Prevalence and clinical features associated to bipolar disorder-migraine comorbidity: a systematic review.

Available online at www.sciencedirect.com

ScienceDirect Comprehensive Psychiatry xx (2014) xxx – xxx www.elsevier.com/locate/comppsych

Prevalence and clinical features associated to bipolar disorder–migraine comorbidity: a systematic review Michele Fornaro a,⁎, Domenico De Berardis b , Concetta De Pasquale a , Luisa Indelicato a , Rocco Pollice c, 1 , Alessandro Valchera d , Giampaolo Perna e , Felice Iasevoli f , Carmine Tomasetti f , Giovanni Martinotti g , Ann Sarah Koshy h , Ole Bernt Fasmer i , Ketil Joachim Oedegaard j a Department of Education Science, University of Catania, Catania, Italy National Health Service, Department of Mental Health, Psychiatric Service of Diagnosis and Treatment, Hospital “G. Mazzini”, ASL 4, Teramo, Italy c Service for Monitoring and early Intervention against psychoLogical and mEntal suffering in young people” (SMILE), L'Aquila University, Italy d Hermanas Hospitalarias, Villa San Giuseppe Hospital, 63100 Ascoli Piceno, Italy e Department of Clinical Neuroscience, Villa San Benedetto Menni, Hermanas Hospitalarias, Albese con Cassano, Como, Italy f Department of Neuroscience, Reproductive Sciences and Odontostomatology, University “Federico II” of Naples, Via Pansini 5, 80131 Naples, Italy g Department of Neurosciences and Imaging, University “G. d'Annunzio” of Chieti, 66013 Chieti, Italy h St. John's National Academy of Health Sciences, Bangalore, India i Department of Clinical Medicine, Section for Psychiatry, University of Bergen, Bergen, Norway j Department of Psychiatry, University of Bergen, Bergen, Norway

b

Abstract Background: The prevalence and clinical features associated with bipolar disorders (BDs)–migraine comorbidity have been reported inconsistently across different studies, therefore warranting a systematic review on the matter. Methods: A systematic review was conducted in accordance with the PRISMA statement searching major electronic databases for documents indexed between January, 2000 and July, 2014. Eligible studies were those including quantitative data on prevalence rates and clinical features associated to BD–migraine comorbidity; case reports excluded. Three authors independently conducted searches, quality assessment of the studies and data extraction. Results: Several cross-sectional studies, and a handful of retrospective follow-up studies or non-systematic reviews assessed the prevalence and/or the clinical correlates of migraine–BD comorbidity. High prevalence rates and a significant burden of BD–migraine comorbidity were common findings, particularly in case of BD-II women (point-prevalence rates up to 77%), migraine with aura (up to 53%) and/or cyclothymic temperament (up to 45% of the cases). Limitations: Some of the biases encountered in a few studies accounted by the present review may nonetheless have hampered the generalizability of the overall conclusions drawn herein. Conclusions: BD–migraine comorbidity may comprise of a sub-phenotype of BDs requiring patient-tailored therapeutic interventions to achieve an optimal outcome. Specifically, additional studies including longitudinal follow-up studies are aimed in order to shed further light on the actual prevalence rates and clinical features associated to BD–migraine comorbidity, with a special emphasis towards the clinically suggestive potential connection between mixed features, bipolar depression, migraine, and increased risk for suicidality. PROSPERO registration number: CRD42014009335. © 2014 Elsevier Inc. All rights reserved.

⁎ Corresponding author at: Teatro Greco n.84, ZIP 95125, Catania, Italy. Tel.: +39 347 4140003; fax: +39 010 3537681. E-mail addresses: [email protected] (M. Fornaro), [email protected] (D. De Berardis), [email protected] (C. De Pasquale), [email protected] (L. Indelicato), [email protected] (R. Pollice), [email protected] http://dx.doi.org/10.1016/j.comppsych.2014.09.020 0010-440X/© 2014 Elsevier Inc. All rights reserved.

(A. Valchera), [email protected] (G. Perna), [email protected] (F. Iasevoli), [email protected] (C. Tomasetti), [email protected] (G. Martinotti), [email protected] (A.S. Koshy), [email protected] (O.B. Fasmer), [email protected] (K.J. Oedegaard). 1 Deceased.

2

M. Fornaro et al. / Comprehensive Psychiatry xx (2014) xxx–xxx

1. Introduction According to the International Classification of Headache Disorders, 2nd edition (ICHD-2), migraine refers to different phenotypes having in common a low threshold to the development of headache among migraineurs, usually being characterized for a recurring pattern, frequent freeinterval, and usually provoked by stereotyped triggers [1]. Notwithstanding, a conclusive definition of migraine remains elusive, especially due to the heterogeneity of some of the frequently associated neurological and psychiatric comorbidities. Specifically, concerning psychiatric comorbidities, several large-scale population-based studies have confirmed clinicians' longstanding suspicions: mood (and Anxiety) disturbances are more prevalent in migraineurs vs. the general population [2]. Much research involving migraine and mood disorders, including prospective, large-scale, population-based studies, has focused on depression, yet a three-fold increase in migraine prevalence in patients with bipolar disorders (BDs) has been documented as well [3]. Specifically, treatment-seeking patients with migraine, and tension-type headache [4] exhibit psychiatric comorbidity (including BD) at rates disproportionately higher than individuals with no history of recurrent migraine [5–9]. However, the actual rates of BD among migraine patients and related clinical features have been reportedly inconsistent, especially when looking at peculiar features such as bipolar subtype, gender, and age or illness severity, at least in comparison with migraine–major depressive disorder (MDD) comorbidity, wherein rates have been documented to range between 3.8 and 57% compared with an average general population lifetime rate of 16% [10–12], or to comorbid generalized anxiety disorder or panic disorder, both found to have higher prevalence rates of migraine comorbidity compared to the general population (N50 vs. 27%) [2]. The purpose of the present systematic review is to examine the prevalence rates of migraine–BD comorbidity and the clinical features associated to both migraine in course of BD and BD in course of migraine. 2. Material and methods Aimed at achieving a high standard of reporting, we followed the procedures indicated by the 2009 update of the Preferred Reporting Items for Systematic reviews and MetaAnalyses (PRISMA) guidelines (http://www.prismastatement.org/) [13]. The protocol of the present systematic review is available through PROSPERO (http://www.crd. york.ac.uk/PROSPERO/), where it was assigned the following registration number: CRD42014009335. 2.1. Eligibility criteria We limited our search to those records including adult subjects with BD, published in peer-reviewed journals.

Limits activated were: species: humans, language: English, all adult: 19+, publication date: from January 1, 2000 to July 25, 2014. Such criteria were adopted following clinical considerations, aimed at providing quality consistency among the studies, yet screening as many sources as possible. Specifically, disregard of contributes indexed before year 2000 essentially followed “diagnostic reliability” considerations. In fact, it was in the 1980 US release of the DSM-III [14] that the term “bipolar disorder” replaced the older term “manic depressive disorder”. The new term, “bipolar disorder” reflected the defining feature of mood polarity rather than simply pointing to the consequences of that polarity: mania and depression, also indicating the distinction between adult and pediatric bipolar disorder diagnoses for the first time. Similarly, the older diagnostic criteria for “migraine” were largely inconsistent with the most recently introduced and broadly adopted ones. Essential contributes published before year 2000 have nonetheless been acknowledged through the text for consistency issues [6,15]. 2.1.1. Information sources and search strategy Sources of information included the following databases: PubMed/Medline, Scopus, PsycLit, PsycInfo, Embase, and Cochrane library. Contact with study authors was attempted when needed, leading to the identification on an additional document [16]. The following key words or their combination were used in the search strategy: “migraine AND bipolar disorders” either in the title and abstract (or in the key words where specified). The adopted PubMed string was: (migraine [Title/Abstract]) AND bipolar disorder[Title/Abstract]) AND (“2000”[Date – Publication] : “2014”[Date – Publication]). 2.2. Study selection Included papers were those reporting epidemiological data about the comorbidity between migraine and BD (and vice versa), with no restriction on etiology or type. Papers covering cases of migraine–BD comorbidity associated with additional disorders (either psychiatric, neurological [including migraine due to epilepsy, Tourette's syndrome or schizophrenia] or to other somatic disorders/diseases) were also evaluated wherever available. When a title and/or an abstract appeared suggestive for inclusion, the full text reprint was obtained and examined to assess its relevance according to our inclusion/exclusion criteria. Excluded papers were case reports, papers not including BD migraine co-occurring cases, those merely focusing on neurobiological, genetic or pharmacological aspects of either migraine or BD; including only children or adolescents; or without an accurate description about the diagnostic definitions of either migraine and BD. 2.3. Data collection process Three authors (MF, LI, and CDP) conducted a two-step literature search, examining all titles and abstracts, accessing the full texts of potentially relevant papers. Upon data


collection and extraction, the appointed authors compared their own results with each other to reach a final consensus based on consensual inclusion and exclusion criteria. Any eventual discrepancy between the principal investigators, blind to each other, was solved by consultation with the senior authors (DDB, AV, GP, FI, CT and GM). Finally, two senior researchers with considerable experience on the topic (KO and OBF) assisted in manuscript revision. Data were sought for the following characteristics: participants, interventions, comparisons, outcomes, and study design (PICOS), as well as funding sources. Specifically, the recorded variables for each article included in the review were: author(s), year publication, study design, sample size, eventual follow-up or control group, outcome measures, conclusions, limitations, quality score, and quality differentiation. 2.4. Risk of bias in individual studies Potential major confounding biases in the studies were ascertained at study level focusing on the following: measurement/diagnostic bias (e.g. lack of reliable diagnostic tools to make the diagnosis of either migraine or BD), confounding bias (e.g. lack of stratification and multivariate control for specific socio-demographic, vital or clinical features), information (especially recall) bias, unrepresentativeness or inhomogeneity of the sample size or lack of control group, selection by indication bias (nonrandom assignment of the exposure where applicable) [17]. 2.5. Summary measures Summary measures were risk ratio (RR) for cohort studies odd ratio (OR) for case-control studies, or difference in means. 2.6. Scoring and ranking of the studies Evaluated data were handled and compared to each other using quality score and quality differentiation adapting the method already used for systematic reviews on BD (but not comorbid migraine) published in high-yield peer-reviewed journals [18]. Specifically, studies were rated for quality using the following eligibility criteria; (i) representativeness of the sample (0–1 points); (ii) presence of BD patients only in the sample (0–2 points); (iii) a priori study design with the goal of evaluating the epidemiology of migraine–BD comorbidity (or vice versa) (0–2 points); (iv) extension of the followup (longitudinal studies)/clinical records (retrospective studies) N1 year (0–2 points); (v) validation of the clinical diagnosis and the used treatment (if applicable) (0–2 points); (vi) inclusion and control of all the available variables for potential confounders/effect modifiers that may had influenced outcome (if applicable) (0–2 points); (vii) reliability of the information gathered for the identification of BD cases/ recall bias (0–2 points); (viii) accuracy of the study was to discern between manic, hypomanic, mixed, and depressive

3

episodes in BD (0–2 points); (ix) appropriateness of the number of comorbid cases reported as results/sample size (0–2 points). Quality rating had 17 as the maximum score. Studies were also differentiated in the following way: (i) good quality: most or all criteria being fulfilled, and when they were not met, the study conclusions were thought to be very unlikely to alter (12–17 points); (ii) moderate quality: some criteria being fulfilled, and where they were not met, the study conclusions were thought to be unlikely to alter (7– 11 points); (iii) poor quality: few criteria fulfilled but the conclusions of the study were thought to be very likely to alter (0–6 points). 2.7. Risk of bias across the studies Any eventual bias affecting cumulative evidence (e.g. publication bias, selective reporting within studies) was assessed through the study evaluation process and accounted in the discussion of the present manuscript.

3. Results 3.1. Study selection The search in PubMed generated 118 papers, and 595 results in Scopus. Thirteen additional results were obtained through search either in Cochrane (n = 7), PsycLit, PsycInfo or reference textbooks/manual search. The combined search strategy yielded a total of 726 results, of which 98 removed after screening because they duplicated other articles. Of the remaining 616 papers, only 115 were further screened; of those later, 51 were excluded because neither titles nor abstracts actually indicated that patients with migraine and BD were included; they were not published in the English language; they had low relevance to the theme of the present review or they included otherwise unclear data regarding the materials, methods, and number of patients analyzed. Therefore, 64 sources (including additional material found via manual search or upon contact with the author) were further reviewed; among these later, 26 were excluded because of their unsatisfactory accuracy in reporting epidemiological data about migraine–BD comorbidity. Finally, 38 references (27 original studies, nine reviews and two edited textbooks) were accounted in our qualitative analysis. Fig. 1 provides a synthetic flow chart of the multistep selection procedure. 3.2. Quality score and quality differentiation results Based on the quality differentiation, original studies were ranked as follows: “poor” (n = 10, mean total score = 4.2), “moderate” (n = 4, mean total score = 8.5) or “good” quality (n = 4, mean total score = 14). No longitudinal follow-up study was found. Details about good quality original studies have been summarized in Tables 1 and 2.

4


Identification

Adaptation of the PRISMA 2009 Flow Diagram [1].

Records identified through database searching (n = 713)

Additional records identified through other sources (n = 13) (Edited books, contact with the authors, manual search)

Records after duplicates removed

Eligibility

Screening

(n = 616)

Records screened (n = 115)

Records excluded (e.g. not related content) (n = 51)

References assessed for eligibility

Sources further excluded (e.g. inaccurate definitions) (n = 26)

Included

(n = 64)

Sources included in qualitative synthesis (n = 38)

Fig. 1. Flow chart of review procedures [13].

3.3. Prevalence rates of BD–migraine comorbidity A 2001 cross-sectional naturalistic study by Fasmer [19] carried on 62 Norwegian outpatients (females = 71%) with either BD-I (n = 14), BD-II (n = 13) or MDD (n = 35), diagnosed according to the International Headache Society (IHS) [1] and Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition (DSM-IV) criteria [20] reported slightly lower prevalence rates of migraine comorbidity in BD vs. MDD (44 vs. 46%) yet documenting a striking prevalence of migraine–BD comorbidity in BD-II cases alone (77%) vs. 14% in BD-I alone (p b .001). The likelihood of having migraine among BD patients experiencing a current MDE vs. MDD controls was ascertained and correlated with the risk of having a positive family history for BD even in Latino adults [21]. The authors highlighted a considerable risk of having migraine among those BD patients (either type-I or type-II) with a positive family history for BD compared to those who did not. In a subsequent collaborative study, the same authors also found

that migraine headache in affectively ill (depressed) Latino adults of Mexican American origin was associated with bipolarity [21]. Specifically, patients with BD were 2.9 times more likely to have migraine than those with MDD (OR = 2.9; CI = 1.6–5.2, p b .001). Of the non-affectively ill (not currently experiencing a MDE) comparison psychiatric patients (including individuals with substance use disorder n = 10, adjustment disorder n = 4, ADHD n = 2 and DSM-IV other Axis-I disorder n = 19), 14.2% had migraine headache. Also, there was a progression of the risk for migraine ranging from 14.2% among the patients who did not meet the criteria for a MDE to 29% among those meeting the criteria for MDD to 54% among those meeting the criteria for BD, suggesting a trend across groups and that migraine could be much more strongly associated with bipolarity rather than unipolar depression. Interestingly, most (n = 81) of the BD patients included in the sample (87 BD and 123 MDD) had type-I BD (BD-II, n = 6), which further highlighted the comorbidity of migraine as a “red flag” for BD, since migraine was documented to be

Table 1 Good quality cross-sectional studies. Aim/hypothesis

Sample size

Main results

Conclusions

Limitations

Quality score

Quality differentiation

Low et al., 2003 [35].

To investigate the prevalence and clinical correlates of comorbid migraine in BD patients vs. non-comorbid bipolar patients.

One hundred eight BD patients (17 BD-II).

BD-II patients, especially females, had a higher prevalence of family history of depression and a lower chance/propensity to be on lithium or other mood stabilizers rather than atypical antidepressants.

BD-II patients with migraine may be at higher risk for MDD misdiagnosis, family history for psychiatric disorder, have an overall higher burden and chance to receive antidepressant alone rather mood stabilizers.

I=1 II = 2 III = 2 IV = 0 V=2 VI = 2 VII = 0 VIII = 2 IX = 1 Total = 12

Good quality.

Ortiz et al., 2010 [8].

To explore the crossprevalence of migraine in BD-I and BD-II patients (study 1), then to explore the prevalence of comorbid psychiatric disorders (including BDs) in migraine patients (study 2).

Study I = 214 subjects (BD-I = 126, BD-II = 61, BD NOS or schizoaffective disorder, n = 27). Study 2, 102 patients, of whom, 0 had a current BD-I, 7 a current BD-II and 5 had a lifetime BD-I vs. 8 with a lifetime BD-II).

Up to 24.5% BD patients had comorbid migraine; those with BD-II had a higher prevalence (34.8%) compared to BD-I (19.1%), p b .005). Psychiatric comorbidities and suicidal behavior were also common.

Migraine is prevalent among BD patients, especially BD-II cases. Risk for suicidal behavior and comorbid Anxiety disorders is also high.

Unrepresentativeness of the BD-II subset. Overrepresentation of females in the included set. Recall bias. Use of outdated IHS criteria for the questionnaire adopted for the diagnosis of migraine. Lack of detailed report about different psychiatric drugs. The diagnosis of migraine was made by a self-assessment questionnaire. Overrepresentation of BD-I and female cases. The interviews were partially conducted in a clinic specialized in the treatment of migraine (chance of Berkson's bias).


Good quality.

BD = bipolar disorder; MDD = major depressive disorder; MDE = major depressive episode; NOS = not otherwise specified.


Author, date

5

6

Table 2 Good quality (retrospective) follow-up studies. Aim/hypothesis

Sample size

Main results

Conclusions

Limitations

Quality score

Quality differentiation

Nguyen and Low 2013 [31].

To carry betweengroup and amonggroup comparisons of migraine in BD according to lifetime mood episode(s)/ combination of mood episode(s) in affective disorder patients (including BD) vs. first-degree relative vs. general population.

N = 36.984 patients (Canadian Community Health Survey 1.2).

Migraine comorbidity in BD was associated with an earlier age of onset of BD itself, especially with depressive polarity of first episode, as well as with a higher number of psychiatric additional comorbidities.

Post-hoc retrospective chart review. Recall bias.


Good quality.

Saunders et al., 2014 [45].

To explore whether the higher comorbidity of migraine in BD women vs. BD men documented by most of the previous evidence would be associated with specific features of illness and psychosocial risk factors that would differ by gender and impact outcome.

BD patients, n = 412, healthy controls, n = 157.

Overall, compared with controls, the adjusted OR of having migraine was 2.0 (95% CI = 1.4–2.8) for manic episodes alone, 1.9 (95% CI = 1.6–2.1) for depressive episodes alone, and 3.0 (95% CI = 2.3–3.9) for subjects with both manic and depressive episodes. Compared to with those with manic episodes alone and depressive episodes alone, the odds of having migraine were significantly increased in subjects with both manic and depressive episodes (OR 1.5 vs. manic episodes alone; 1.8 vs. depressive episodes alone). Among 306 BD-I patients, n = 223 had no comorbid migraine, n = 83 had comorbid migraine. Migraine was comorbid in 29 BD-II patients and absent in 37 BD-II cases. History of suicide attempt was recorded in 110 (39%) cases of non-comorbid migraine vs. 57 (45%) of

Since migraine was more common in BD subjects (31%) vs. controls (6%) and had an elevated risk in BD women compared to men (OR = 3.5; 95% CI, 2.1–5.8), the presence of migraine, especially in course of mixed symptoms, may be

Self-reported, physician made, diagnoses of migraine (recall bias). The sample included patients with BD followed over time, meaning that the most severe end of the bipolar spectrum may have not been represented; risk of Berkson's bias. Control


Good quality.


Author, date

prompt clinicians to account migraine as a common comorbidity which may ultimately affect the long-term course of BD itself.

group size smaller that studied population. Overrepresentation of BD-I cases.


cases with comorbid migraine. Rapid cycling pattern was recorded in 89 (31%) cases without migraine vs. 52 (41%) cases with comorbid migraine. Interestingly, this was also the first study specifically exploring mixed symptoms, which were found in 102 (36%) of non-migraine cases vs. 63 (50%) of comorbid cases. With respect to gender distribution, there was no statistically significant difference between rates of migraine in the BD-II group vs. BD-I one in the women (OR = 1.6; 95%, 0.8–3.0, yet group differences between women with BD and migraine vs. those without comorbid migraine included a marginally higher rate of mixed symptoms (OR = .19, 95% CI, 1.0–3.7). BD = Bipolar Disorder; MDD = Major Depressive Disorder; MDE = Major Depressive Episode.

7

8


considerably more prevalent in BD-II vs. BD-I cases [22]. Noteworthy, while most of the previous investigations already studied the relationship between migraine and affective disorders [2,6,15,19,22–29], and other authors already reported a high prevalence of migraine among both BD and MDD patients, this study [21] was the first one documenting a statistically significant difference in terms of prevalence as a function of polarity compared to other studies exploring such relationship [6,19,22]. The lifetime prevalence of migraine in psychiatric patients (including BD cases) compared with those of first-degree relatives and the general population was explored by Baptista et al. [30] in Venezuelan adult subjects. Overall, migraine was recorded in 42.4% of general population females (n = 219) vs. 57.6% of general population males (n = 297); the prevalence of migraine in BD (n = 191) was as high as 63.4% in females (n = 121) vs. 36.6% in males (n = 70); all differences, p b .001. Unfortunately, no further distinction has been reported in this multi-purpose study with regard to BD-I and BD-II comparisons, and the diagnosis of BD-II was also broad, including subjects with cyclothymic and hyperthymic temperaments, while the diagnosis of migraine was generic. The cross-prevalence of migraine and BD was investigated also in Canadian adults [8] in two related community-based studies. The first study explored the prevalence of migraine in BD patients (n = 323, mainly BD-I or BD-II cases). Thirty nine BD-I patients out 204 BDs (or 19.2% of the total), had migraine vs. 165 (80.8%). BD-II cases with comorbid migraine were 32 (34.8%) vs. 60 (65.2%) out of 92 BD-II patients. Nguyen and Low. also explored the prevalence of migraine comorbidity and mood episodes (including those of BDs) in a nationally (Canadian) representative (n = 36,984) population-based study [31]. Specifically, although the association of migraine and BD was already documented in previous reports, the variability in related findings across studies suggested the need for authors to examine mood episodes separately (manic episodes alone vs. depressive episodes alone vs. manic and depressive episodes vs. controls with no lifetime history of mood episodes, including record of each socio-demographic and clinical features associated to every mood episode) to determine which of these later may be specifically associated with migraine. Logistic regression analyses controlling for age, sex and education level comparing the lifetime prevalence of migraine between controls and each combination of mood episodes, and then among the different combinations of mood episodes, pointed out that migraine was associated with lower socio-economic status, earlier age of onset of affective illness, suicidality and use of mental health resource. Specifically, with regard to BD, migraine was comorbid in 9.3% of controls without mood episodes (70.8% were female and 14.6% had suicidal ideation), in 19.5% of manic episodes alone (65.9% females; 47% lifetime suicidal ideation) and in as much as 28.5% of manic and depressive episodes (71.3% females; suicidal ideation = 18% compared to 20% of non-migraineurs). Overall, compared with

controls, the adjusted OR of having migraine was 2.0 (95% CI = 1.4–2.8) for manic episodes alone, 1.9 (95% CI = 1.6–2.1) for depressive episodes alone, and 3.0 (95% CI = 2.3–3.9) for subjects with both manic and depressive episodes. Compared with those with manic episodes alone and depressive episodes alone, the odds of having migraine were significantly increased in subjects with both manic and depressive episodes (OR 1.5 vs. manic episodes alone; 1.8 vs. depressive episodes alone). The first report about the comorbidity profiles of chronic migraine among Asian patients came from Chen et al. [32], who analyzed the Taiwan National Health Insurance Research Database covering the period within 2006–2009 (http://nhird.nhri. org.tw/en/). Among other findings, the only result about BD patients with chronic migraine was that they had an RR = 1.81, p b .022 compared to BD patients with other clinical presentations of migraine. 3.4. Main clinical features associated with migraine–BD comorbidity A 2001 study by Fasmer and Oedegaard [22] reported the clinical characteristics of outpatients with major affective disorders (either BD or MDD) having migraine as a comorbidity (n = 53, of whom BD-I = 6 and BD II = 23) vs. patients without comorbid migraine (n = 49, of whom BD-I = 16 and BD- II = 5). The study documented a higher occurrence of BD- II (43 vs. 10%) and a lower frequency of BD-I (11 vs. 33%) in those with comorbid migraine than those without comorbid migraine. Also, a family history of migraine was recorded in up to 49% of affective cases with migraine vs. 18% of affective cases without migraine (p = .001). In contrast to previous findings [33], the study at issue [22] failed to detect any statistically significant difference in terms of atypical symptoms between bipolar depressed patients with comorbid migraine vs. non-comorbid patients (as well as any significant difference for melancholic or psychotic features), yet, a statistically significant difference for increased frequency of irritability (p = .018) and suspiciousness (p = .019) was documented during MDEs in major affective disorder patients with comorbid migraine vs. affective controls. Remarkably, no additional reported analysis further discriminated between bipolar (I vs. II) and MDD cases. The authors nonetheless proposed that irritability may be both a part of the prodromal phase that often precedes migraine headache [34] and a prominent symptom during (hypo-)manic or mixed affective episodes, accepting that an increased presence of irritability among migraine patients (including those with DSM-IV-defined MDD), could reflect a linkage to BD, as already proposed in 1999 by Mahmood et al. [6]. Additional evidence about the prevalence rates of migraine in BD outpatients was reported in 2003 by Low et al. [35] who assessed 108 BD cases (F = 73, M = 35, mean age = 44.8 years) through a migraine questionnaire (adapted by the “Guidelines for the Diagnosis and Management of Migraine in Clinical Practice” published


by the Canadian Medical Association Journal in 1997 [36]) which enabled recording of clinical and demographic information. Overall, the lifetime prevalence of migraine was found to be 39.8% (43.8% in females vs. 31.4% among males). Specifically, the BD-II subgroup (n = 17) had a migraine–comorbidity prevalence of 64.7%. Of these BD-II patients, n = 12 (70.5%) were women, of whom 9 (75%) had comorbid migraine. BD-II patients with comorbid migraine had a younger age of onset of depression, higher levels of education and fewer psychiatric hospitalizations; these patients also had a more prevalent (though not statistically significantly different) family history for migraine (64.9 vs. 25.9%), psychiatric family history compared to non-migraineur BD-IIs and, usually, an index episode of depression rather than mania [35]. Migraine with aura was common (53.5%), and only 16.3% of those with migraine visited a neurologist, yet in 93% of the cases the patients took medications for migraine, of whom, 81% used (or selfreported to abuse) over-the-counter pain killer medications for migraine. BD–migraine patients were also more often on atypical antidepressants rather than mood stabilizers (after controlling for gender, as women had a propensity to be on lithium). Rates of previous misdiagnosis were also high, with up to 23.3% bipolar patients with migraine being previously diagnosed as MDD compared to non-migraine BDs (6.2%), p b .01 [35]. Further information about the pharmacological prescription trends in migraine–BD was reported by retrospective studies. A post-hoc analysis of the 2006 Norwegian Prescription Database (NPD) (www.reseptregisteret.no) including 4,640,219 subjects, investigated whether a suggestive causal confirmation of the link between migraine and BD should be outlined by a common prescription pattern among patients with comorbid BD and migraine [37]. A total of 81,225 persons (1.8% of the population) received medications for migraine and 19,517 (0.45% of the population) received a mood-stabilizing agent for BD; 843 persons (0.02% of the population) received both types of drugs. The OR expressing the relationship between the concurrent use of both categories of medications was 2.55 (95% CI, 2.38–2.73, p b .001, z score = 26.44), significant for all mood stabilizers (lithium: OR = 1.82, CI = 1.58– 2.10, p b .001, z score = 8.31; carbamazepine: OR = 2.48, 95% CI = 2.01–3.06, p b .001, z score = 8.42; valproic acid: OR = 2.26, 95% CI = 1.89–2.70, p b .001, z score = 8.96; lamotrigine: OR = 3.50, 95% CI = 3.14–3.90, p b .001, z score = 22.68). Also, the association was higher for men (OR = 3.16, 95% CI = 2.74–3.66, p b .001, z score 15.53) than for women (OR = 2.21, 95% CI = 2.04–2.39, p b .001, z score = 19.61), actually supporting the hypothesis that a common trend of prescription for migraine and BD in comorbid cases may be the expression of common pathophysiological underpinning. A post-hoc analysis of the 2006 release of the NPD was also reported with regard to antidepressant prescription trends in those subjects with comorbid migraine and BD [37]. Antidepressant drugs were

9

found to be prescribed in 257,700 persons (or 5.6% of the population), and 11,269 persons (0.24% of the population) were prescribed with both antidepressant and migraine-killer medications. The prescription of antidepressants was found to be significantly increased among patients receiving a prescription for migraine-killer drugs (OR = 2.82, 95% CI = 2.76–2.88); χ 2 p b .001), and this association was stronger for men than for women, although teenage women carried the highest risk for this co-morbid constellation (OR = 3.89, CI = 3.17–4.77; χ 2 p b .001). It is worth noticing that in the same year, Holland et al. [9] reported low prevalence rates of comorbid migraine and BD in their chart review study involving 1083 adult outpatients (wherein BD patients numbered 169; of whom 46 self-reported migraine; and prevalence of comorbid migraine and BD = 4.7%). Such prevalence rates were in line with those (2%) reported in the same year by Weber et al. [38] based on the US National Hospital Discharge Survey (NHDS) covering 5,587,707 (BD cases, n = 27,054) cases followed within the period 1979– 2006. In order to better characterize the different clinical features of migraine in relationship with affective disorders, including BD, Oedegaard and his collaborators [23] investigated the clinical correlates of migraine with aura (n = 57, of whom BD-I = 5 [9%] and BD-II = 21[37%]); migraine without aura (n = 41, of whom BD-I = 4[10%] and BD-II = 14[34%]); migraine aura without headache (n = 18, of whom BD-I = 4[22%] and BD-II = 5[28%]) and no migraine (n = 85, of whom BD-I = 23[27%] and BD- II = 11[13%]). The authors reported that the frequency of suicide attempts in the current study was high among the patients having migraine with aura (53%) and against this background it was surprising to find that only 3 (17%) of the eighteen patients having migraine aura without headache had made a suicide attempt. An additional significant finding was that only 28% of the patients having migraine without aura had an affective temperament compared to 56% of the patients having migraine with aura (these were later characterized by an overrepresentation of cyclothymic temperament). Having migraine with aura was also strongly associated with an early age of onset (OR: 11.4, age b 15 years, OR: 5.6, age 15–24 years), and upon multivariate analysis, this effect explained the differences between the two migraine groups better than the variables suicide attempt (OR: 3.2, p = ns) and affective temperaments (OR: 4.0, p =.047). The relationship between migraine and mental disorders was also investigated in a large populationbased German sample (n = 4181), which included only 40 BD patients (0.8%), with no clear distinction being reported between BD-I and BD-II cases [39]. Nonetheless, after adjusting for socio-demographics confounders, past-year migraine (no further distinction being reported about the type of migraine) was positively associated with any past-year mood disorder, including BD (adjusted OR = 2.78; 95% CI, 1.29–5.97). With respect to anxiety disorders, panic disorder was detected in as much as 50% of migraineurs (vs. 26% of non-migraineurs). Overall anxiety disorders were common

10


in BD patients with comorbid migraine [19]. In a 2008 Canadian study [8], migraine was associated with social phobia (SP) (OR = 3.8; 95% CI = 1.9–7.4), panic disorder (PD) (OR = 4.3; 95% CI = 2.3–7.8), obsessive– compulsive disorder (OCD) (IR = 2.1; 95% CI = 0.9–4.6), generalized anxiety disorder (GAD) (OR = 2.1; 95% CI = 1.1– 3.7), diagnostic subtype (OR = 2.2; 95% CI = 1.2–3.9), history of suicidal behavior (OR = 1.7; 95% CI = 1.0–3.0) and lower body mass index (KW = 4.5, p b .05). Interestingly, the study found that the increased prevalence of migraine among BD-II patients was not related to increased rates of BD-II among women [χ 2 (3) = 3.68, p = .29]. Conversely, the second part of the study [8] included 102 migraine patients, who were assessed for their lifetime and current prevalence of comorbid psychiatric disorders, including BD-I and BD-II. In this sample (75.5% women and 24.5% men), the most common diagnosis was migraine without aura (83.3%). Notably, a diagnosis of migraine preceded a psychiatric diagnosis (if any) in 78.6% (n = 59) of the cases of the affected subjects (n = 75), whereas a prior psychiatric diagnosis (if any) followed by a diagnosis of migraine was made in only 14.6% of the sample (n = 11), while both diagnoses were made during the same year in 6.7% of the cases (n = 5). Overall, the results from the two studies combined, pointed out that 24.5% of BD patients had comorbid migraine; those with BD-II had a higher prevalence (34.4%) compared to BD-I (19.1%), p b .005. Moreover, BD patients with comorbid migraine had a significantly higher rate of suicidal behavior, SP, PD, GAD and OCD (all p b .005). The impact of migraine on the clinical course of BD-I and BD-II was investigated by Brietzke et al. [40] comparing BD individuals with comorbid migraine (n = 115 or 33.9% of the total, of whom 88.7% were diagnosed as BD-I) vs. BD cases without migraine (n = 224, BD-I cases = 91.1% of this subset). Patients with comorbid migraine had more mood episodes (10 vs. 3%, p = .009), especially those with depressive polarity in first episode (53 vs. 48.2%, though not with a statistically significant difference) and were more often female than male (82.6 vs. 65.2%, p = .004). Also, BD patients with comorbid migraine had at least one (additional) psychiatric comorbidity in 72.6 vs. 47.4% of the cases ( p = .001), of which, at least one anxiety disorder (66.1 vs. 38.8%, p b .001), at least one general medical comorbidity (64.3 vs. 44.2%, p b .001), although none of the given medical comorbidities (obesity, seizure/epilepsy, asthma, diabetes mellitus, hypothyroidism and systemic hypertension) reached a statistically significant difference itself. No association was found between migraine and previous suicide attempts. Remarkably, while this study was almost exclusively focused on BD-I cases and no clear information was reported on BD-II cases; nonetheless, the authors concluded that migraine comorbidity may be associated with poorer outcomes in BD. BD–migraine comorbidity was associated with higher prevalence of psychiatric and general medical comorbidities, even if a difference between migraine comorbid and non-comorbid

BD patients was not observed after Bonferroni correction with respect to number of lifetime hospitalizations for depression/mania, rates of rapid cycling or history of suicide. A conference paper by Tavormina [41] reported the results of a private practice cross-sectional study carried on 423 consecutive newly admitted outpatients assessed for bipolar spectrum disorders according to 1999 Akiskal's prototypes [42]. Most of the eligible patients (n = 400) presented somatic symptoms, including colitis in 45% of the cases, gastritis (25%) and migraine in 8% of the cases. The author concluded that somatic syndromes represent an important pointer for the diagnosis of bipolar spectrum disorders at first visit. Nonetheless, the lack of details about the adopted methods hinders a better appreciation of the actual validity of the stated results and the claimed clinical implications. With regard to specific socio-demographic features associated to migraine–BD comorbidity, further insights came from two Canadian retrospective studies published within 2006–2007. Specifically, the Canadian Community Health Survey and Well-Being (82-617-XIE) component of the 2002 Canadian Community Health Survey (CCHS) (http://www23.statcan. gc.ca/imdb-bmdi/pub/5015-eng.htm) was analyzed by McIntyre et al. in a retrospective study published in 2006 [43] to investigate the prevalence and prognostic implications of predetermined comorbid general medical disorders, including migraine, among persons who screened positive for a lifetime manic episode (BD-I cases). Along with chronic fatigue, asthma, chronic bronchitis, multiple chemical sensitivities, hypertension and gastric ulcer, migraine was significantly more frequent in BD vs. general population (all, p b .05). In a companion-paper [24] using the same database as data source (total, N = 36,984), the persons with BD (2.4% of the sample) where found to have a relatively higher prevalence of migraine vs. general population (24.8 vs. 10.3%; p b .05), with a sex-specific prevalence of comorbid BD and migraine being 14.9% for males and 34.7% for females. Also, BD males with comorbid migraine were more likely to live in a low income household ( p b .05), to receive welfare assistance ( p b .05) and to report an earlier age of onset of BD ( p b .05). On the other side, BD females with comorbid migraine were more often comorbid with medical conditions ( p b .05) and were more frequently users of mental health care services ( p b .05). Overall, the odds of migraine in persons screened positive for BD-I were 40% higher than the corresponding odds for people screened negative for BD (OR = 1.40; 95% CI = 1.30 to 1.90). Notably, since the diagnostic instruments used by the CCHS were much more sensitive for euphoric manic presentations (BD-I) than BD-II and other “softer” expressions of BD, both reports based on the CCHS dataset focused just on BD-I cases. The 2002 CCHS database was analyzed also by Jette et al. in 2007 [44] to determine the prevalence of various conditions (including BD-I) associated with migraine. Overall, the prevalence of physician-diagnosed migraine was 15.2% for females and 6.1% for males. Those with BD (n = 938) and comorbid migraine were also


distinguished between those cases with further mental health comorbidities and those without. Among other findings about their medical history, the BD sub-group with BD-I and migraine who also had further mental health comorbidities reported use of antidepressant in 30.8% of the cases vs. 10.3% for BD–migraine without additional psychiatric comorbidities vs. 3.9% of BD patients with no migraine and, no additional psychiatric disorder vs. 17.6% of BD patients with an additional mental health disorder but without migraine. Finally, a recent retrospective study by Saunders et al. involving 412 BD patients and 157 healthy controls [45] suggested that the higher comorbidity rates of migraine seen in BD women vs. BD men, as already documented by most of the previous evidence, would be associated with specific features of illness and psychosocial risk factors that would differ by gender and impact outcome. Among 306 BD-I patients, 223 had no comorbid migraine. Migraine was comorbid in 29 BD-II patients and absent in 37 BD-II patients. History of suicide attempt was recorded in 110 (39%) cases with no comorbid migraine vs. 57 (45%) cases with comorbid migraine while a rapid cycling pattern was seen in 89 (31%) cases without migraine vs. 52 (41%) cases with comorbid migraine. Interestingly, this was the first study specifically exploring mixed symptoms, which were found in 102 (36%) of non-migraine cases vs. 63 (50%) of comorbid cases. Also, with respect to gender distribution, there was no statistically significant difference between rates of migraine in the BD-II group vs. BD-I one in the women (OR = 1.6; 95%, 0.8–3.0), yet group differences between women with BD and migraine vs. those without comorbid migraine included a marginally higher rate of mixed symptoms (OR = .19, 95% CI, 1.0–3.7). Since migraine was more common in BD subjects (31%) vs. controls (6%) and had an elevated risk in BD women compared to BD men (OR = 3.5; 95% CI, 2.1–5.8), the authors stated that the presence of migraine, especially in course of mixed symptoms, should prompt clinicians to account migraine as a common comorbidity which may ultimately affect the long-term course of BD itself. 3.4.1. Affective temperaments and soft-bipolar traits associated with BD–migraine comorbidity Whenever assessed, rates of cyclothymic temperament in BD patients with migraine were found higher compared to hyperthymic temperament based on their own respective Akiskalian definitions [46,47], especially for BD-II outpatients in whom diagnosis encompassed either patients with overt lifetime hypomania or cyclothymic temperament alone in some studies [19]. The presence of affective temperament (s) was seen in 45% of comorbid–migraine patients vs. 22% of non-comorbid patients. (p = .015) [22]. The clinically suggestive question whether patients experiencing a MDD with comorbid migraine may share a “bipolar spectrum trait” vs. non-migraine depressed cases was first addressed by a report made available in 2005 by Oedegaard et al. [25], carried out on 201 patients with an index episode of either

11

(hypo)mania or depression (BD-I, n = 36; BD-II, n = 51; MDD, n = 144 [of whom, n = 63 were MDD with migraine vs. n = 51 MDD without migraine]), with or without the following comorbidities: asthma, allergic rhinitis, allergic dermatitis, other allergies, thyroid disorders, attention deficit hyperactivity disorder (ADHD), Raynaud's syndrome, dyslexia and left or mixed handedness, and affective temperaments. Up to 70% of migraine MDD self-reported increased irritability during depression and nearly one fourth (22%) stated to have experienced seasonal variation in their depressive symptoms, also showing a higher prevalence of anxiety comorbidities (statistically significant differences were seen only for agoraphobia) and a higher occurrence of cyclothymic temperament. The relationship between migraine and type A behavior pattern (TABP), characterized by fairly stable urgency, impatience, irritability and competitiveness (a putative “bipolar trait” in patients with affective disorders) was documented by Norwegian researchers in a cohort of 23 BD-II patients and 42 MDD controls [48]. BDII patients had significantly higher type A scores, measured by the means of the Jenkins Activity Survey [JAS] [Form C] [49] compared to MDD controls (Wilks lambda = .851, F = 2.62, df [degrees of freedom] = 4.60, p b .05). Higher JAS scores were significantly associated with cyclothymic temperament (Wilks lambda = .728, F = 4.30, df = 4, p = .005) whereas lower JAD scores were associated with depressive temperament (Wilks lambda = .747, F = 3.13, df = 4, p = .026). Interestingly, TABP was not associated with migraine despite reports that migraine itself was associated with BD-II by the same authors in their own previous studies, and TABP largely overlaps with cyclothymic temperament and BD-II themselves. Nonetheless, it is worth noting that despite a broad definition of BD-II, the study was based on a small sample size and lacking a blind cross-sectional assessment of affective diagnoses (including BD-II itself). 3.4.2. Handedness and laterality preference in bipolar patients with comorbid migraine A higher occurrence of handedness/laterality prevalence in affective patients with comorbid migraine was documented since the 1980s [50], although not ubiquitously replicated by subsequent cross-sectional studies [22]. While the unrepresentativeness of the sample and confounding biases affecting some studies may have contributed to this inconsistent scenario, there is nonetheless a substantial body of data concerning laterality of brain function in patients with affective disorders, showing differences in the functioning of the two hemispheres in patients with bipolar depression [51–53], possibly linked to temporal lobe epilepsy too [54]. While patients with migraine experiencing pain exclusively on one side of the head, with no side-shift from one attack to the next, seem to be uncommon [55], many patients nonetheless complain pain predominantly located on one side of the head [56]. Therefore, in order to further investigate this topic, Fasmer and Oedegaard

12


performed a post-hoc analysis on 47 (out of 102) in- and outpatients with BD (or MDD) [57] based on the methods and instruments that were already reported in a previous study that they published [22]. Twenty three out 47 patients included in the study subset were diagnosed with BD-II (vs. 24 MDD). In the BD-II group, 13 had migraine with aura, 7 had migraine without aura and 3 had migraine aura without headache. Bilateral headache was seen in 8 BD-II patients vs. unilateral headache in 12 BD-II migraineurs, while left-sided headache was reported in 9 BD- II patients vs. right sided headache in 3 BD-II patients. Remarkably, as for their previous study [22], the definition of BD-II was broader than the one adopted by DSM-IV criteria, actually including the criteria proposed by Akiskal and collaborators [58]. Nonetheless, the authors concluded that the differential pattern seen in BD-II vs. MDD patients with regard to laterality of pain may suggest a differential functional impairment of the cerebral hemispheres in patients with migraine depending on their baseline affective disorder. The hypothesis of a differential connection between handedness preference across affective diagnoses was already suggested since the late 1980s [50] and supported by electrophysiological [59] and ECT studies [60]. This was tested by Fasmer and collaborators in 2005 [61], carrying a post-hoc analysis based on one of their own 2005 studies (sample size, n = 201, of whom n = 114 MDD) [25]. Specifically, hand preference was assessed by the Edinburgh inventory [62], and the patients were classified as having either right, left or mixed handedness. Right, left and mixed handedness were seen in 78 vs. 6 vs. 17% BD-I patients (n = 36), and 59 vs. 8 vs. 33% of BD-II cases (n = 51). Concerning affective temperaments, right vs. left vs. mixed handedness respectively were as follows: 62 vs. 13 vs. 26% for cyclothymic one; 44 vs. 11 vs. 44% had hyperthymic temperament; 50 vs. 25 vs. 25% were irritable, and 78 vs. 4 vs. 17% were depressive cases. BD-II patients had the greatest prevalence of non-right handedness (41%) compared to BD-I (23%) and MDD (26%), although not with at statistically significant different level. While the authors concluded that non-right handedness is associated with migraine and bipolar affective temperaments (“soft bipolarity”) in their affective disorders sample, they did not report stratified results or data about between-group comparison (including BD-I vs. BD-II). 3.4.3. Subjective painful somatic symptoms in migraine–BD patients Both the 2013 study by Hung et al. [63] and the 2013 report by Birgenheir et al., [64] covered the clinical aspects of pain conditions, including migraine (also) in course of BD. Specifically, the study by Birgenheir et al. [64] crosssectionally evaluated the prevalence of non-cancer pain (namely, arthritis, back pain, chronic pain, migraine, headache, psychogenic and neuropathic pain) in BD (and schizophrenia) US veterans (n = 5,195,551). Painful US veterans with BD (n = 96,186) were significantly more likely to have migraine (n = 4677 or 4.9%; adjusted OR =

4.67, 95% CI = 4.53–4.82). On the other hand, the study by Hung et al. [63] was specifically designed to compare the impacts of migraine and MDEs on depression, including those associated to BD, reporting a very high prevalence of migraine in BD-II (70.0% of the cases), followed by MDD (49.1%). Overall, patients with migraine had a higher risk of being in a current MDE compared to those without migraine (69.3 vs. 30.9%, OR = 5.5, 95% CI = 2.75–9.28, p b .001). Similarly, patients with migraine had a higher risk of a past suicide attempt history than those without (30.7 vs. 13.7%, OR = 2.79, 95% CI = 1.40–5.57, p b .001). Also, the regression model adopted by the authors indicated that the variance explained by migraine was significantly greater that those explained by MDE in physical and pain symptoms. Remarkably, none of the papers provided additional stratification of results with respect to BD-I vs. BD-II comparisons; specifically, the paper by Hung et al. [63] included only BD-II and MDD cases but not BD-I patients as affective diagnoses. 3.4.4. Longitudinal studies Although our search strategy failed to detect any prospective follow-up study specifically addressing the comorbidity between migraine and BD, a research report published in 2006 was nonetheless considered worth noting. Specifically, Angst et al. [65] tested seven different definitions of atypical depression in a prospective study of a large community sample followed for over 20 years, aiming at identifying a syndrome with a high correlation with female sex and BD-II (which in turn have been reported to be associated with migraine). Among other socio-demographic and clinical variables, the authors investigated the prevalence of migraine headache in atypical depression, finding an association with both atypical depression itself and female gender, along with SP, binge eating, neurasthenia and subjective cognitive impairment, but not tension headache which was unrelated with atypical depression/BD interface. In contrast to previous evidence, the study by Angst et al. (2006) failed to detect any significant association of major triadic atypical depression (most reliable definition for atypical depression among the seven alternative ones discussed in their text) of BD-II (atypical BD-II comorbid with migraine = 14.9% vs. non-atypical BD-II with comorbid migraine = 8.6%), but it was significantly more frequent in women, as migraine is. 3.5. Risk of bias within and across studies Comorbidities are best studied in representative samples because the prevalence of disease and the association among disorders may be sometimes altered in clinic-based samples or primary care settings [66]. Also, a Berkson's bias leading to under-estimation (or over-estimation) of the rate of occurrence of migraine–BD comorbidity may have occurred across the studies since most severe cases usually tend to seek medical care instead of participating in populationbased studies instead of and this is seen more in clinic-based


ones [67]. Moreover, incidence rates of a number of comorbid conditions increase with the frequency of migraine attacks, being higher for episodic vs. chronic migraine [68]. An additional bias potentially often encountered in such studies and accounted for in this systematic review, especially retrospective ones, is recall bias, due to the chance of neglect or poor regard of some BD patients towards their own previous (hypo-)manic episodes, which may ultimately affect even the prevalence estimates actually documented for migraine–BD comorbidity itself. Furthermore, a gender-related publication bias may have been occurred due to the unrepresentativeness of females subjects (clearly) documented to be follicular phase of the menstrual cycle [69]; similarly a potential confounding bias may be related to under-reported epileptic spikes eventually associated to some cases with migraine–BD comorbidity [70]. The afore mentioned biases potentially occurring within the studies may also affect between-studies comparisons, especially due to the lack of clearly detectable effect sizes/ z-scores in most of the selected studies. There is also inconsistency of the diagnostic criteria adopted for the definition of BD-II (sometimes broadened to include affective temperaments too) or IHS editions for the diagnosis of migraine, as well the inaccurate distinction between BD-I and BD-II or specific types of migraines by some studies, especially in low-powered and post-hoc studies.

4. Discussion 4.1. Study limitations Most of the limitations of the present systematic review (first of its kind at writing time) are intrinsically related to the potential biases of the included studies and should be taken into account in the interpretations of the results discussed herein. Specifically, ORs and RRs of the given studies included in the present review may differ across different studied samples according to gender, presence of any affective temperament, previous history of suicide, associated medical and/or psychiatric clinical pictures. Nonetheless, we submit that the publication period covered by the present systematic review was sufficient to reduce the chance of a publication bias as much as possible (though longitudinal studies are particularly needed). Similarly, it is worth notice that while the stringent selection criteria adopted through this review greatly reduced the number of contributes considered of adequate quality, this systematic selection ideally emphasized just the most rigorous and clinically relevant results on the topic. 4.2. Summary of evidence: brief pathophysiological hypothesis and clinical implications Based on the main findings of the present review, the prevalence of comorbid migraine and BD was remarkably high, with point-prevalence rates over 70% documented by

13

early studies carried out on Norwegian BD-II outpatients; nonetheless, it must be remarked that very low prevalence rates (around 2%) were reported as well, though major methodological constraints may affect the validity of such low-prevalence reports [9,38]. Similarly, rates of lifetime suicide attempts in BD patients with comorbid migraine were inconsistently reported, being high in most of the studies, especially for BD-II cases [8], although not replicated in some BD-I samples [40], at least with respect to migraine patients with aura without headache [25]. Among other clinically suggestive findings, irritability was repeatedly reported in course of migraine–BD comorbidity [22], being possibly associated with a seasonal pattern of affective illness and with cyclothymic temperament [48], particularly in case of comorbid migraine with aura [25], which has already been associated with a lifetime history of suicidal attempts [8]. The high prevalence rates of cyclothymic temperament documented by some studies may reflect an over-representation of BD-II females included in some study samples [19]. Since pain signals are conveyed through trigeminovascular projections to areas of the brain that are competent to produce migraine symptoms as well as depression, the comorbidity between migraine and BD may be due to a common dysregulation at the 5-hydroxytryptamine (5-HT) type-1B and/or 5-HT1D serotonergic [71–74] and/or D3 and/or D4 dopaminergic receptors [75,76]. A dysregulation may be involved in the pathogenesis of migraine as well as BD and related disorders, in the broader sense, therefore including panic attacks and hypomania too [77,78]. Also, both migraine and depression may be mediated by the cascade of neuronal events associated with central sensitization and its clinical marker, allodynia [79]. A putative common underlying pathophysiology between some migraine and BD cases may therefore contribute to explain the significant prescription trend overlap existing between migraine and BD [37], despite the warning about the risk for serotonin syndrome made by the US Food and Drug Administration for depressed patients with migraine exposed at higher doses of serotonergic drugs [80]. Environmental and genetic risk factors may produce a latent brain state that precipitates co-existing conditions sharing a common pathophysiologic mechanism. Specifically, regarding migraine and BD comorbidity, it has been proposed that several brain areas, including the hypothalamic, limbic and cortical regions may be activated during migraine and a depressive episode, whether or not associated to MDD or BD [81]. Indeed, genetic risk factors for the migraine/BD comorbidity have been studied in both genome-wide linkage and genome-wide association studies yielding significant findings for this phenotype [82]. Regrettably, longitudinal studies examining whether either BD or migraine impacts each other and prevalence rates for both conditions are nonetheless lacking. Thus, any hypothesis about a syndromic relationship between migraine and BD should also be taken cautiously.

14


Apart from pathophysiological hypotheses in support of the migraine–BD comorbidity, which are indeed beyond the scopes of the present review, it is nonetheless interesting to highlight that migraine was also associated to an earlier age of onset of the affective disorder, including BD [33] whenever comorbid, by most of the studies covered in the present review [22,25,35]. Additionally, the comorbidity between migraine and BD was reported to be preferentially associated with an index episode of depression rather than mania [40] and with more overall depressive episodes vs. mania [64], which may further increase the chance of receiving improper antidepressant monotherapies [44] rather than mood stabilizers including valproate [37] or lithium [35,37] and may thus increase the risk of abuse of pain-killer medications available over the counter. This scenario, along with the frequent occurrence of irritability—a plausible mixed feature of bipolar depression according to the specifiers recently introduced by the DSM-5 [83]—may also suggest a lower tolerance toward subjective pain among those BD-II migraineurs with cyclothymic temperament, which may in turn affect the adherence towards prescribed medications [84] and possibly concur to poor therapeutic outcome on antidepressants for bipolar depression or even to an increased risk for suicidality [85], especially in the view of a considerable risk for misdiagnosis in course migraine–BD comorbidity (Low et al., 2003). The choice of the optimal mood-stabilizer is also crucial in this population. Specifically, we submit that the eventual presence of “mixed features” associated to migraine–BD comorbidity may preferentially respond to valproate rather lithium, though the actual usefulness of sodium-channel antagonists for migraine (including lamotrigine largely prescribed to BD-II patients, especially in the presence of a seasonal pattern too) is still a matter of debate [86]. Indeed, enhancing the diagnostic accuracy and clinicians' attention to migraine– BD comorbidity should also ideally emphasize the interest for future pharmacological studies on the matter. With respect to overall medical and psychiatric comorbidities, with the sole exception of the report by Low et al. [35], evidence indicated a higher number of lifetime hospitalization and lower level of education and income [31,43], as well as a higher prevalence of family history for BD in case of comorbid migraine [21]. Moreover, migraine may precede BD [8] and the somatic presentations may be a “first-visit hallmark” for some bipolar outpatients [41] especially in case of severe somatic pain [64], thus urging attention considering the stigma still associated to BD worldwide, and that the evidence covered by the present review almost entirely focused on Western samples, thus precluding the appreciation of any pathoplastic differential effect of migraine on BD across cultures, if ever existing [87]. 4.3. Conclusions Psychiatric disorders, including BD, are commonly comorbid with migraine. Therefore any practitioner treating

one disorder is likely to encounter another, as already observed by Kraepelin [88]. While more methodologically rigorous studies are needed, especially longitudinal ones assessing the actual impact of mixed features in course of bipolar depression associated to migraine and the sensitive issue of related suicidality, both clinicians and researchers should bare attention to the co-occurrence of migraine and BD, especially in BD-II women with cyclothymic temperament, as this may ultimately also influence the therapeutic outcome towards prescribed medications. References [1] Headache Classification Subcommittee of the International Headache S. The International Classification of Headache Disorders: 2nd edition. Cephalalgia 2004;24(Suppl 1):9-160. [2] Merikangas KR, Angst J, Isler H. Migraine and psychopathology. Results of the Zurich cohort study of young adults. Arch Gen Psychiatry 1990;47:849-53. [3] Merikangas KR, Merikangas JR, Angst J. Headache syndromes and psychiatric disorders: association and familial transmission. J Psychiatr Res 1993;27:197-210. [4] Williams DE, Raczynski JM, Domino J, Davig JP. Psychophysiological and MMPI personality assessment of headaches: an integrative approach. Headache 1993;33:149-54. [5] Buse DC, Silberstein SD, Manack AN, Papapetropoulos S, Lipton RB. Psychiatric comorbidities of episodic and chronic migraine. J Neurol 2013;260:1960-9. [6] Mahmood T, Romans S, Silverstone T. Prevalence of migraine in bipolar disorder. J Affect Disord 1999;52:239-41. [7] Smitherman TA, Rains JC, Penzien DB. Psychiatric comorbidities and migraine chronification. Curr Pain Headache Rep 2009;13:326-31. [8] Ortiz A, Cervantes P, Zlotnik G, van de Velde C, Slaney C, Garnham J, et al. Cross-prevalence of migraine and bipolar disorder. Bipolar Disord 2010;12:397-403. [9] Holland J, Agius M, Zaman R. Prevalence of co-morbid bipolar disorder and migraine in a regional hospital psychiatric outpatient department. Psychiatr Danub 2011;23(Suppl 1):S23-4. [10] Juang KD, Wang SJ, Fuh JL, Lu SR, Su TP. Comorbidity of depressive and anxiety disorders in chronic daily headache and its subtypes. Headache 2000;40:818-23. [11] Mitsikostas DD, Thomas AM. Comorbidity of headache and depressive disorders. Cephalalgia 1999;19:211-7. [12] Oedegaard KJ, Neckelmann D, Mykletun A, Dahl AA, Zwart JA, Hagen K, et al. Migraine with and without aura: association with depression and anxiety disorder in a population-based study. The HUNT Study. Cephalalgia 2006;26:1-6. [13] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. [14] APA. Diagnostic and statistical manual for mental disordersIn: & Association AP, editor. Third ed. (DSM-III). ; 1980. [Washington, D.C.]. [15] Breslau N, Davis GC. Migraine, major depression and panic disorder: a prospective epidemiologic study of young adults. Cephalalgia 1992;12:85-90. [16] Fasmer OB, Akiskal H, Kelsoe J, Oedegaard K. Clinical and pathophysiological relations between migraine and mood disorders. Curr Psychiatry Rev 2009:93-109. [17] McGauran N, Wieseler B, Kreis J, Schuler YB, Kolsch H, Kaiser T. Reporting bias in medical research—a narrative review. Trials 2010;11:37. [18] Pompili M, Gonda X, Serafini G, Innamorati M, Sher L, Amore M, et al. Epidemiology of suicide in bipolar disorders: a systematic review of the literature. Bipolar Disord 2013;15:457-90.

M. Fornaro et al. / Comprehensive Psychiatry xx (2014) xxx–xxx [19] Fasmer OB. The prevalence of migraine in patients with bipolar and unipolar affective disorders. Cephalalgia 2001;21:894-9. [20] APA. Diagnostic and statistical manual for mental disordersFourth ed. (DSM-IV). ; 1994 [Washington D.C., USA]. [21] Dilsaver SC, Benazzi F, Oedegaard KJ, Fasmer OB, Akiskal KK, Akiskal HS. Migraine headache in affectively ill latino adults of Mexican American origin is associated with bipolarity. Primary care companion to the Journal of clinical psychiatry, 11; 2009302-6. [22] Fasmer OB, Oedegaard KJ. Clinical characteristics of patients with major affective disorders and comorbid migraine. World J Biol Psychiatry 2001;2:149-55. [23] Datta SS, Kumar S. Hypomania as an aura in migraine. Neurol India 2006;54:205-6. [24] McIntyre RS, Konarski JZ, Wilkins K, Bouffard B, Soczynska JK, Kennedy SH. The prevalence and impact of migraine headache in bipolar disorder: results from the Canadian Community Health Survey. Headache 2006;46:973-82. [25] Oedegaard KJ, Angst J, Neckelmann D, Fasmer OB. Migraine aura without headache compared to migraine with aura in patients with affective disorders. J Headache Pain 2005;6:378-86. [26] Hamelsky SW, Lipton RB. Psychiatric comorbidity of migraine. Headache 2006;46:1327-33. [27] Peterlin BL, Ward TN. Neuropsychiatric aspects of migraine. Curr Psychiatry Rep 2005;7:371-5. [28] Nuyen J, Schellevis FG, Satariano WA, Spreeuwenberg PM, Birkner MD, van den Bos GA, et al. Comorbidity was associated with neurologic and psychiatric diseases: a general practice-based controlled study. J Clin Epidemiol 2006;59:1274-84. [29] Endicott NA. Psychophysiological correlates of ‘bipolarity’. J Affect Disord 1989;17:47-56. [30] Baptista T, Uzcategui E, Arape Y, Serrano A, Mazzarella X, Quiroz S, et al. Migraine life-time prevalence in mental disorders: concurrent comparisons with first-degree relatives and the general population. Invest Clin 2012;53:38-51. [31] Nguyen TV, Low NC. Comorbidity of migraine and mood episodes in a nationally representative population-based sample. Headache 2013;53:498-506. [32] Chen YC, Tang CH, Ng K, Wang SJ. Comorbidity profiles of chronic migraine sufferers in a national database in Taiwan. J Headache Pain 2012;13:311-9. [33] Merikangas KR, Stevens DE. Comorbidity of migraine and psychiatric disorders. Neurol Clin 1997;15:115-23. [34] Davidoff RA. Migraine: manifestations, pathogenesis, and management. Philadelphia, PA, USA: F.A. Davis Company; 1995. [35] Low NC, Du Fort GG, Cervantes P. Prevalence, clinical correlates, and treatment of migraine in bipolar disorder. Headache 2003;43:940-9. [36] Pryse-Phillips WE, Dodick DW, Edmeads JG, Gawel MJ, Nelson RF, Purdy RA, et al. Guidelines for the diagnosis and management of migraine in clinical practice. Canadian Headache Society. CMAJ 1997;156:1273-87. [37] Oedegaard KJ, Riise T, Dilsaver SC, Lund A, Akiskal HS, Fasmer OB, et al. A pharmaco-epidemiological study of migraine and antidepressant medications: complete one year data from the Norwegian population. J Affect Disord 2011;129:198-204. [38] Weber NS, Fisher JA, Cowan DN, Niebuhr DW. Psychiatric and general medical conditions comorbid with bipolar disorder in the National Hospital Discharge Survey. Psychiatr Serv 2011;62:1152-8. [39] Ratcliffe GE, Enns MW, Jacobi F, Belik SL, Sareen J. The relationship between migraine and mental disorders in a population-based sample. Gen Hosp Psychiatry 2009;31:14-9. [40] Brietzke E, Moreira CL, Duarte SV, Nery FG, Kapczinski F, Miranda Scippa A, et al. Impact of comorbid migraine on the clinical course of bipolar disorder. Compr Psychiatry 2012;53:809-12. [41] Tavormina G. Are somatisation symptoms important evidence for an early diagnosis of bipolar spectrum mood disorders? Psychiatr Danub 2011;23(Suppl 1):S13-4.

15

[42] Akiskal HS, Pinto O. The evolving bipolar spectrum. Prototypes I, II, III, and IV. Psychiatr Clin North Am 1999;22:517-34 [vii]. [43] McIntyre RS, Konarski JZ, Soczynska JK, Wilkins K, Panjwani G, Bouffard B, et al. Medical comorbidity in bipolar disorder: implications for functional outcomes and health service utilization. Psychiatr Serv 2006;57:1140-4. [44] Jette N, Patten S, Williams J, Becker W, Wiebe S. Comorbidity of migraine and psychiatric disorders—a national population-based study. Headache 2008;48:501-16. [45] Saunders E, Nazir R, Kamali M, Ryan K, Evans S, Langenecker S, et al. Gender differences, clinical correlates, and longitudinal outcome of bipolar disorder with comorbid migraine. J Clin Psychiatry 2014, http://dx.doi.org/ 10.4088/JCP.13m08623 [Online ahead of print: April 15, 2014]. [46] Cassano GB, Akiskal HS, Perugi G, Musetti L, Savino M. The importance of measures of affective temperaments in genetic studies of mood disorders. J Psychiatr Res 1992;26:257-68. [47] Akiskal HS, Mallya G. Criteria for the “soft” bipolar spectrum: treatment implications. Psychopharmacol Bull 1987;23:68-73. [48] Oedegaard KJ, Neckelmann D, Fasmer OB. Type A behaviour differentiates bipolar II from unipolar depressed patients. J Affect Disord 2006;90:7-13. [49] Shekelle RB, Gale M, Norusis M. Type A score (Jenkins Activity Survey) and risk of recurrent coronary heart disease in the aspirin myocardial infarction study. Am J Cardiol 1985;56:221-5. [50] Messinger HB, Messinger MI, Graham JR. Migraine and lefthandedness: is there a connection? Cephalalgia 1988;8:237-44. [51] Bruder GE, Stewart JW, Towey JP, Friedman D, Tenke CE, Voglmaier MM, et al. Abnormal cerebral laterality in bipolar depression: convergence of behavioral and brain event-related potential findings. Biol Psychiatry 1992;32:33-47. [52] Bellani M, Marzi CA, Savazzi S, Perlini C, Cerruti S, Ferro A, et al. Laterality effects in schizophrenia and bipolar disorder. Exp Brain Res 2010;201:339-44. [53] Lenhart RE, Katkin ES. Psychophysiological evidence for cerebral laterality effects in a high-risk sample of students with subsyndromal bipolar depressive disorder. Am J Psychiatry 1986;143:602-7. [54] Altshuler LL, Devinsky O, Post RM, Theodore W. Depression, anxiety, and temporal lobe epilepsy. Laterality of focus and symptoms. Arch Neurol 1990;47:284-8. [55] Sjaastad O, Fredriksen TA, Sand T, Antonaci F. Unilaterality of headache in classic migraine. Cephalalgia 1989;9:71-7. [56] Brandt J, Celentano D, Stewart W, Linet M, Folstein MF. Personality and emotional disorder in a community sample of migraine headache sufferers. Am J Psychiatry 1990;147:303-8. [57] Fasmer OB, Oedegaard KJ. Laterality of pain in migraine with comorbid unipolar depressive and bipolar II disorders. Bipolar Disord 2002;4:290-5. [58] Akiskal HS, Bourgeois ML, Angst J, Post R, Moller H, Hirschfeld R. Re-evaluating the prevalence of and diagnostic composition within the broad clinical spectrum of bipolar disorders. J Affect Disord 2000;59 (Suppl 1):S5-S30. [59] Kropp P, Siniatchkin M, Stephani U, Gerber WD. Migraine–evidence for a disturbance of cerebral maturation in man? Neurosci Lett 1999;276:181-4. [60] Sackeim HA, Decina P, Epstein D, Bruder GE, Malitz S. Possible reversed affective lateralization in a case of bipolar disorder. Am J Psychiatry 1983;140:1191-3. [61] Fasmer OB, Akiskal HS, Hugdahl K, Oedegaard KJ. Non-righthandedness is associated with migraine and soft bipolarity in patients with mood disorders. J Affect Disord 2008;108:217-24. [62] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97-113. [63] Hung CI, Liu CY, Wang SJ. Migraine predicts physical and pain symptoms among psychiatric outpatients. J Headache Pain 2013;14:19. [64] Birgenheir DG, Ilgen MA, Bohnert AS, Abraham KM, Bowersox NW, Austin K, et al. Pain conditions among veterans with schizophrenia or bipolar disorder. Gen Hosp Psychiatry 2013;35:480-4.

16


[65] Angst J, Gamma A, Benazzi F, Silverstein B, Ajdacic-Gross V, Eich D, et al. Atypical depressive syndromes in varying definitions. Eur Arch Psychiatry Clin Neurosci 2006;256:44-54. [66] Wittchen HU, Lieb R, Wunderlich U, Schuster P. Comorbidity in primary care: presentation and consequences. J Clin Psychiatry 1999;60(Suppl 7):29-36 [discussion 7–8]. [67] Berkson J. Limitation of the application of fourfold table analysis to hospital data. Biometrics Bull 1946;2:47-53. [68] Aamodt AH, Stovner LJ, Langhammer A, Hagen K, Zwart JA. Is headache related to asthma, hay fever, and chronic bronchitis? The Head-HUNT Study. Headache 2007;47:204-12. [69] Lay CL, Payne R. Recognition and treatment of menstrual migraine. Neurologist 2007;13:197-204. [70] Belcastro V, Striano P, Parisi P. Interictal and periictal headache in patients with epilepsy: migraine-triggered seizures or epilepsytriggered headache? Eur J Neurol 2013;20:1333-4. [71] Ferrari MD, Saxena PR. On serotonin and migraine: a clinical and pharmacological review. Cephalalgia 1993;13:151-65. [72] Chugani DC, Niimura K, Chaturvedi S, Muzik O, Fakhouri M, Lee ML, et al. Increased brain serotonin synthesis in migraine. Neurology 1999;53:1473-9. [73] Wang W, Timsit-Berthier M, Schoenen J. Intensity dependence of auditory evoked potentials is pronounced in migraine: an indication of cortical potentiation and low serotonergic neurotransmission? Neurology 1996;46:1404-9. [74] Hamel E. Serotonin and migraine: biology and clinical implications. Cephalalgia 2007;27:1293-300. [75] Peroutka SJ. Dopamine and migraine. Neurology 1997;49:650-6. [76] Barbanti P, Fabbrini G, Ricci A, Pascali MP, Bronzetti E, Amenta F, et al. Migraine patients show an increased density of dopamine D3 and D4 receptors on lymphocytes. Cephalalgia 2000;20:15-9.

[77] Gardner K. The genetic basis of migraine: how much do we know? Can J Neurol Sci 1999;26(Suppl 3):S37-43. [78] Gelernter J. Genetics of bipolar affective disorder: time for another reinvention? Am J Hum Genet 1995;56:1262-6. [79] Finocchi C, Villani V, Casucci G. Therapeutic strategies in migraine patients with mood and anxiety disorders: clinical evidence. Neurol Sci 2010;31(Suppl 1):S95-8. [80] Boyer EW, Shannon M. The serotonin syndrome. N Engl J Med 2005;352:1112-20. [81] Burstein R, Jakubowski M. Neural substrate of depression during migraine. Neurol Sci 2009;30(Suppl 1):S27-31. [82] Oedegaard KJ, Greenwood TA, Johansson S, Jacobsen KK, Halmoy A, Fasmer OB, et al. A genome-wide association study of bipolar disorder and comorbid migraine. Genes Brain Behav 2010;9:673-80. [83] APA. Diagnostic and statistical manual for mental disordersFifth ed. (DSM-5). ; 2013 [Washington, D.C., USA]. [84] Fornaro M, De Berardis D, Iasevoli F, Pistorio ML, D'Angelo E, Mungo S, et al. Treatment adherence towards prescribed medications in bipolar-II acute depressed patients: relationship with cyclothymic temperament and "therapeutic sensation seeking" in response towards subjective intolerance to pain. J Affect Disord 2013;151:596-604. [85] Fornaro M, Martino M, De Pasquale C, Moussaoui D. The argument of antidepressant drugs in the treatment of bipolar depression: mixed evidence or mixed states? Expert Opin Pharmacother 2012;13:2037-51. [86] Chiossi L, Negro A, Capi M, Lionetto L, Martelletti P. Sodium channel antagonists for the treatment of migraine. Expert Opin Pharmacother 2014;15:1697-706. [87] Fornaro M, Clementi N, Fornaro P. Medicine and psychiatry in Western culture: ancient Greek myths and modern prejudices. Ann Gen Psychiatry 2009;8:21. [88] Kraepelin E. Manic-depressive insanity and paranoia; 1921.

Prevalence and clinical features associated with bipolar disorder polypharmacy: a systematic review.

The prevalence and predictors of bipolar and borderline personality disorders comorbidity: Systematic review and meta-analysis.

Clinical features, comorbidity, and cognitive impairment in elderly bipolar patients.

Migraine headache and bipolar disorder comorbidity: A systematic review of the literature and clinical implications.

Prevalence and associated factors of tuberculosis and diabetes mellitus comorbidity: A systematic review.

A meta-analysis and systematic review of the comorbidity between irritable bowel syndrome and bipolar disorder.

Clinical Features and Treatment of Penile Schwannoma: A Systematic Review.

The prevalence of pain in bipolar disorder: a systematic review and large-scale meta-analysis.

Cellular models to study bipolar disorder: A systematic review.

Prevalence rates and clinical implications of bipolar disorder "with mixed features" as defined by DSM-5.

The prevalence of bipolar disorder in general primary care samples: a systematic review.

Clinical tooth preparations and associated measuring methods: a systematic review.

Childhood maltreatment and unfavourable clinical outcomes in bipolar disorder: a systematic review and meta-analysis.

Shared clinical associations between obesity and impulsivity in rapid cycling bipolar disorder: a systematic review.

Hyperprolactinemia and medications for bipolar disorder: systematic review of a neglected issue in clinical practice.

Bipolar disorder and stigma: a systematic review of the literature.

Clinical Outcomes Associated with Home Mechanical Ventilation: A Systematic Review.

Clinical risk factors for bipolar disorders: a systematic review of prospective studies.

Clinical Features and Treatments of Syphilitic Uveitis: A Systematic Review and Meta-Analysis.

Comorbidity of bipolar disorder and eating disorders.

Bipolar disorder in idiopathic dystonia: clinical features and possible neurobiology.

Clinical features of prefrail older individuals and emerging peripheral biomarkers: a systematic review.

Clinical features and treatment of vulvar Merkel cell carcinoma: a systematic review.

Clinical Features and Surgical Treatment of Schwannoma Affecting the Base of the Tongue: A Systematic Review.