New relapse of multiple sclerosis and neuromyelitis optica as a potential adverse event of AstraZeneca AZD1222 vaccination for COVID-19.

We report on nine patients (eight cases of MS and one case of NMOSD) who presented a disease relapse in close temporal association with their first AZD1222 vaccination dose against COVID-19. These patients had been stable for a median period of six years, with no evidence of disease activity and no change in their medication. After a median of 13 days (7 to 25 days) from vaccination, they developed a new relapse with increased disability and new lesions on magnetic resonance imaging. Although this association may be rare, it might be an adverse event of AZD1222.

Introduction

The key point for controlling the coronavirus disease 2019 (COVID-19) pandemic is vaccination. Several safe and effective vaccines have now been developed against COVID-19, which, together with wearing a face mask, social distancing and frequent hand sanitation, can potentially slow the spread of the virus (Center for Disease Control and Prevention, 2021). Patients with multiple sclerosis (MS) are correctly encouraged to be vaccinated against severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), the virus responsible for COVID-19 (National, 2021). Given the urgency generated by the COVID-19 pandemic, COVID-19 vaccines have been granted emergency-use authorization in several countries on the basis of results from ongoing trials (Coyle et al., 2021). There has been no trial on any vaccine for patients diagnosed with autoimmune diseases such as MS or neuromyelitis optica spectrum disorders (NMOSD). There are some recommendations regarding the disease-modifying drug (DMD) that they are taking at the time of receiving the vaccine (MSIFInternationalMS, 2021). Hesitancy to receiving vaccinations has been reported among patients with MS (Diem et al., 2021), with particular concern and unwillingness observed in relation to COVID-19 vaccination (Xiang et al., 2021). Vaccines rely on both the innate and the adaptive arm of the immune system, and interact in a complex and complementary manner to generate the immunological memory. (Marshall et al., 2018). According to MS societies and specialists’ recommendations, “there is no vaccine preference for those living with MS. A fever can make your MS symptoms worse temporarily, but they should return to prior levels after the fever is gone. Even if you have side effects, it's important to get the second dose of the vaccine for it to be effective” (National, 2021).

Brazil has been hit hard by COVID-19. The initial vaccines approved by the Brazilian regulatory agency ANVISA were CoronaVac (Sinovac Life Sciences, Beijing, China), an inactivated virus vaccine (Zhang et al., 2021), and AstraZeneca AZD1222 (developed in Oxford, United Kingdom), consisting of a replication-deficient chimpanzee adenoviral vector ChAdOx1 that contains the SARS-CoV-2 structural surface spike protein. (Chagla, 2021; Voysey et al., 2021) Later on, other vaccines were also approved for controlling the COVID-19 pandemic in Brazil. In this context, here we report on nine patients with previously controlled MS or NMOSD who developed a new relapse in close temporal relationship with AZD1222 vaccination in Brazil. At the time of this report, only these two vaccines had been approved in Brazil.

Methods

This study was approved by the Ethics Committee at Universidade Metropolitana de Santos, SP, Brazil. Due to the characteristics and importance of the adverse event, the Committee exempted the patients from signing a consent statement, although confidentiality remained guaranteed. The exception was for Figure 2, which is presented with the signed consent of the patient. Neurologists working at demyelinating disease units who belong to a discussion group in WhatsApp started to post cases of patients with MS or NMOSD who had no evidence of disease activity (NEDA) (Havrdova et al., 2009) for years, but then, a few days after anti-SARS-CoV-2 vaccination, presented a demyelinating relapse. A separate study group was created and neurologists who wanted to report on their cases sent detailed information on this adverse event in a file created for this purpose. Additional Ethics Committee approval was obtained whenever necessary. Patients were included if they had not presented a clinical relapse, a new lesion on MRI or an increased or new neurological disability for at least one year. The patient should be using the same medication for MS for the whole time he/she was stable. The new relapse, occurring post-vaccine, should be in close temporal relation to the inoculation and alter the neurological examination and the magnetic resonance image (MRI). The present report is essentially descriptive.

Results

Table 1 summarizes the data on the nine patients who presented a new relapse after a median period of six years of NEDA for MS and 10 years of stability for NMOSD. Briefly, there were three men and six women, of median age 33 years, comprising eight cases of MS and one case of NMOSD. All the patients presented a relapse after a median interval of 13 days (range: 7 to 25 days), following their first dose of AZD1222. None of the patients with MS had COVID-19 prior to vaccination, but the patient with NMOSD had been infected with SARS-CoV-2 eight months before her vaccination and had recovered well.

Table 1

Summarized data on patients with MS or NMOSD who presented a relapse in close temporal relation to the AZD1222 vaccination for protection against COVID-19.

Sex	Age (years)	Disease	Duration of controlled disease (years)	Medication in use	General AEs to first dose of AZD1222	Days elapsed between vaccine and relapse	Clinical manifestation of relapse	Magnetic resonance	Treatment	Recovery*	General AEs to second dose of AZD1222
F	22	RRMS	5	Fingolimod	Headache	7	Facial paralysis, hemiparesis, ataxia	Non-Gd tumefactive lesion (Fig 2)	Pulsotherapy methylprednisolone	Not yet recovered	Not done
F	32	RRMS	2	Dimethyl fumarate	Myalgia, fever	10	Loss of vision and papillitis in the left eye	New Gd+ lesions in the left eye	Pulsotherapy methylprednisolone Immunoglobulin	Partial recovery in 3 weeks	Not done
M	35	SPMS	3	Natalizumab	Myalgia, fever, fatigue	7	Worsening of disability, could not walk, severe weakness of both legs	High lesion load, new lesions	Oral prednisone	Not yet recovered	Not done
F	30	RRMS	1	Natalizumab	Myalgia, chills, fever	25	Right hemiparesis	New Gd+ lesions	Pulsotherapy methylprednisolone	Recovered in 3 weeks	Not done
F	42	RRMS	3	Fingolimod	Chills	15	Rapidly progressive weakness in both arms, grade III at its worst	New Gd+ lesions in spinal cord, T2 level	Pulsotherapy methylprednisolone	Recovered in 3 weeks	No adverse events
M	35	RRMS	4	Teriflunomide	Chills, pain on the site of injection	20	Incoordination of right arm and hand	New Gd+ lesions in brainstem	Pulsotherapy methylprednisolone	Not yet recovered	Not done
M	51	PPMS	2	No drug	Headache	25	Hypoesthesia in both arms	New Gd+ lesions in cervical cord	No treatment	Not yet recovered	Not done
F	32	RRMS	6	Glatiramer acetate	Headache	7	Motor and sensitive deficits in right leg and foot	New Gd+ lesions + new lesions	Pulsotherapy methylprednisolone	Not yet recovered	Not done
F	62	NMOSD	8	Azathioprine	Unspecific	7	Loss of vision in the left eye	New Gd+ lesions in left optic nerve	Pulsotherapy methylprednisolone	Recovered vision in 3 weeks	Not done

Abbreviations: M: male, F: female; RRMS: relapse-remitting multiple sclerosis; SPMS: secondary progressive MS; PPMS: primary progressive MS; NMOSD: neuromyelitis optica spectrum disorders; AE: adverse event; Gd: gadolinium.

Recovery* - when “not yet recovered” is stated, it means that the patient is still with some degree of neurological disability after at least 40 days from the onset of the relapse.

There were two patients inoculated with CoronaVac who were excluded from the study for not fulfilling the abovementioned inclusion criteria. The first patient had shown a small new T2 lesion in her MRI six months before vaccination. Therefore, she was not considered to be in NEDA for at least one year and her presence in this study could not be justified. The second patient did not show MRI changes during the post-vaccine relapse, which was mostly sensory and resolved within three weeks without treatment. Again, inclusion of this patient could have been a confounding factor in a group with distinctive relapses.

The changes to the nine patients’ disability status during the relapse are shown in Fig. 1 . None of the drugs used for MS treatment and none of the clinical presentations of the disease seemed to influence the onset of a relapse (Table 1). None of the patients had interrupted their treatment or changed therapy at any time over the years in which they had remained stable regarding their MS/NMOSD .

Fig. 1

Increase in disability of each patient (A = baseline; B = relapse post-vaccine) 95% CI for mean difference: (−2.017; −0.608); t-test of mean difference = 0 (vs ≠ 0): t-value = −4.41 p-value = 0.003.

Magnetic resonance image (MRI) of Patient 1 listed in the table. In relation to an MRI done 11 months previously, a tumefactive lesion appeared at the middle cerebellar peduncle to the right, and in the pontine tegmentum to the left, extending to the topography of the trigeminal nucleus, without contrast enhancement.

Discussion

Here, we discuss the cases of nine patients with MS or NMOSD with NEDA for long periods, i.e. 1–8 years, who had relapses as early as 7–15 days after administration of the first dose of the AstraZeneca AZD1222 vaccine against COVID-19. The most common features were new Gd+ lesions and higher lesion loads, shown on MRI.

Appearance of autoimmunity after administration of vaccines has already been described in the literature, including vaccines against SARS-CoV-2. Usually, these patients develop self-limiting symptoms that may include Guillain-Barré syndrome (GBS) (Arya et al., 2019; Martín Arias et al., 2015), autoimmune thrombocytopenia (Greinacher et al., 2021; Kuter, 2021; See et al., 2021), encephalitis (Ribeiro et al., 2021) or lupus (Vista et al., 2012), among others. Moreover, the concept of autoimmune/inflammatory syndrome induced by adjuvants (ASIA), described by Shoenfeld in 2011, (Shoenfeld and Agmon-Levin, 2011) has been extensively debated. A cohort study on 500 subjects who developed ASIA identified that around 35% had neurological manifestations, 73% arthralgia and 65% chronic fatigue (Watad et al., 2019). It is noteworthy that some of these patients had ASIA after receiving the influenza vaccine, for which the adjuvant is the same one used in AZD1222.

Vaccines may be developed on different bases, using living attenuated, non-replicating or inactivated pathogens, for example. Specific components may also be introduced, such as purified proteins and, more recently, nucleic-acid vaccines made of mRNA, plasmid DNA or viral DNA vectors that code for a specific antigen. This latter category forms the basis for the current COVID-19 vaccines, for which the antigen is either the full length or a portion of the envelope SPIKE protein (Gebre et al., 2021). However, the presence of the adjuvants, another component of vaccines with an unquestionably essential function, must not be neglected.

In association, the antigen and the adjuvants have the aim of inducing a robust and long-lasting specific response from both humoral and cellular components of the immune system. For this, three signals are essential: (1) specific antigens need to be recognized by the major histocompatibility complex – T cell receptor (MHC-TCR) interaction; (2) costimulatory engagement need to occur to favor TCR signaling pathways; and (3) pro-inflammatory signals supplied by cytokines lead to functional involvement of the T lymphocytes towards the Th1, Th2 or Th17 phenotypes (Smith-Garvin et al., 2009).

Although the first signal is very specific, i.e., through recognition of the specific vaccinal epitope, the second and third signals are much more unspecific (Owen et al., 2021). Adjuvants are known to activate innate immunity receptors, such as toll-like receptors (TLRs) (He et al., 2015). The TLR signaling pathway culminates in phosphorylation of the transcription factor NF-KB, for which responsive elements are found in more than 1500 inflammatory genes (Taniguchi and Karin, 2018). These elements include cytokines, chemokines, enzymes and also antigen presentation molecules like MHC I and II (Amaral et al., 2020). Moreover, the Nod-like receptor (NLR) triggers assembly of an inflammasome, which further activates caspase-1, for release of IL-1β and IL-18 by means of pyroptosis (Wong et al., 2021). Together, these molecules correlate with the mild flu-like symptoms that may be observed soon after administration of most vaccines. However, although essential, given that these molecules supply T lymphocytes with enough stimuli for B and T clonal activation and expansion for an appropriate immune response, they may also trigger undesired responses, such as autoimmune reactions.

The human organism contains a very broad repertoire of lymphocytes. These lymphocytes are usually under the control of peripheral mechanisms of tolerance, a role that is mostly played by CD4+Foxp3+ regulatory cells (Arpaia et al., 2015). In the case of immune-mediated diseases, such as MS and NMOSD, this repertoire is enriched by anti-myelin or anti-aquaporin-4 antigen-specific lymphocytes and, besides Treg suppression, they are also targets for immunomodulatory therapy.

The inflammatory effect of adjuvants may blunt the mechanisms of tolerance, thereby allowing these self-reacting clones to surge. Thus, lymphocytes with specificities unrelated to the vaccinal antigen may be activated by the abundance of cytokines produced. This phenomenon is known as bystander activation and may be observed under many different circumstances (Wong et al., 2021). In fact, investigation of the effects of adjuvants for induction of autoimmunity dates back to studies by Thomas Rivers, who took the observations made by Louis Pasteur after inoculation of rhesus monkeys with brain extracts and correlated these with the discovery of Freund´s adjuvant, by Jules Freund, as previously reviewed (Van Epps, 2005). Whereas Louis Pasteur needed multiple inoculations for induction of paralysis, Rivers needed only one inoculation if he used brain extracts in association with Freund´s adjuvant.

Importantly, it is known that both IL-1 and IL-6 are potent blockers of Treg function (Korn et al., 2008), while they favor involvement of the encephalitogenic Th17 subtype (Ivanov et al., 2006). The AZD1222 vaccine contains the adjuvant MF59®, which is an oil-in-water emulsion containing squalene (4.3%) in citric acid buffer with the surfactants Tween 80 (0.5%) and Span 85 (0.5%). This emulsion has been clearly shown to induce inflammation, given that it may induce secretion of cytokines such as IL-6 and IL-8, as well as the chemokines CCL2, CCL3 and CCL4 (Ko and Kang, 2018).

Here, we hypothesize that bystander activation of reminiscent self-reactive clones may have been reactivated in our patients. Considering that they were previously in a state of NEDA for long periods, it is plausible to think that self-reacting clones were well controlled either through regulatory mechanisms or through the therapies used, such as dimethyl fumarate, glatiramer acetate or teriflunomide. The inflammation caused by the adjuvant, MF-59, may have interrupted this control, thus favoring clonal expansion and disease relapse. The relatively short period after vaccination, i.e. 7–15 days, may be strong evidence for this.

Recent reports have also shown that COVID-19 severity correlates with a broad spectrum of self-reacting antibodies, including against central nervous system antigens. Unexpectedly, these responses were found to be compartmentalized to the CNS. This was further confirmed in a murine experimental model. The candidate antigens are NRG3, SYNJ2 and DPYSL2 (Song et al., 2021). Because no reports have so far demonstrated epitope similarities between SARS-CoV-2 and neuro-antigens, molecular mimicry seems not to be the case, and the mechanisms still need to be unraveled. Moreover, AZD1222 is composed of SPIKE alone, and not a fully viral particle.

Another interesting report using an experimental model has shown that the SARS-CoV-2 SPIKE protein is able to cross the blood-brain barrier by means of transcytosis via ACE-2 interaction and, furthermore, can reach the brain parenchyma, via the olfactory bulb. The AstraZeneca AZD1222 vaccine is composed of a non-replicating adenovirus with DNA sequences of the SPIKE protein, which is similar to other vaccine platforms against COVID-19. It seems less likely that this would occur after vaccination, although it remains a possibility.

AstraZeneca AZD1222 has been successfully used worldwide for controlling the COVID-19 pandemic, although there have been a few cases of immune thrombotic thrombocytopenia associated with its use. These cases were mediated by platelet-activating antibodies against platelet factor 4, thus mimicking autoimmune heparin-induced thrombocytopenia (Greinacher et al., 2021).

To the best of our knowledge, there have been no previous reports on the onset of MS and NMOSD relapses in close temporal association with AZD1222 vaccination in patients whose disease was otherwise well controlled. The very low number of cases reported here would qualify this adverse event, if indeed it is an adverse event, as extremely rare. Pfizer-BioNTech COVID-19 has been reported not to increase relapse rate in MS in two different studies involving large number of patients from Israel (Achiron et al., 2021; Lotan et al., 2021). Likewise, no increase in relapse rate was observed in NMOSD in a study involving patients from several countries who had received four different COVID-19 vaccines (Lotan et al., 2021).

Our motivation in reporting these cases was that there may be other such cases around the world. Moreover, there are no recommendations regarding the merits of the second dose of vaccine for these patients. For the moment, since the second dose of AZD1222 leads to a relatively small increase in protection against COVID-19, we have suggested to these patients that they should not take it.

Our aim in presenting this report was to bring this subject to discussion, especially with regard to cases of vaccination of autoimmune patients. We strongly support vaccination, and we believe in its importance for controlling the pandemic. We still believe that its benefits greatly outweigh the risks, given that the prevalence of adverse effects is very low. However, we also believe not only that this subject is relevant, but also that it must be brought to discussion between patients and physicians, in order to have a more transparent and trustful relationship.

Funding

This research did not receive any grant from funding agencies in the public, commercial or not-for-profit sectors.

CRediT authorship contribution statement

Yara D Fragoso: Visualization, Writing – review & editing, Writing – original draft, Supervision. Sidney Gomes: Writing – review & editing. Marcus Vinicius M Gonçalves: Writing – review & editing. Euldes Mendes Junior: Writing – review & editing. Bianca Etelvina S de Oliveira: Writing – review & editing. Cristiane Franklin Rocha: Writing – review & editing. Gutemberg A Cruz dos Santos: Writing – review & editing. Carlos Bernardo Tauil: Writing – review & editing. Raquel Vassao Araujo: Writing – review & editing. Jean Pierre S Peron: Writing – review & editing, Writing – original draft.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. No private or public funding was provided for this study.