Graves’ Disease before and after the Severe Acute Respiratory Syndrome Coronavirus 2 Pandemic: Experience of a Clinical Center Open Access

Graves’ disease (GD) is an autoimmune disease caused by autoantibodies against the thyroid-stimulating hormone receptor (TRABs) and is the most common cause of hyperthyroidism. Although evidence shows that genetics is responsible for approximately 80% of patients’ susceptibility to the disease, its development is related to an interplay between genetic and nongenetic factors, including viral infections [1‒6]. When hyperthyroidism is confirmed, measurement of serum levels of TRABs is the most reliable tool to confirm GD, although its measurement can be negative in mild forms of the disease. Increased radioactive iodine uptake on thyroid scintigraphy and/or a hypervascular and hypoechoic enlarged gland on ultrasound may also support this diagnosis [1, 3‒5, 7]. The main goal of GD treatment is to restore normal thyroid function, which can be achieved with medical therapy or with definitive treatment with radioactive iodine (I131) ablation or surgery [1‒4, 7].

Coronavirus disease 2019 (COVID-19) is an infectious disease caused by a newly identified coronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). On March 2, 2020, the first case of COVID-19 was officially diagnosed in Portugal, and 9 days later, a pandemic was declared by the World Health Organization (WHO) [8‒11]. Since the pandemic, various studies have shown an increased risk of thyroid dysfunction induced by SARS-CoV-2 infection and many cases of new onset GD have been reported. Some studies also show a bidirectional relationship between GD and COVID-19 [6, 12‒15]. This relationship may be explained by the pathophysiology of SARS-CoV-2 infection, since angiotensin-converting enzyme 2 (ACE2) is important in the internalization of SARS-CoV-2 into host cells and thyroid gland shows one of the highest levels of ACE2 expressions, making it a susceptible target to SARS-CoV-2 injury. This direct injury to the thyroid cells and the cytokine storm associated with the infection may induce dysfunction to the thyroid gland [11, 15, 16]. The relationship between COVID-19 vaccination and the development of GD has also been under investigation, with some studies reporting cases of GD occurring after vaccine inoculation. However, the connection between these events remains to be clarified [15, 17‒21].

Our main objectives were to determine whether the emergence of the SARS-CoV-2 pandemic was associated with an increase in the number of new GD cases in our center and to compare the characteristics of patients with GD diagnosed before and after the start of the SARS-CoV-2 pandemic. In patients with GD diagnosed after the start of SARS-CoV-2 pandemic, we aimed to compare GD characteristics between those who had SARS-CoV-2 infection and/or vaccine prior to GD and those in whom GD had no reported relation to SARS-CoV-2 infection and/or vaccination.

Observational and retrospective study conducted at the Endocrinology Department of Braga Local Health Unit. Adult patients newly diagnosed with GD, who attended their first consultation at the Endocrinology Department of Braga Local Health Unit between January 1, 2018, and December 31, 2022, were eligible to participate. We considered the presence of positive TRABs and/or a compatible scintigraphy as a diagnostic criterion for GD. Exclusion criteria included age under 18 years old, uncertain etiology of hyperthyroidism/thyrotoxicosis, relapsed cases of GD, pregnancy, and patients whose medical records were mostly incomplete. Patients with newly diagnosed GD were divided into two groups: group 1 which included patients whose GD symptoms started until December 31, 2019 (GD before the pandemic [GDbP]), and group 2 in which GD symptoms started between June 1, 2020, and December 31, 2022 (GD after the pandemic [GDaP]). Although the first reported case of SARS-CoV-2 infection in Portugal was on March 2, 2020, there is still some uncertainty about the exact date in which the pandemic began. Therefore, we assumed an empirical suspend time window of 5 months from January to May 2020, in which no patients were included. Group 2 was further subdivided into four subgroups: subgroup 2.1 – patients without any reported relation to SARS-CoV-2 infection or vaccination (control group [GDnIV]); subgroup 2.2 – GD symptoms following a documented SARS-CoV-2 infection (GDwI); subgroup 2.3 – GD symptoms following any SARS-CoV-2 vaccination (GDwV); subgroup 2.4 – GD symptoms following both a documented SARS-CoV-2 infection and vaccination (GDwIV).

For each patient, we collected the following data: demographic information (sex and age at diagnosis); history of any previously diagnosed autoimmune condition; laboratory measurements expressed in multiples of the upper limit of normal of the respective laboratory reference: free tri-iodothyronine, free thyroxine (FT4), TRABs, anti-thyroglobulin (anti-Tg), and antithyroid peroxidase antibodies; maximum standardized uptake value in thyroid scintigraphy. We opted to use multiples of the upper limit of normal in relation to the analytical parameters, since the data come from different laboratories, with different units and normal ranges. Other data collected include maximum dosage of antithyroid drugs (ATD) with equivalence to methimazole dosage (the equivalent potence of methimazole to propylthiouracil was considered to be approximately 1:20 [22]), number of weeks until ATD first dosage reduction and suspension, definitive treatment in the first 12 months of disease, and diagnosis of Graves’ orbitopathy.

For patients in group 2, we also collected the following data: presence of SARS-CoV-2 infection diagnosis, number of days between the onset of GD symptoms and the immediately preceding SARS-CoV-2 infection and/or the immediately preceding SARS-CoV-2 vaccination, and also the vaccine brand. The data were collected by consultation of the patient’s electronic medical records.

Data were analyzed using the Statistical Package for the Social Sciences (IBM SPSS^®) 29.0 version, and a p value < 0.05 was considered as statistically significant. For continuous quantitative variables, the presence of a normal distribution was assessed through histogram analysis and the Shapiro-Wilk test and confirmed by evaluating skewness and kurtosis. As none of our variables presented a normal distribution, for continuous variables with a non-normal distribution, we present the median (MD), the interquartile range with 25th and 75th percentiles, and minimum and maximum values. For categorical variables, the absolute number and percentage are presented. To compare differences in continuous variables between independent groups with non-normal distribution, we used the Mann-Whitney U test. To compare differences between independent groups in categorical variables, the chi-square test (χ²) and Fisher’s test (φ) were used.

In the analyzed period of time, there were 6,556 first consultations in our department (3,777 in the pre-pandemic period and 2,779 in the post-pandemic period) and 296 patients were diagnosed with GD. Forty-six patients were excluded from the study: 37 due to initial symptoms within the 5-month window period and 9 due to pregnancy, leaving a total of 250 included patients (N = 108 [43.2%] in group 1, GDbP; N = 142 [56.8%] in group 2, GDaP). The number of certain variables may vary from the sample size N due to missing data.

Table 1 describes and compares the data from groups GDbP and GDaP and subgroups GDnIV, GDwI, GDwV, GdwIV (2.1–2.4). In our department, there was a significant increase of new cases of GD in the post-pandemic period (p < 0.001).

When we compared the groups GDbP and GDaP (Table 1), there were no significant differences regarding patients’ personal history, GD’s diagnosis, and the presence of thyroid autoantibodies, including TRABs. Most patients were treated with methimazole; and despite similar methimazole dosages, GDbP group had a longer period before starting drug therapy, took longer to reduce ATD dosage, and needed more time to suspend it.

No differences were observed between the GDbP and GDaP groups in terms of the incidence and severity of Graves’ orbitopathy. Among patients diagnosed with orbitopathy, 22.2% of the GDbP group and 29.4% of the GDaP group presented with active orbitopathy.

When comparing group GDbP and subgroups GDnIV, GDwI, GDwV, and GdwIV (Table 1), there were no major differences between groups regarding sociodemographic characteristics and autoimmune disease history. When compared to pre-pandemic GD patients, anti-Tg values were significantly higher in subgroup GDwI when compared to group GDbP. Group GDbP took longer to initiate and to suspend drug therapy, when compared to groups GDnIV, GDwV, and GdwIV. Group GDbP also took longer to reduce ATD dosage when compared with group GDwV.

Lastly, when we compared the subgroup GdnIV with the subgroups GDwI, GDwV, and GdwIV, we found a statistical difference in FT4 values between GDnIV and GDwI and between GDnIV and GDwV. No other statistical differences were observed in analytical parameters. Regarding the duration of treatment, when compared to the control subgroup, patients of group GDwV required a lower period of time to reduce dosage, with no other significant different values. In subgroups GDwI and GDwV, the symptoms developed 90 days after COVID-19 infection (MD: 90; P25 45, P75 147) and 105 days after vaccination (MD: 105; P25 31, P75 183), respectively. On GDwV, the majority of patients received Comirnaty (Pfizer) vaccine (n = 34, 65.4%), followed by Spikevax (Moderna) (n = 8, 15.4%) and Vaxzevria (AstraZeneca) (n = 6, 11.5%), while a fraction of patients received Jcovden (Janssen) (n = 4, 7.7%).

We have data on remission or relapse in 55 patients (50.9%) in the GDbP group and 70 patients (49.3%) in the GDaP group. The remaining patients either received definitive treatment during the initial disease episode, continued with treatment, or were lost to follow-up. The relapse rate was 40.0% and 27.1% in GDbP and GDaP, respectively (p > 0.05).

In our center, in the post-pandemic period, the number of new GD cases approximately doubled, when compared with the immediate pre-pandemic period. Various triggers may precipitate the onset of an autoimmune disease such as GD, including viral infections and emotional stress. Evidence, mainly from case reports and case reviews, suggest a potential relationship between SARS-CoV-2 infection or vaccination and GD [12, 23‒25]. Although the pathophysiologic mechanisms behind it have not been well established, a possible explanation involves the ACE2 receptors, which serve as the entry point for SARS-CoV-2 into our cells. The high expression of these receptors in the thyroid gland makes it a target for the virus [11, 16, 26]. On the other hand, the SARS-CoV-2 pandemic combined a higher incidence of viral infections with elevated levels of anxiety due to the infection itself and the significant consequences of the pandemic in people's lives, such as lockdowns. This created the perfect storm for the emergence of new cases of thyroid dysfunction [6, 11, 16].

We found a limited number of papers analyzing the incidence of new GD cases with the SARS-CoV-2 pandemic. In accordance with our study, Donner et al. [27] also detected an increased incidence of new GD cases in the post-pandemic era in pediatric patients. Barajas Galindo et al. [28] also showed an increased incidence within the adult population, when comparing the years 2017–2018 with 2020, showing an even higher increase when comparing with 2021, which may indicate that large-scale vaccination against COVID-19, which started in the beginning of that same year, also had some impact. Endo et al. [29] focused only on new DG cases with a history of previous SARS-CoV-2 immunization; however, an increased incidence was not found. The most popular hypothesis connecting immunization against COVID-19 and GD is related to the vaccine adjuvants, an entity called ASIA (Autoimmune syndrome induced by adjuvants) [19, 26]. Vaccine adjuvants are used to enhance our immunological response to the inoculated antigen and are usually well tolerated by the majority; however, in predispose individuals, they might induce autoimmune responses [30].

When we compared groups GDbP and GDaP regarding the analytical parameters, such as free tri-iodothyronine, FT4, and maximum standardized uptake value, significant differences were not found. In contrast with our research, some studies showed increased FT4 titers in post-pandemic patients [28]. In relation to treatment duration, our study showed that post-pandemic patients took less time to initiate, taper and discontinue drug therapy, when compared to group GDbP. Although there are no studies directly comparing these aspects, these treatment duration discrepancies may be attributed to various possibilities. On one hand, preventive measures were advised, commending the continuation of ATD therapy for previously diagnosed GD patients during the pandemic, to mitigate the risk of complications and relapses, given the limited access to healthcare [31]. On the other hand, COVID-19-related GD patients may have been more symptomatic, prompting earlier healthcare access and thus initiation of ATD therapy.

In our research, subgroup GDwI presented a significant higher MD value of anti-Tg. Few studies have provided data on autoantibodies, and among those that did, they primarily focused on comparing TRABs, revealing little consensus, as some Increased autoantibodies may result from potential immunological effects of SARS-CoV-2 infection. Mohammadi et al. [32] suggests that inflammatory mediators produced in response to the virus lead to activated T helper 2 cells that promote autoantibody production against thyroid-stimulating hormone receptors, thyroglobulin, and thyroid peroxidase. Several studies have also reported similarity between the principal proteins of SARS-CoV-2, such as the spike protein, the nucleoprotein, and the membrane protein, with thyroid peroxidase sequences, which may lead to cross-reaction [19, 33].

In alignment with the previous analysis regarding treatment duration, all subgroups (except the GDwI subgroup) took notably less time for initiating and discontinuing ATD therapy when compared with group GDbP. This may be explained by the possibility that GD patients related to COVID-19 (infection or vaccination) could have been more symptomatic than GDbP patients due to the presence of higher levels of FT4 (as suggested in other studies [28]), resulting from a more severe thyrotoxicosis, likely leading to earlier healthcare access, resulting in a prompter initiation, reduction, and suspension of ATD therapy. As the GDnIV subgroup also took less time to initiating and discontinuing therapy, this difference may simply be due to greater demand for healthcare and greater patient adherence to treatment. While almost all subgroups required less time to initiate and discontinue ATD therapy, only the GDwV subgroup needed significantly less time to reduce the ATD dosage, taking approximately half the time compared to the GDbP group. This may suggest a potential advantage among GD patients whose condition was triggered by prior immunization against COVID-19.

Finally, when we compared subgroup GDnIV with subgroups GDwI, GDwV, and GdwIV, we found a significant higher level of FT4 in GDwI and GDwV, when compared to GDnIV patients. Similar results were found in studies that compared pre-pandemic GD patients with GD patients with a history of COVID-19 infection or vaccination [26, 28, 29, 34]. Although we did not find statistically significant differences in the relapse rate between GDbP and GDaP groups, GDbP’ patients have a longer follow-up period, which may bias the results.

Our findings suggest that COVID-19 infection-related GD and SARS-CoV-2 vaccine-related cases may follow a different pathophysiological pathway compared to classic GD. Further research is required to validate these findings and elucidate the underlying mechanisms driving these differences.

Our study presents several strengths. To our knowledge, this is the first study to investigate the relationship between the SARS-CoV-2 pandemic and GD in Portugal. Our research presents a comparison of multiple variables, including sociodemographic data, analytical parameters, and treatment type and duration, distinguishing it from analyses conducted so far. Another strength of our study is the fact that all COVID-19 infections are documented in the National System of Epidemiologic Vigilance (SINAVE) and consequently laboratory tested, minimizing the miss-allocation of patients. We avoided classification bias by excluding patients whose GD symptoms started within the 5-month period between January and May 2020, ensuring a clearer separation between the GDbP and GDaP groups. Additionally, the absence of a temporal criterion for classification of GD patients as triggered by SARS-CoV-2 infection or vaccination constitutes a significant advantage in our study. Currently, there is no globally accepted temporal criterion for such assignments [29], so this flexibility allowed us to avoid patients’ allocation bias to our subgroups and include a larger number of patients, resulting in a more substantial and meaningful sample size. However, we are aware that this might also translate into a possible limitation, as some patients, despite having a positive history for COVID-19 infection or vaccination, might not have exhibited a correlation due to the significant time interval between the COVID-19 event and the onset of GD symptoms.

Our study also presents some limitations to consider. Firstly, as this is a retrospective study, the authors are limited to the existing data in medical records, which limits the data we could collect. For example, we did not have data available on patients' smoking habits or severity of symptoms. We also cannot conclude whether access to consultations due to GD at our department was faster in the post-pandemic period compared to the pre-pandemic period since the number of doctors and first consultations varied significantly in the 2 periods analyzed. We cannot exclude that there were differences in access to subsequent medical appointments between pre- and post-pandemic periods and between different doctors, which may have influenced the follow-up of patients and the adjustments in medical therapy. Other limitations of our research are a possible classification bias in patients without SARS-CoV-2 infection confirmation in GDnIV as COVID-19 can be an asymptomatic disease and the fact that our sample size does not allow us to draw some important conclusions. These constraints highlight areas deserving attention and potential improvement in future studies, aiming for a more comprehensive understanding of the relationship between the SARS-CoV-2 pandemic and GD.

This study can serve as a pilot investigation that may raise questions about possible pathophysiological and behavioral patterns in GD post-COVID-19 infection or vaccination, thereby suggesting the need for additional research. A larger patient sample could potentially reveal more prominent differences between the groups.

In conclusion, in our center, the SARS-CoV-2 pandemic was associated with a significant increase in new cases of GD. In the post-pandemic period, possibly given the immunological effects of SARS-CoV-2 infection and/or vaccine, patients with GD had higher FT4 titers and anti-Tg compared to those without a COVID-19-related history, as well as a shorter period until initiation of medical therapy. However, they took less time to reduce and suspend ATD, which may indicate an initially more symptomatic disease but a shorter duration in long term, possibly explained by different pathophysiologic path comparing to classical GD. It is important to continue investigations to clarify the impact of SAR-CoV-2 infection and vaccination on thyroid disorders such as GD.

This study protocol was reviewed and approved by the Ethics Committee of the Braga Local Health Unit (Approval No. 169_2023) and by the Ethics Committee of Research in Life and Health Sciences (Approval No. CEICVS 137_2023). Written informed consent was obtained for all participants, through an online questionnaire sent by email.

Maria Joana Santos was a member of the journal’s Editorial Board at the time of submission. Juliana Marques Sá and Ana Sofia Correia have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

J.M.S. and M.J.S. conceived and designed the study. J.M.S. and A.S.C. obtained the informed consent, collected data, and performed statistical analysis. M.J.S. supervised all these steps. All authors discussed the results and contributed to the final manuscript, whose final version was reviewed by M.J.S., and J.M.S. and A.S.C. contributed equally.

Juliana Marques Sá and Ana Sofia Correia have first authorship.

The data collected in the preparation of this work are not available for sharing, as the local Ethics Committee requires their destruction after the article is published. Further inquiries can be directed to the corresponding author.