Pharmacovigilance signals from active surveillance of mRNA platform vaccines (tozinameran and elasomeran)

Abstract

Introduction

Two severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) messenger RNA (mRNA) vaccines, tozinameran/BNT162b2 (Comirnaty®, Pfizer-BioNTech) and elasomeran/mRNA-1273 (Spikevax®, Moderna), were approved by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) at the end of 2020, less than a year after the start of the coronavirus disease 2019 (COVID-19) pandemic. In France, the health authorities have requested an intensive vaccination campaign, accompanied by a reinforced and active pharmacovigilance surveillance. This surveillance and analysis of real-life data, based on spontaneous reports received by the French Network of Regional PharmacoVigilance Centers (RFCRPV), has enabled to identify numerous pharmacovigilance signals. Some of them, such as myocarditis and heavy menstrual bleeding, have been confirmed as adverse effects of these vaccines.

Method

We propose a descriptive review of the main pharmacovigilance signals identified by the RFCRPV concerning vaccines from the mRNA platform.

Results

Most pharmacovigilance signals were common to both mRNA vaccines: myocarditis, menstrual disorders, acquired haemophilia, Parsonage-Turner syndrome, rhizomelic pseudo-polyarthritis and hearing disorders. Other signals were more specific, such as arterial hypertension with tozinameran or delayed reaction site injection with elasomeran.

Conclusion

This non-exhaustive review illustrates the experience of RFCRPV in identifying and monitoring pharmacovigilance signals related to mRNA vaccines in France during the COVID-19 pandemics, and the crucial role of pharmacological and clinical expertise in this area. It also highlights the predominant contribution of spontaneous reporting in the generation of pharmacovigilance signals, particularly for serious and rare adverse events not detected before marketing.

Abbreviations

ACR

American College of Rheumatology

APTT

activated partial thromboplastin time

ASIA

autoimmune/autoinflammatory syndrome induced by adjuvants

CDC

Centers for Disease Control and Prevention

COVID-19

coronavirus disease 2019

EMA

European Medicines Agency

ESC

European Society of Cardiology

ESH

European Society of Hypertension

EULAR

European Alliance of Associations for Rheumatology

FDA

Food and Drug Administration

PRAC

Pharmacovigilance Risk Assessment Committee

PRP

pseudo-rhizomelic polyarthritis

RFCRPV

French Network of Regional Pharmacovigilance Centers

ROR

reporting odds ratio

SMPC

summary of product characteristic

VAERS

Vaccine Adverse Event Reporting System

WHO

World Health Organization

Two messenger RNA (mRNA) vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), tozinameran/BNT162b2 (Comirnaty®, Pfizer-BioNTech) and elasomeran/mRNA-1273 (Spikevax®, Moderna), were approved by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) at the end of 2020, less than a year after the start of the coronavirus disease 2019 (COVID-19) pandemic. In France, the health authorities have requested an intensive vaccination campaign accompanied by a reinforced and active pharmacovigilance surveillance. This monitoring and analysis of real-life data, based on spontaneous reports received by the French Network of Regional PharmacoVigilance Centers (RFCRPV), has allowed the identification of numerous pharmacovigilance signals. Some of them, such as myocarditis (detailed elsewhere) and heavy menstrual bleeding, have been confirmed as adverse effects of these vaccines. Other potential signals identified by the RFCRPV have also made it possible to inform healthcare professionals in their choice to pursuit or not mRNA vaccination schedule.

Here we propose a review of the main pharmacovigilance signals identified by the RFCRPV [1] concerning vaccines from the mRNA platform.

Common signals for the mRNA vaccines tozinameran (Comirnaty®) and elasomeran (Spikevax®)

Menstrual disorders

In the broadest sense, menstrual disorders include several symptoms, mainly metrorrhagia, menorrhagia (or a combination of both), amenorrhea, and cycle abnormalities (either lengthened or shortened). An exploratory assessment is only necessary in case of persistent disorders or when associated with a significant blood loss or even a general physical deterioration.

In the early summer of 2021, the first spontaneous reports of bleeding were received, in the context of a strong activity on social networks following the publications of Clancy and Merchant [2], [3]. The arguments resulting from the analysis of these first cases of women's spontaneous reports were insufficient to establish a causal link. Causality assessment is indeed complex for an event that approximately 30% of women will spontaneously present during their life [4], all the more that several factors are also impacting the regulation of the menstrual cycle, particularly stress and fatigue. Recent publications have also shown the role of inflammation in this area [5], [6]. However, the reported cases of positive reintroduction and abnormal bleeding in postmenopausal women had to be taken into account [7]. Therefore, from July 2021, the pharmacovigilance centers in charge of the surveillance of these vaccines have proposed to monitor specifically these events. During the summer of 2021, the number of reports for these events increased steadily for the two mRNA vaccines and led to a first detailed expertise in the reports presented in September 2021. In particular, the PV experts proposed the consultation of the French college of gynecologists-obstetricians, the drafting of recommendations, and the realization of quantitative and qualitative additional analyses, complemented by a literature review, to be presented in December 2021. Several aspects were investigated after this period, concerning menstrual disorders in the under-18 population, gynaecological bleeding in postmenopausal women for tozinameran, and then endometriosis for elasomeran. In 2021, the EMA considered that the evidence gathered was not yet sufficiently convincing and did not validate the signal. The French pharmacovigilance centres continued their reinforced surveillance of these events, as illustrated by a new quantitative report in April 2022, followed by a work carried out in the summer of 2022 after an information campaign encouraging the reporting of serious cases [8]. These analyses of more than 15,000 reports of menstrual disorders have allowed several re-examinations of the potential signal by the EMA. After a new lack of conclusion from the EMA in June 2022, the signal of menstrual disorders associated with vaccination with tozinameran and elasomeran vaccines was finally validated by the EMA in October 2022, in front of all the French data analyzed since 2021 and with the favourable vote of many other European countries. The pathophysiological mechanism of these disorders is not yet established.

Acquired haemophilia

Acquired haemophilia is a rare condition (estimated prevalence of 1–1.5 cases per 1,000,000 adults-years), with a predominance in subjects over 60. This isolated, acquired factor VIII deficiency is related to the presence of inhibitory autoantibodies directed against this coagulation factor. Acquired haemophilia is quite different from haemophilia, a constitutional deficiency of genetic origin. In two third of the cases, the search for an aetiological and progressive pathological context (lupus, rheumatoid arthritis, cancer) proves negative [9]. Given the population concerned, the presence of an antithrombotic treatment such as vitamin K antagonists or platelet inhibitors (aspirin, clopidogrel) may delay the diagnosis because bleeding complications tend to be falsely attributed to antithrombotic drugs. The diagnosis is most often evoked clinically, but can only be confirmed by an isolated prolongation of the activated partial thromboplastin time (APTT) with an isolated factor VIII deficiency and the identification of anti-factor VIII antibodies.

In France, since the beginning of the vaccination campaign and until February 10, 2022, 18 cases of acquired haemophilia have been recorded with tozinameran, occurring in 10 women and 8 men, with an average age of 75 years. In five cases, it occurred after the first dose (D1) (three men and two women between the ages of 25 and 90, with onset times of 4, 10, 12, 32 and 90 days) and in 12 cases, after the second dose (D2) (four men and eight women between the ages of 45 and 90, with onset times of 2, 13, 15 [2 cases], 17, 21, 22, 23, 30, 49, 57 and 60 days). Finally, in one case, it occurred two months after the first booster dose (R1) in a 70-year-old man. All cases required hospitalization with emergency management. In three cases, the evolution was fatal, corresponding to a mortality rate of 16.6%, consistent with that reported in the literature (8–20%) [10].

For elasomeran, five cases of acquired haemophilia were identified (end of August 2022), one after D1 (male, 90 years, 2 days to onset), three after D2 (female, 88 years, 57 days to onset; female, 53 years, 2 days to onset; male, 76 years, 67 days to onset) and one after R1 (male, 84 years, 10 days to onset), with an average time to onset of 30 days. These cases occurred in two women and three men (median age: 76 years). Three cases had a favourable outcome or were in the recovery process and two cases are still unresolved [11].

Among the 23 reported cases, there were three patients with a history of immunity (two with rheumatoid arthritis, one with lupus erythematosus), two with lymphopathy or monoclonal gammopathy of undetermined significance, and one patient with a relapse of a previously known acquired haemophilia. For the 17 other patients, the aetiological workup (autoimmune, infectious, neoplastic) remained negative. This type of event has also been reported with other vaccines [12], in particular with the influenza vaccine [13] and the H1N1 vaccine [14]. In addition, several cases of acquired haemophilia in patients vaccinated with tozinameran or elasomeran have also been reported in the literature [15], [16], [17], [18], [19], [20], [21], [22].

A search of the World Health Organisation (WHO) global database, Vigibase® (May 30, 2022) retrieved 120 cases of “acquired haemophilia” with tozinameran and 37 cases with elasomeran (mainly from the United States of America, France and Italy). From a mechanistic point of view, it was interesting to note that cases of acquired haemophilia have also been reported after COVID-19 infection [23], [24]. In the same way, an experimental work was recently implemented to evaluate whether the antibody response induced by the vaccine against the SARS-CoV-2 spike protein could have factor VIII inhibitory functions [25].

In conclusion, the role of the vaccine in the occurrence of acquired haemophilia cannot be excluded. The number of cases reported in France, the cases retrieved in Vigibase® and the scientific articles published are in favour of a potential signal. Thus, even if it is a rare event, it may be recommended to look for an acquired haemophilia in case of any sudden onset or bleeding of unusual intensity.

Parsonage-Turner syndrome

Parsonage-Turner syndrome, also called amyotrophic neuralgia of the shoulder, is a rare condition (more frequent in men) whose etiopathogeny is poorly understood.

The dysimmune hypothesis remains plausible in the presence of infectious or vaccine triggers, and cases of Parsonage-Turner syndrome have been reported after infection with SARS-CoV-2. Post-vaccination forms account for 10 to 15% of the cases [26]. Since the beginning of the vaccination campaign in France and until February 10, 2022, 59 cases of Parsonage-Turner syndrome (43 cases for tozinameran and 16 cases for elasomeran) have been reported [27], [28]. The cases were reviewed and analyzed with an expert pharmacologist and neurologist commissioned to assess the neurological events from COVID-19 vaccines monitoring. The diagnosis was confirmed in 30 cases, based on the clinical picture and additional examinations, and men were mainly concerned (n  = 20, 67%). The diagnosis was not confirmed in 29 cases because of incompatible delay in onset or insufficient evidence to confirm the diagnosis. Among confirmed cases of Parsonage-Turner syndrome, median age was 51 years (range: 25–75 years) for tozinameran, with a greater prevalence between 30 and 49 years. For elasomeran, median age was 33 years and cases were more frequent after 65 years (36%). The evolution was favourable in 10 cases (36%). Eight cases occurred after D1 (28%), 16 cases after D2 (6%) and 6 cases after R1 (16%). Finally, in eight cases, particular forms were noted, both in their clinical expression (a focal form, an atypical form, an early form) and in their context of occurrence. Indeed, we observed:

• •a relapse in a patient with a history of Parsonage-Turner from which she had completely recovered;
• •one case of contralateral Parsonage-Turner syndrome of the vaccinated arm;
• •one case of post-partum Parsonage-Turner syndrome (a risk period for this syndrome);
• •three cases occurring in the context of a trauma or intense work (ski fall, gardening effort, ceiling plasterboard installation).

The literature contains several case reports of Parsonage-Turner syndrome after vaccination with tozinameran and elasomeran. Mahajan et al. reported the case of a man presenting sudden severe painful symptoms at D7, secondary followed by peripheral muscle weakness. The electromyography (EMG) confirmed the diagnosis of Parsonage-Turner syndrome, and the clinical picture improved with corticosteroid therapy [29]. Queler et al. reported the occurrence of Parsonage-Turner syndrome with painful symptoms in the left forearm 13 hours after D1 performed in the right upper limb. A recent history of Lyme disease (which may have caused amyotrophic neuralgia) treated with doxycycline was retrieved in the previous two months [30]. Diaz-Segarra et al. reported a non-algesic amyotrophic neuralgia 9 days after vaccine administration, which was clearly improved by corticosteroid therapy [31]. Coffman et al. reported an amyotrophic neuralgia 15 days after D2, with a favourable evolution in 3 months [32]. Shields et al. reported four cases of amyotrophic neuralgia with an onset time of 7 days (one case), 15 days (two cases) and 2 months (one case), respectively, confirmed secondarily by EMG [33]. Koh et al. reported two cases of Parsonage-Turner syndrome (one case 25 days after D1, with a favourable evolution under corticosteroid therapy; one case 4 days after D2, with a spontaneous favourable evolution) [34].

In conclusion, despite the low number of Parsonage-Turner syndrome cases reported in France, these data clearly suggest that the role of the vaccine cannot be excluded. Furthermore, the occurrence of similar cases with other COVID-19 vaccines and the reports from the literature are also supporting a potential signal.

Pseudo-rhizomelic polyarthritis (PRP)

PRP, characterized by pain and inflammatory stiffness of the scapular and/or pelvic girdle, is a significant clinical and biological inflammatory syndrome occurring in people over 50. It is a benign autonomous entity, but can also result from various pathologies, sometimes serious and requiring urgent care, such as Horton's disease, which is frequently associated with PRP. PRP is exceptional before 50. The incidence (20 to 100 cases per 100,000 person-years) increases with age, with a prevalence of approximately 700 per 100,000 person-years over 50. PRP is more common in women than in men, but this difference decreases with age. It is more common in Caucasians. In addition to the involvement of the girdles (rhizomes), peripheral manifestations are frequent, sometimes edematous, which can mimic other inflammatory rheumatisms. Corticosteroid therapy with 15 to 25 mg of prednisone equivalent is usually spectacularly effective.

Since the beginning of the vaccination campaign and until November 11, 2021, 70 cases of PRP for tozinameran and 12 cases for elasomeran have been reported [35], [36]. The median age was 72 years (range: 50–87 years) for tozinameran and 76 years for elasomeran. Cases occurred similarly in both genders, with 51.5% of women. In 37 cases, the symptoms occurred after D1 (mean time (SD): 10.9 (9.7) days for tozinameran and 6.8 (6.0) days for elasomeran) and in 44 cases after D2 (mean time (SD): 24.6 (30.2) days for tozinameran and 43.8 (45.4) days for elasomeran). One patient had a history of rheumatoid arthritis with a flare of PRP after vaccination. Three patients had a pre-existing PRP flare. Nine patients had concomitant Horton's disease. Two other patients presented an anicteric cholestasis, a classic sign of PRP. Most observations met diagnostic criteria (34 cases meeting European Alliance of Associations for Rheumatology [EULAR]/American College of Rheumatology [ACR] criteria) or benefited from specialized advice or follow-up. The response to corticosteroid therapy was almost constant when prescribed early at the proper dosage.

In the scientific literature, few case reports have been published [37], [38]. Interestingly, an analysis performed on Vigibase® found 290 cases of PRP and 9 cases of PRP associated with Horton's disease [39]. A disproportionality analysis showed that COVID-19 vaccines (mainly mRNA) were associated with an increase in PRP cases (reporting odds ratio [ROR]: 2.3, 95% CI: 2.0–2.6). However, a comparison with the influenza vaccine did not find an increase in these cases with COVID-19 vaccines (ROR: 0.2, 95% CI: 0.2–0.2).

In conclusion, the number of cases reported to the French pharmacovigilance system was fairly typical of post-vaccination PRP, even though the number of reports seems low as compared to the number of injections in people over 50 years (over 41.4 million). This constitutes a potential signal to be investigated, without calling into question the benefit–risk ratio of the COVID-19 vaccines. Nevertheless, the occurrence of post-vaccine PRP deserves to be recognized early, given its good response to corticosteroids. In addition, PRP may sometimes be accompanied by Horton's disease, which should be diagnosed as soon as possible given the risk of ophthalmological complications and the need to implement a corticosteroid therapy, with doses of prednisone up to 40–60 mg.

Hearing impairment [40], [41]

Hearing loss is characterized by a decrease in the ability to perceive sound. Sudden onset hearing loss is defined as the occurrence of a sensorineural hearing loss of at least 30 dB on three successive audiometric frequencies within 72 hours. In front of any sudden onset deafness, it is advisable to eliminate a retro-cochlear pathology in the cerebellopontine angle, vestibular schwannoma being the most frequent in this case. The pathophysiological mechanisms underlying the occurrence of idiopathic sensorineural hearing loss have been described as viral, vascular, pressure-induced, autoimmune, and genetic. Different degrees of hearing loss can be observed in sudden hearing loss.

In France, the degree of hearing loss is based on the classification of the International Bureau of Audiophonology, which classes hearing losses from mild to profound, according to the average unweighted hearing loss at four conversational frequencies: 0.5 kHz, 1 kHz, 2 kHz and 4 kHz. To be relevant, the hearing loss had to be greater than or equal to 30 dB Hearing Level), and should be measured in bone conduction for pure sensorineural hearing loss and in air conduction for mixed hearing loss. The deafness is characterized as mild for a loss of 21–40 dB, moderate for 41–70 dB, severe for 71–90 dB, profound beyond 90 dB and total (cophosis) for a loss of 120 dB. There are two main types of hearing loss: conductive deafness due to a disorder of the external or middle ear, and sensorineural deafness due to a cochlea or auditory nerve disorder. They may be associated with vertigo.

The incidence of sensorineural hearing loss is approximately 5–27 per 100,000 person-years in the United States. Sudden onset hearing is most often unilateral, but can be bilateral in the presence of a genetic or autoimmune context. In these cases, the bilateral hearing loss may not be simultaneous.

Since the beginning of the vaccination campaign, questions about a possible link between COVID-19 vaccination and deafness and/or tinnitus and reports were recorded by the RFCRPV, often accompanied by requests of guidance concerning further vaccination. We therefore analysed all reports of deafness recorded in the French national pharmacovigilance database from the start of vaccination until February 2, 2022. The choice to analyse cases of “deafness” as a first intention was motivated by the possibility of recovering auditory examinations and, in particular, the audiogram and/or medical report. These elements were intended to facilitate the confirmation of deafness cases and the establishment of a link with the vaccine, taking into account the multiple factors that can generate temporary deafness. We only considered sensorineural and mixed hearing loss. The various data in the literature suggested that a 21-day delay in onset would be “compatible” with an established link with the vaccine. Two medical experts reviewed all the cases. Among the 379 cases of deafness recorded over this period, we retained 171 cases, including 142 cases under tozinameran and 29 cases under elasomeran. A positive reintroduction was observed in 8 cases (5%), bilateral deafness in 26 cases (15%) and a known need for a hearing device in 17 cases (10%). There was no gender effect, and the age group 30–64 years was predominant (69% of the cases). Medical history was cited in about one third of the cases (cardiovascular and/or otoneurological, autoimmune). Deafness was more reported after D1 for elasomeran (p <0.063) and more unilateral deafness were reported with tozinameran (p <0.01). Considering the number of vaccine doses administered, the rate of deafness reports remained similar for both vaccines (1.45 per 1,000,000 person-years for tozinameran and 1.67 per 1,000,000 person-years for elasomeran).

In the literature, several case reports have described cases of deafness with COVID-19 vaccines. Jeong and Choi reported three cases of unilateral deafness, including one case after ChAdOx1-S (Vaxzevria®) and two cases after tozinameran [42]. Ekobena et al. reported four cases of transient audio-vestibular symptoms after vaccination with tozinameran or elasomeran. The onset time was quite variable, and there was one positive rechallenge and one negative rechallenge [43]. Wishova et al. published in 2021 the most extensive case series of post-vaccination hearing impairment. This was a retrospective study based on the patient consultation database of an otorhinolaryngology speciality clinic in California. They compared the rate of diagnosis of “idiopathic hearing loss” without evidence of a specific aetiology (autoimmune, infectious, etc.) over 30 days (February to March 2021) in 2019, 2020, and 2021. The incidence of deafness cases within the active threads of the clinic consultation was estimated respectively at 1.6% in 2019, 2.4% in 2020 and 3.9% in 2021. Among the 1325 deafness consultations in 2021, 30 cases occurred after a COVID-19 vaccine, in 19 men and 11 women, with a mean (SD) delay of 10.2 (9.0) days (range: 1–42 days). There were 12 cases under tozinameran, and 18 cases after elasomeran, with 1 case of positive reintroduction, 7 cases post D1, 7 patients with associated symptoms (vertigo, tinnitus, etc.). There were five cases with a history of Meniere's disease and/or autoimmune auditory pathology. The main limitations were related to the difficulty of access to consultations during the COVID-19 period in 2020–2021 and the underestimation of the number of subjects with hearing impairment [44]. Pharmaco-epidemiological data are represented by two studies carried out on American and Israeli databases. Foremeister et al. performed a study on 555 incident cases of sudden deafness reported to the Centers for Disease Control and Prevention (CDC) Vaccine Adverse Event Reporting System (VAERS) over 7 months (December 14, 2020, to July 16, 2021). This work was preceded by a preliminary analysis (December 14, 2020, to March 2, 2021) of 40 cases by the same authors. The authors estimated the event incidence from 0.6 to 28 cases per 100,000 person-years (comparable for the three vaccines), as compared to 11 to 77 per 100,000 person-years for the general population before the pandemic. A more detailed analysis of 21 cases did not enable to identify particular risks factors for the occurrence of this event [45], [46]. Yanir et al. conducted a retrospective study in 2021 on an Israeli medical health database including approximately 4.7 million subjects between December 20, 2020, and March 31, 2021. In total, 91 cases of acute deafness after tozinameran vaccination were identified within three weeks after D1, and 79 cases after D2.

The estimated incidence of deafness per 100,000 person-years was 14.49 (95% CI: 11.69–17.29) in 2018, and 15.68 (12.79–18.58) in 2019. It was 22.44 (11.77–33.11) after D1 and 22.90 (11.68–34.12) after D2. The attributable risk was estimated at 0.91 per 100,000 person-years (0.29–1.38) after D1 and 0.61 per 100,000 person-years (−0.07–1.12) after D2, as compared with the baseline incidence in 2018. The incidence was significantly higher in women > 65 years and men aged 16–44 years [47].

This analysis identified rare cases of deafness related to mRNA vaccines. This was a minimum estimate since only medically well-documented cases were retained. The estimated incidence does not call into question the benefit/risk ratio of COVID-19 vaccines in a pandemic context. Nevertheless, in the absence of any aetiology that could explain a case of sudden deafness, health professionals should be aware of a possible relation with mRNA vaccines. This knowledge may facilitate a rapid management with an appropriate treatment, which could improve the prognosis. This should also of interest to guide a collegial discussion concerning the strategy for further mRNA vaccinations, while taking account of the various risk factors of the patient.

Specific signal for tozinameran

High Blood Pressure (HBP)

The reinforced pharmacovigilance follow-up allowed the rapid identification of a signal on the occurrence of hypertension after tozinameran vaccination. An in-depth analysis of 73 cases of hypertension reported since the beginning of the vaccination campaign was carried out and published in report No. 4 listing cases up to February 4, 2021 [48]. These cases were classified as severe for 49.3%. Two primary contexts of occurrence are mainly retrieved: either post-vaccination, within the first 15 minutes after injection, or delayed, within 12 hours to 5 days after vaccination. We also noted a marked and transient increase, for a few hours or a few days, in blood pressure levels, associated with symptoms such as headache, dizziness, feeling of malaise, asthenia, or epistaxis, in patients with no history or well-balanced hypertension. These cases have required medical attention (through emergency management, hospitalization, initiation or adjustment of antihypertensive treatment). This immediate or delayed increase in blood pressure, of short duration and favourable evolution, does not call into question the benefit/risk ratio of tozinameran vaccine, but constitutes a signal to be monitored. This French signal was transmitted to the EMA on February 18, 2021, and renewed in March and April 2021. The pharmacovigilance report No. 9 (covering data until March 11, 2021) listed a total of 207 cases of hypertension and confirmed the signal already shared with the EMA [49]. Finally, the report No. 14 (until April 29, 2021) described a total of 415 cases corresponding to grade 3 hypertension in 37 cases, grade 2 in 9 cases, according to the European Society of Cardiology and the European Society of Hypertension (ESC/ESH) 2018 recommendations, with the same conclusions as the previous reports [50].

An adrenergic mechanism could explain a transient blood pressure increase during vaccine administration, as classically evoked for other vaccinations or medical procedures. However, for delayed blood pressure increases, occurring a few days after vaccination, a different mechanism should be suggested. It may be related to the strength of the immune stimulation induced by the mRNA vaccine through the cytokines, in the light of current research on the interactions between immunity, inflammation, lymphocytes and hypertension [51], [52].

Specific signals for elasomeran

Delayed injection site reaction

After the first vaccinations in the United States, the possibility of a very rare occurrence of anaphylactic shocks was identified, initially based on the hypothesis of an allergy to PEG, a component of mRNA vaccines. Strict recommendations for post-injection monitoring and collaboration with allergology departments were then implemented in France. However, delayed skin reactions occurring within ten days after the vaccination at the injection site were the most frequently reported to the French pharmacovigilance system. These data were quickly monitored and validated in April 2021. As early as March 2021, a letter to the editor in the New England Journal of Medicine warned about the occurrence of these delayed reactions in a series of 12 patients, while reassuring that the rash was not severe and resolved spontaneously [53]. Now referred to as the “COVID arm”, this rash is characterized by pain, erythema, pruritus, and induration at the injection site occurring on average seven days after injection. In the phase III clinical trial of the elasomeran vaccine, 0.8% and 0.2% of patients experienced delayed injection site reactions after the first and second dose, respectively [54]. The underlying mechanism is thought to be a T-cell-mediated hypersensitivity reaction. However, these reactions do not constitute a contraindication to a subsequent dose. At the European level, the signal was validated in April 2021 and added to the Spikevax® summary of product characteristic (SmPC) in October 2021, after validation by the Pharmacovigilance Risk Assessment Committee (PRAC).

Thyroiditis

Subacute thyroiditis was one of the events of particular interest monitored since the beginning of the vaccination campaign in France. These events included a wide range of thyroid diseases linked to inflammatory or infectious aetiologies, with very different clinical and biological evolutions. Thyroiditis is an inflammation of the thyroid gland, frequent and mostly benign [55], [56].

Several questions from clinical practitioners have led us to review the cases of thyroiditis after vaccination with elasomeran. Thus, as of January 6, 2022, we retained 29 cases of thyroid disorders out of the 18 million injected doses and the 18482 reported cases. Among these 29 cases, seven concerned the occurrence or recurrence of a cyst, mass, or goitre, one the occurrence of thyroid neck pain, one hypothyroidism without further clarification, five of Basedow's disease (three recurrences and two de novo) and 15 of thyroiditis. As of August 25, 2022, 20 additional cases of thyroid disorders have been reported, including two thyroiditis. The persons concerned were relatively young (mean age: 44 years, median: 42 years), with a majority of women (60%). Slightly more than half of them had a history of thyroid disorders, and the events after vaccination were considered as a recurrence of a pre-existing pathology, that occurred with an average delay of 7 days (range: 1–19 days). For the remaining seven cases (thyroiditis), the first symptoms also appeared rapidly, within 15 days after vaccination, but the diagnosis was made a few weeks later. In addition, two cases were diagnosed with De Quervain's thyroiditis. Of the other five cases, two had positive anti-thyroid peroxidase antibodies [57].

As hypothyroid disorders are common, these cases did not allow us to determine an excess risk with the vaccine. Nevertheless, the characteristics of twenty published cases of thyroiditis, subacute or not, reported with mRNA vaccines are consistent with those of our series [58], [59]. In the same way, the authors of these publications suggested, in addition to a potential risk of post-vaccination recurrence in patients with a history of thyroid disorders such as Hashimoto's thyroiditis or Graves’ disease, the cases of de novo thyroiditis, particularly subacute thyroiditis, which was transient and not serious in most of the cases. Two main mechanisms were proposed to explain these thyroid disorders:

• •an inflammatory/immune reaction to the vaccine excipients used to increase the response to the vaccine (autoimmune/autoinflammatory syndrome induced by adjuvants (ASIA) syndrome, already described for other vaccines (papillomavirus, hepatitis B, influenza…) [60];
• •an interaction at the thyroid level with the Spike protein due to the presence of numerous angiotensin II receptors likely to bind the protein, but also an interaction between the antibodies generated by the vaccine and the thyroid antigens, leading to transient thyroiditis by binding to receptors on the surface of the thyroid [61], [62].

Finally, COVID-19 disease may also be the cause of thyroiditis [63].

In final, there are some arguments supporting that thyroid disorders such as thyroiditis, occurring after vaccination and apart from recurrences of pre-existing disorders, can potentially be favoured by vaccination. This does not call into question the benefit–risk ratio of the vaccination. However, when faced with neck pain, swelling of the neck, or intense fatigue or palpitations, practitioners should look for a possible thyroiditis. At the time of writing, cases of thyroiditis were still under particular surveillance, both at the national and the European level.

Conclusion

This non-exhaustive review illustrates the experience of RFCRPV in identifying and monitoring pharmacovigilance signals related to mRNA vaccines in France during the COVID-19 pandemics, and the crucial role of pharmacological and clinical expertise in this area. Frequency of reports through the ANSM was adapted up to a weekly basis to provide real-time insights to healthcare professionals and to guide their strategy concerning the continuation of the vaccination schedule. This paper also highlights the predominant contribution of spontaneous reporting in the generation of pharmacovigilance signals, particularly for serious and rare adverse events not detected before marketing.

Disclosure of interest

The authors declare that they have no competing interest.