicting values for treatment outcome. It is however worth noting that the Authors excluded patients with comorbid personality disorders and severe psychiatric comorbidities. Furthermore, they set the response threshold to a ≥ 30% reduction in monthly migraine days, which makes it impossible to compare the studies. In partial agreement with the present findings, a multicenter real-life study reported that the responsiveness to erenumab in high frequency migraine and CM was negatively associated with the presence of psychiatric comorbidities (48). Another study (49) showed that the responsiveness to galcanezumab in CM patients with comorbid anxiety and/or depression varied according to the dose administered. It must be noted that in both these two studies, psychiatric comorbidities were simply assessed in terms of presence/absence based on medical records. Finally, a study (50) reported that a treatment with fremanezumab resulted effective in reducing the number of headache days per month in patients with CM and comorbid depression. However, it is important to consider that here depression was evaluated via a self-reported questionnaire. To the best of our knowledge, our study is the first to perform a thorough and detailed evaluation of psychological predictors of unfavorable response to anti-CGRP monoclonal antibody in CM patients.
The neurobiological reasons behind the poorer response to erenumab observed in the CM patients with a higher burden of psychiatric diseases are not immediate. Erenumab is a large molecule that does not cross the blood brain barrier and therefore it is thought to be devoid of direct activity on the central nervous system (51, 52). It is however known that unmanaged stress and anxiety represents risk factors for the development of a hyperexcitable trigeminal system and favour the development of persistent central sensitization (53, 54). It is therefore stimulating to speculate that, in patients with a higher emotional/psychiatric load, erenumab may indeed counteract peripheral sensitization by blocking the activation of the GCRP receptor, but the consequent indirect inhibitory effect on central sensitization may be attenuated or inhibited by the simultaneous effect of psychiatric comorbidities on the opposite direction. Future studies should better explore this hypothesis, for instance including specific measures aimed to determine the extent to which patients have symptoms that may be related to central sensitisation or central sensitivity syndromes, such as the Central Sensitivity Inventory (55).
Some limitations in our study suggest caution in the interpretation of results and call for further ad hoc studies. We did not collect any follow-up about patients’ psychological state after erenumab treatment, which would have possibly contributed interesting additional information about the potential of erenumab to modify personality characteristics. We assessed childhood trauma and current stressors using retrospective self-report questionnaires, which may be affected by recall bias. Finally, though properly calculated form a statistical point of view, the sample size was relatively small, which could have limited the interpretation of our findings. Hence, future multicentric studies on larger CM populations are needed.