1. Introduction
Headache disorders are ranked as the fourth most common among neurological disorders in terms of disease prevalence (age standardized) and years lived with the disease (YLD) in both sexes [
1]. Headache disorders usually affect both young and older adults (40%), thus resulting in the loss of many working hours (absenteeism) and lower productivity (presenteeism) [
2]. Among all headache disorders, migraine has the third highest prevalence of all medical illnesses. According to the World Health Organization (WHO), disability due to migraine is greater even than disability attributed to cardiovascular disorders [
3,
4,
5,
6].
Multiple Sclerosis (MS) is an inflammatory disorder of the central nervous system with a high prevalence, ranking sixth in frequency among neurological disorders. Not only MS as a primary disease but also comorbidities frequently associated with MS, affect the quality of life of these patients, resulting in reduced productivity, as measured in disability-adjusted life years (DALYs) [
7].
A definite association between headache disorders, and specifically migraine and MS, is not yet proven. Recent studies evaluated the incidence of headache disorders to be up to 64% in MS patients [
8,
9,
10,
11]. Many hypotheses have been speculated about the pathophysiological mechanism of headache in MS patients and the various co-factors involved. Experimental cortical demyelination that accelerates cortical spreading depression, the presence of meningeal and cortical B-cell follicles, and the specific location of lesions attempt to explain headache incidence in MS patients [
12,
13,
14,
15]. Freedman and Gray, who have studied the presence of headache in patients with MS during an attack, showed that nearly half of the patients had brain stem involvement [
16]. Epidemiological studies confirm the aforementioned results and imaging studies have indicated that midbrain/periaqueductal (PAG) MS lesions are associated with an increased incidence of headache [
17,
18,
19]. PAG as well as other midbrain structures and their connections to the rostral ventromedial medulla and the dorsolateral pontomesencephalic tegmentum have been associated with the occurrence of headache, by decreasing the firing of the nociresponsive neurons of the dorsal horn [
17,
20]. Additionally, retrospective observational studies that included patients who underwent deep brain electrical stimulation showed that migraine-like attacks were generated via the stimulation of PAG [
21,
22]. Other locations that have also been related to headache are the substantia nigra, the red nucleus and the hypothalamus, all of which are linked to PAG by afferent and efferent signaling [
17]. Among the co-factors, MS therapies are considered to play a role.
With the development of research regarding B-cell implications in the pathophysiology of MS, new depleting therapies that target either B-cells alone or both T- and B-cells have been introduced in MS clinical practice (Rituximab, Ocrelizumab, Ofatumumab, Cladribine and Ublituximab). Common adverse events of B-cell therapies usually include lymphopenia, susceptibility to infections and an increased incidence of malignancies. The majority of clinical trials focus on severe adverse events (SAEs), such as the aforementioned. Minor side effects such as headache and dizziness are usually not systematically reported and are often underrated, resulting in missing data.
B-cell therapies that have been used in MS as well as in many other diseases (e.g., lymphoma, rheumatoid arthritis, chronic lymphocytic leukemia, pemphigus, etc.), have been insufficiently related to headache as an adverse event. However, this finding occurred mostly as an Infuse Related Reaction (IRR), as reported by the Summary of Product Characteristics (SmPC) of each drug [
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33]. Headache was noted largely in the real-world evidence status and not in the pre-market clinical studies, resulting in missing data regarding pain characteristics that would allow classification.
Headache as an adverse event has been assessed, but not meticulously reported, in many clinical trials among the different available MS therapies. A recent meta-analysis, which included all studies on interferon-beta (INF-β) in MS [
34], showed increased headache incidence over placebo in patients receiving INF-β (Relative Risk 1.16 [95% Confidence Interval 1.02–1.33],
p-value = 0.02), who have reported de novo headaches, the aggravation of pre-existing headaches or change in the clinical features. INF-β was one of the most widely prescribed immunomodulatory MS treatment until now. According to the International Classification of Headache Disorders 3rd edition (ICHD-3) criteria [
35], this type of headache is categorized as secondary (§ 8.1.10 Headache attributed to long-term use of non-headache medication) and studies speculate potential underlying pathogenetic mechanisms related to INF-β, as the headache attack has a close temporal relation to the subcutaneous injection of the drug. In vitro studies show that IFN-β influences the neuronal excitability in neocortical pyramidal neurons [
36,
37], which seems to play an important role in the pathogenesis of primary headaches [
38]. Cytokine level changes (e.g., serum tumor necrosis factor alfa and IL-10) should also be evaluated, as high serum levels are reported in headache patients [
37,
39,
40]. Patients receiving fingolimod in a FREEDOMS trial had an increased incidence of headache compared to placebo (26.6% vs. 23%,
p-value = 0.007) [
41]. Furthermore, fingolimod is associated with posterior reversible encephalopathy syndrome (PRES) and there are several cases that support a temporal correlation of fingolimod treatment initiation with new and persistent headaches [
42,
43]. A direct endothelial modulatory effect has been hypothesized by the authors as a mechanism of action [
43,
44,
45,
46].
A possible common denominator might be the increase in IL-10 levels. Munno et al. [
39] showed that IL-10 levels were increased in patients during migraine attacks and subsequently decreased after sumatriptan treatment. Both INF-β and fingolimod result in an increase in IL-10 expression in circulating B-cells [
47,
48,
49] and myeloid cells [
50].
Taking the above into consideration together with the higher IL-10 levels produced by the reconstituted B-cells particularly following anti-CD20 treatment [
50], we hypothesize that anti-B-cell therapies might thereby contribute to increased headache incidence in patients with MS.
However, to date, no previous studies exist that review the incidence of headache as an adverse event in MS patients receiving different anti-B-cell therapies. To provide additional evidence, we conducted a systematical review and meta-analysis to investigate the association of B-cell targeted therapies with headache incidence in MS patients.
3. Results
We identified 258 records that were eligible for inclusion, including 30 RCTs (
Figure 1). The included trials evaluated four interventions (Rituximab, Ocrelizumab, Cladribine, Ofatumumab). There were no data available concerning headache as an adverse event in clinical trials of Ublituximab.
Summary of Study Retrieval and Identification for Meta-analysis according to PRISMA statement.
In total, nine studies [
59,
60,
61,
62,
63,
64,
65,
66,
67] (3785 patients) were included in the final analysis, whose characteristics are shown in
Table 1.
Cladribine (a synthetic purine nucleoside analog that inhibits DNA synthesis and ribonucleotide reductase) was assessed in three studies (2107 participants), Ocrelizumab (a humanized anti-CD20 monoclonal antibody) in two studies (866 participants), Ofatumumab (a fully human anti-CD20 monoclonal antibody) in two studies (269 participants) and Rituximab, a chimeric anti-CD20 monoclonal antibody, also in two studies (543 participants), as presented in
Table 2.
The presence of headache was not associated, in general, with B-cell depleting therapy (RR 1.12 [95% CI 0.96–1.30],
p-value = 0.15, I
2 = 9.32%, Q = 7.42,
p = 0.492) (
Figure 2).
By extracting the Risk Ratio of headache incidence from the included randomized control trials for each drug, a statistically significant association between headache and treatment with Cladribine was uncovered (RR 1.20, [95% CI 1.006–1.42],
p-value = 0.042, I
2 = 0%, Q = 1.07,
p = 0.586) (
Table 2,
Figure 3).
A sub-group analysis of the included studies for every monoclonal antibody against the B-cell line (i.e., Ofatumumab, Rituximab and Ocrelizumab) and for each drug separately showed no apparent connection between the development of headache and the different treatment options (
Supplement Figures S1–S4).
According to ROB-II, all studies fall into the category of some concerns of bias. The main reason is the absence of a finalized pre-specified analysis plan before unblinding the outcome data that were available for headache assessment analysis. As a result, there were some concerns in the domain of bias in the selection of the reported result. In all other ROB-II domains there was low risk of bias (
Supplement Figure S5 and S6).
Most of the headache-related data we present here were extracted from clinicaltrials.gov and were not available in the respective trial’s publications. There is a concern of possible publication bias, especially since headaches are considered secondary adverse effects, but this concern is minimized due to the obligatory reporting in clinicaltrials.gov. The analysis of our results indicates that it probably does not have a big effect in this case (Egger’s test
p-value = 0.18,
Supplement Figure S7).
4. Discussion
We performed a meta-analysis and compared the relative risk of headache in patients with MS under treatment with B-cell depletion therapies to those receiving placebo and found that B-cell targeted therapies are not associated overall with an increased risk of developing secondary headache. Furthermore, we conducted a Risk Ratio analysis separately for each drug included in the original analysis. No apparent correlation was found between the development of headache and monoclonal antibodies against B-cells (
Supplement Figures S1–S4). Interestingly, an association of Cladribine and headache incidence was identified (
Figure 3). As comparison between B-cell therapies and other MS treatments (head-to-head clinical trials) was unfeasible, due to cross-over studies and many confounding factors. A placebo comparator was a priori decided for the study analysis.
The route of administration plays an important role, especially when it involves parenteral administration, injection related adverse events and high frequency dosing regimen. B-cell therapy studies are more appropriate for investigating headache incidence, as they do not bear the risk of injection-associated headache, as opposed to interferons that are administered subcutaneously with a high frequency. Thus, the above results depict the active-drug’s action more, rather than the route of administration.
Although headache is among the commonly reported side effects of Cladribine, an induction mechanism has not been proposed so far. The modulation of nociception by perturbations introduced in purinergic signaling cascades may explain headaches. The contribution of purinergic signaling in pain conduction encompasses both vasomotor, neuronal and cortical processes, closely following the dissemination of purinergic receptors in each tissue [
68]. Of note, the ATP-mediated migrainogenic activation of trigeminal nerves has been shown to be regulated by the calcitonin gene-related peptide (CGRP) [
69]. Experimental models have furthermore determined that the purinergic modulation of nociception or the algogenic activation of the trigeminal ganglia may occur either directly, i.e., via activation of algogenic P2-calcium receptors [
70], or indirectly by modulating the nitroxidergic system peripherally [
71].
In addition, early studies of radioligand binding affinity demonstrated that Cladribine binds as an agonist to the adenosine receptor sub-types A
1 and A
2A [
72]. This may be relevant to the emergence of headache in Cladribine-treated patients, as A
2A receptor-mediated signaling is thought to be central to the pathophysiology of headaches [
73]. Interestingly, A
1 receptor mRNA and protein levels were found to be reduced in MS brain tissue [
74] and A
1 receptor deficiency increased proinflammatory responses and aggravated experimental allergic encephalomyelitis in mice [
75]. This may suggest that transcriptional control or the transcript degradation of the A
1 receptor gene is perturbed during MS-associated neuroinflammation. It is therefore conceivable that A
2A receptors are disproportionately expressed in the CNS of MS patients, resulting in excess A
2A receptor-mediated signaling upon treatment with Cladribine. A possible purinergic signaling cascade and disproportional expression of A
2A over A
1 receptors, as mechanism of action, should be investigated through further studies.
This study is not without caveats. Limitations of this study include the heterogeneity of the drugs under investigation, the different categorization of headache disorders and MS criteria and variability in methods used. More specifically, the chronological spectrum of the included studies in the analysis ranges from 2008 to 2019. During this period, different diagnostic criteria for both MS and headache disorders were applied. Furthermore, all the included studies share in common that they report every type of headache together as one group, named in general as “head pain”, without sub-dividing into further categories such as migraine, TTH or other type of primary or secondary headache disorder.
As stated previously, minor adverse events such as headache were not the primary study interest of the clinical trials. For this reason, several studies have disclosed that the data were collected by a non-systematic assessment. Consequently, there were missing data in the clinical trials of our interest.
We should also note that even though all the treatments analyzed in this study have a reported direct or indirect B-cell mediated mechanism of action, they do not share similar pharmacodynamics. Even the three so called “anti-CD20 monoclonal antibodies” (Rituximab, Ocrelizumab, Ofatumumab), despite sharing the “same” target (B-cells that express CD20), differ in the exact neuroimmunomodulatory effects. Moreover, Cladribine is a synthetic purine nucleoside analog that inhibits DNA synthesis and ribonucleotide reductase. Once inside the cell, Cladribine is activated mostly in lymphocytes, after being triphosphorylated by the enzyme deoxyadenosine kinase (dCK). Activated, the triphosphorylated Cladribine is incorporated into mitochondrial and nuclear DNA, which triggers apoptosis [
76]. Due to the extremely specific ratio of dCK to 5′-NTase needed to activate and accumulate enough Cladribine to induce apoptosis, only lymphocytes are uniquely vulnerable [
77]. Within the lymphocyte pool, Cladribine targets B-cells more than T cells. The CD3+ T cells remain suppressed longer than the CD19+ B-cells, and the CD4+ cells are affected more than the CD8+ cells.
Overall, the shift to B-cell therapies in the management of MS provided more and better tolerated treatment options. However, as newly introduced players emerge in the field of MS therapeutics, increased pharmacovigilance is needed. With every new treatment introduced, the expectation of a more individualized, targeted and better tolerated medical care raises.