This review offers an overview of the PSDs reported in the literature on migraine, the variables associated with them and the determinants that influence their onset and course. The most frequently studied PSDs were related to eight areas: emotional problems, reduced vitality and fatigue, pain, difficulties at work, reduced physical and reduced mental health, poor social functioning and increased global disability. A few variables were identified as associated with a handful of PSDs, namely global disability, emotional problems, pain and headaches frequency. On the contrary, there is more evidence that migraine-specific treatments improve emotional problems, physical and mental health, difficulties with employment and global disability. We found no studies evaluating possible determinants of worsening of PSDs and the presence of pain only was identified as a determinant of PSD onset.
To our knowledge, there is no previous attempt to systematically address PSDs relevant to migraine patients according to a definition based upon the ICF. A previous review [13
] provided evidence that gender and social role expectations, as well as coping strategies, are different and this determines differences in response to pain. In our review, we expanded the scope of problems reported by migraineurs, and added information on the course and factors that influence the improvement of these difficulties. Our results also cast light on the conceptualization of disability found in migraine studies. The Migraine Disability Assessment Schedule (MIDAS) [71
] is a reference point to assess disability in migraineurs and is used in the majority of publications dealing with outcome of migraine. However, it covers only a small part of the entire burden of living with migraine. A major result of our study was the focus on several areas that may be relevant to describe the problems experienced by patients with migraine, and that we believe should be investigated.
The determinants of improvement identified here can be roughly divided in two areas. The first includes variables referred to features of the disease itself, e.g., frequency of headaches and presence of pain. In general, limited evidence was found for the effect of these determinants, although they covered the full span of PSDs. The observed trend acknowledges that reduced headache frequency and pain decrease have a positive effect on improvement in vitality and fatigue—for which strong evidence was derived—emotional problems, physical and mental health, social functioning, work ability and global disability. The second area deals with prophylactic and symptomatic treatments. Studies on symptomatic medication herein included were focussed on different kinds of triptans rather than anti-inflammatory agents, and there is strong evidence that these medications determine an improvement in emotional problems and work efficiency. However, the most important determinants of PSD improvement found in this literature review were prophylactic medications, and there is strong evidence that these medications positively affect emotional problems, improve work efficiency, global disability, physical and mental health.
While several pharmacological studies have been published in the past years, limited data exist on non-pharmacological treatments. There is only sparse evidence on the effect of complementary treatment and psychological therapy: two studies showed efficacy of massage therapy and yoga for the improvement of anxiety and mood problems [44
], while one study found that home-based behavioral training might be effective to improve patient functioning [59
]. The fact that such non-pharmacological interventions did not frequently occur in this literature review is in part due to the search strategy that gave primary relevance to longitudinal intervention studies, i.e., clinical trials of acute or prophylactic medication. Surgery is not a common procedure for migraine, rather it is indicated for the treatment of chronic cluster headache [72
] and there are few experience on its use to treat patients with drug-resistant chronic migraine [73
] also with comorbidity to depression [74
]. As a consequence, the extent to which non-pharmacological treatments might improve migraineurs’ difficulties is still an open question that should be addressed in future research.
The articles included in the present review were mostly reports of clinical trials, which provided a strong control over study variables. In daily practice, however, clinicians have to deal with acceptance and adherence to treatment, which is a relevant issue for both prophylactic and acute migraine treatments [75
]. Among migraineurs, the issue of non-adherence to treatment may have different implications, varying from inadequate timing in taking triptans, to not accepting prophylactic medications, to the overuse of symptomatic ones. Medication overuse, jointly with comorbidity to mood problems [78
], might determine worse health outcomes. In fact, problems with adherence to migraine treatment might be further on amplified by low treatment adherence which in mood disorders is around 40 % [80
]. Multidisciplinary treatment has been proposed as a strategy for improving adherence to treatment, but results are conflicting [60
]. The implication of this is that our results mostly report facts that have been generated in the ideal contexts of clinical trial, but the situation of patients in real-life settings might be quite different.
Headache frequency is a determinant that deserves a separate comment. Frequency was found to be associated with the most relevant PSDs, but with strong evidence only for improvements in vitality, physical and mental health. It should be pointed out that reduction of headache frequency—which is not a PSD itself—is generally the primary endpoint (for e.g., in clinical trials on the use of prophylactic medications), while PSDs such as vitality were viewed as secondary outcome measures. Considering the aim and the methodology of the present review, the implication of this fact is that we are likely to underestimate the causal relationship between the treatment intervention tested in clinical studies, the magnitude of their effect on primary endpoints, and the improvements in PSDs that we focus in this review.
Some limitations of this study should be mentioned. Even though our search was extensive, we cannot be sure that all relevant articles were located. The fact that no determinant of PSD worsening was found is likely to be due to a publication bias, with studies that report negative results not being published. Some aspects of the included studies may have influenced our results, in particular the fact that there is an over-representation of data derived from the MIDAS [71
] and the Short-Form 36 Health Survey (SF-36) [84
]: taken together, they were used in 28 of the 51 included studies. The MIDAS may provide indications of either global disability or problems with work, household and leisure activities, which were herein described as separate PSDs. Similarly, SF-36-derived data were either reported as summary scores and therefore described as general health scores, or as subscales (e.g., vitality and social functioning) and thus reported as separate PSDs. Since these instruments are almost always used, the same PSDs are almost always reported, so that information about other PSDs are less frequently reported and remain almost unknown. However, the representation of PSDs associated with migraine is partial. The reason for this is that while there is a homogeneity due to the amount of data derived from MIDAS and SF-36, the number of PSDs is not describable in a synthetic way if the purpose is to avoid the reporting of known issues such as increased disability and reduced quality of life. With our synthesis, we tried to balance the opposite needs of being comprehensive and synthetic. Investigators in the field of migraine are therefore encouraged to evaluate other kinds of daily difficulties not included in commonly used instruments, and to include outcome measures that are able to capture the burden of migraine in a comprehensive way.
As we were interested to evaluate the course over time and determinants of PSDs’ course over time and the determinants of PSD change over time, we finally included several clinical trials: this is likely to reduce the ecological validity of our results, as subjects participating in clinical trials are exposed to a situation that is not the same commonly found in daily clinical practice.
A comment is also needed for pain, which was reported as a PSD in a relatively limited number of studies, e.g., when pain severity was directly assessed (for e.g., with the visual analog scale contained in MIDAS). In other studies, pain was considered to be a determinant, for example when reduction in pain was not directly measured, but was used to create groups of subjects for between-subjects analysis (e.g., the percentage of subjects achieving pain relief or experiencing pain reduction by a given timeframe following the intake of a medication).
Finally, the heterogeneity across studies should also be taken into account. The studies were very different with respect to sample size, number and duration of follow-up as well as study designs that provide different levels of control over confounding variables. Disease duration was reported in a small number of studies, hence preventing the evaluation of the effect of length of exposure to different PSDs, which may be influenced by disease duration.