New aspects in the pathogenesis and management of subacute thyroiditis

Susceptibility to SAT has been considered to be associated with the occurrence of certain types of human leukocyte antigens (HLA) since 1975. In that year, Nyulassy et al. [1] first reported the increased frequency of HLA-B*35 in patients with SAT. Since then, the significant correlation between SAT and HLA-B*35 has been confirmed in several different populations [2‐6]. The presence of HLA-B*35 was observed in up to 70% of SAT patients [2, 7], but it was considered doubtful that HLA-B*35-negative patients had virtually no genetic background of SAT. Thus, in the meantime, some authors analyzed other potential associations between HLA and SAT, but their results were highly inconsistent and did not provide reliable conclusions [7‐9]. Those authors applied much older, especially serological methods whose accuracy is substantially lower than that of the recently used high resolution method. Recent improvement of diagnostic methods brought the novel insight into the genetic background. In 2020, our research team demonstrated the new association between HLA and SAT in the Caucasian population [10]. In addition to the previously described correlation with HLA-B*35, SAT was proved to be associated with the presence of HLA-B*18:01, -DRB1*01 and -C*04:01 [10]. In Caucasians, HLA-C*04:01 is in a linkage disequilibrium with HLA-B*35:01/02/03, i.e. these two alleles commonly occur together because of the close location of their loci [11]. Therefore, HLA-C*04:01 alone cannot be treated as an independent SAT risk factor due to its linkage with the previously described HLA-B*35. However, the presence of any of the two haplotypes should be considered as a marker of genetic susceptibility to SAT [10]. No such linkage disequilibrium has been described for HLA-B*18:01 or -DRB1*01 so these alleles should be considered new, completely independent SAT risk factors. The study provided the set of four alleles whose assessment allows confirming the genetic predisposition for SAT in almost all patients [10].

As the studies on the novel genetic background of SAT were progressing, the new data on the influence of the HLA alleles on the clinical course of SAT were emerging. Currently, several new publications on this important issue are available and are discussed below.

Previous viral infection (occurring approximately 2–6 weeks earlier) is considered a SAT triggering factor in genetically predisposed individuals. The types of viruses associated with the occurrence of SAT include Coxsackie viruses, Echo viruses, adenoviruses, influenza viruses, mumps and rubella viruses, parvovirus B19, orthomyxovirus, HIV, Epstein-Barr virus, hepatitis E and measles virus [12, 13]. A few cases of SAT associated with dengue virus infection have recently been described [14]. During the last year, we have been dealing with the SARS-CoV-2 pandemic spreading around the world. This RNA virus causes COVID-19. So far, over 100 million people have been affected around the world, and this number will be undoubtedly increasing further. Prior to the current pandemic, no association of coronaviruses with the development of SAT was reported [12]. Observations on the pre-existing SARS-CoV-1 virus indicated possible damage to thyroid follicular cells and subsequent fibrosis, which corresponded to destructive thyroiditis [15]. The development of thyroid diseases in the course of severe acute respiratory syndrome (SARS) has been associated with various mechanisms of thyroid damage, including excessive immune response, immunodeficiency associated with infection, or direct cellular damage [16]. SARS-CoV-2 exhibits significant tissue tropism, including the high affinity to the thyroid tissue. The key factor in SARS-CoV-2 infection is an angiotensin converting enzyme 2 (ACE2), which serves as a 'receptor' enabling virus to enter the cell. Thyroid cells are rich in ACE2 [17‐19]. Rotondi et al. has recently demonstrated that the mRNA encoding the ACE-2 receptor is highly expressed in thyroid follicular cells, making them a potential target for SARS-CoV-2 entry [20]. Due to the increased affinity of SARS-CoV-2 to the thyroid gland, attention was drawn to the potential risk of developing several potential forms of thyroiditis. Actually, some cases of Graves' disease (GD) [21] and Hashimoto's thyroiditis [16] have been described, but SAT appeared to be the most common COVID-19-related thyroiditis [19]. Despite the very recent onset of the COVID-19 pandemic, many cases of SAT in direct relation to SARS-CoV-2 infection have already been described. Thus, SARS-CoV-2 should be included into the list of viral triggering factors of SAT. Moreover, as the pandemic is still progressing, we should consider SARS-CoV-2 as the most important triggering factor of SAT at the present time.

Very interesting observation was published by Nakagawa et al. [22] who reported SAT in 71-year-old Japanese female with psoriatic arthritis treated initially with secukinumab, an interleukin (IL)-17A inhibitor. Due to treatment failure, the therapy was changed to adalimumab (ADA), a tumor necrosis factor (TNF)-α inhibitor, which triggered SAT occurrence. Treatment was again modified to ixekizumab, an IL-17A inhibitor, without SAT relapse [22]. This very recent case constitutes a confirmation of previous reports on SAT developing during treatment with TNF inhibitors. Those cases include patients with various rheumatic diseases treated with such TNF inhibitors as ADA, etanercept, and infliximab [23‐25]. The time lag from the beginning of the therapy to the onset of SAT varied from 5 days [22] to a few years [23, 24]. There have been no reports of SAT associated with other biologic disease-modifying anti-rheumatic drugs (DMARDs) or targeted synthetic DMARDs, such as an IL-6 receptor inhibitor (tocilizumab), CTLA-4 Ig (abatacept), IL-17 inhibitors, IL-23 inhibitors or JAK inhibitors [22]. Moreover, several previous reports demonstrated that IFNα-inhibitor therapy for hepatitis C can also induce SAT [22, 26]. The postulated mechanism has included the hypothesis that TNF inhibition leads to increased production of IFN-α by plasmacytoid dendric cells, which promotes lymphocyte migration and inflammatory reaction [27]. It was also suggested that the mechanism of SAT triggered by TNF inhibitors may be related to cytokine imbalance [22].

It has been well known for decades that the most common complaints of patients with SAT include anterior neck pain – usually radiating to the jaw, ear and upper mediastinum – as well as fever, rising especially at night. Fatigue, malaise and muscle pain can also occur. Many patients present clinical and/or biochemical manifestation of thyrotoxicosis, with usually low to moderate severity [29]. In the minority of cases, signs and symptoms of thyrotoxicosis may dominate the clinical presentation, with weight loss, tremor and palpitations [29]. The symptoms described above are considered typical for the first phase of SAT. Thyrotoxicosis phase results from the destruction of thyroid follicles and release of thyroid hormones. Typical course includes also the second phase, when pain and fever resolve and the third phase with subsequent hypothyroidism. Persistent hypothyroidism is rare [29].

The most characteristic laboratory finding is a high erythrocyte sedimentation rate (ESR), sometimes reaching even three-digit value. C reactive protein (CRP) is elevated in many cases, although it is a less typical marker. White blood count (WBC) may also be increased. Anti-thyroid antibodies are believed to be usually normal. The ultrasound (US) pattern of SAT include hypoechoic and heterogeneous areas with blurred margins, poorly vascularized on color Doppler [33].

Although the natural course of the disease is believed to be often self-limiting, patients frequently suffer from severe symptoms which persist even for months and not rarely resolve only after glucocorticoid (GC) treatment.

The clinical characteristics of the disease has been changing significantly for the recent years. More and more cases of painless SAT have been reported [29, 34, 35], with frequency reaching 6.25% in our studies published in 2019 [29], but increasing instantly during the pandemic, including virtually the majority of patients with SAT hospitalized due to COVID-19 [18, 30]. The latter observation was primarily assigned to the frequent application of analgesics and nonsteroidal anti-inflammatory drugs (NSAID) to COVID-19 patients [36]. However, further analyses revealed painless course even in patients with no such treatment [18]. Thus, the neck pain – symptom regarded previously as the key diagnostic criterion – appeared to be not always present.

Fever was also reported to occur less often than it was believed, and was frequently associated with microhaematuria [29]. However, microhaematuria was also observed in some patients with normal body temperature and it resolved spontaneously with SAT resolution [29].

As it was previously mentioned, some cases of children with SAT have been reported [31, 32]. The most spectacular case, published very recently, regarded 5-year-old boy with airway compromise caused by SAT [31]. The boy suffered from fever, painful neck swelling, dysphagia and hoarseness for a few weeks and finally he was admitted to hospital due to dyspnea. US and fine needle aspiration biopsy (FNAB) findings were typical for SAT and magnetic resonance imaging confirmed trachea compression by the enlarged thyroid. Steroid treatment led to symptom resolution [31]. Thus, one should remember about the possibility of SAT occurrence in children, as well as of the fact that symptoms of airway compression may be a manifestation of SAT.

Several years ago, the absence of thyroid antibodies was considered typical for SAT. However, elevated levels of anti-thyroid antibodies, including thyroid peroxidase antibodies (aTPO), thyroglobulin antibodies (aTg), and even thyrotropin receptor antibodies (TRAb) are more often present [29, 37]. The frequency of increased aTPO, aTg and TRAb levels has recently been reported by our research team as 15.52%, 33.33% and 6.0%, respectively [29]. Similar results were reported by Nishikara et al. who observed the presence of aTg and aTPO in 52.5% and 15.6% of samples, respectively [38]. The authors demonstrated that in some patients the aTg concentrations decreased or disappeared within 4 months to 6 years [38]. This observation confirms our previous hypothesis that in many patients the occurrence of thyroid antibodies in SAT may be considered a simple consequence of the release of thyroid antigens due to the gland damage during SAT [29]. However, it can also reflect the increasing incidence of autoimmune thyroid disease (AITD) and the phenomenon of overlapping of SAT and AITD. The crucial significance of co-occurrence of particular HLA alleles and simultaneous presence of SAT and GD has been recently reported in our studies [37].

Despite the progress in availability of diagnostic tools, the differential diagnosis between thyrotoxicosis in GD and SAT may be challenging, especially in resource-limited settings, including undeveloped regions of the world, but also recently due to the limited accessibility of diagnostic tests, caused by COVID-19 pandemic, observed even in some developed countries. Many authors analyzed the usefulness of several surrogate markers based on simple blood count analysis. The potential usefulness of such parameters as platelet-lymphocyte ratio (PLR) [39‐42], monocyte-eosinophil ratio (Mo/Eo) combined with free triiodothyronine (FT3) to free thyroxine (FT4) ratio (FT4/FT3) [43], neutrophil–lymphocyte ratio (NLR) [39, 41, 42] has been postulated. Taşkaldiran et al. observed that in Turkish patients with SAT, PLR and NLR were significantly higher than in patients with GD, thyroid adenomas or in healthy control groups [39]. Another study provided similar observation that PLR and NLR values in SAT patients are higher as compared with healthy control subjects [42]. In a large cohort from India, PLR values were also higher in SAT as compared to GD, however, they were significantly lower than in the control euthyroid group [40]. The authors concluded that PLR is a useful marker to differentiate SAT and GD and proposed a PLR cut-off value of 70.4 with demonstrated sensitivity of 86% and specificity of 74% [40]. Cengiz et al. aimed to assess the usefulness of NLR and PLR in the follow up of SAT patients [41]. They proposed cut off values of 2.4 (80% sensitivity and 51% specificity) for NLR and 146.84 (83% sensitivity and 54% specificity) for PLR for the acute phase of the disease. Taking into account the discrepancies in the obtained results, we can conclude that NRL and PLR can be useful in differential diagnosis of SAT and GD and in the follow up of SAT treatment, but currently no universal cut-off value can be unequivocally recommended. Hu et al. evaluated other potential markers and reported that Mo/Eo and FT4/FT3 ratios were significantly higher in SAT than in GD, and the cut-off values for fT4/fT3, Mo/Eo ratios and Mo/Eo ratio + fT4/fT3 for diagnosing GD were ≤ 2.841, ≤ 8.813 and > 0.644, respectively [43]. Although still no universal cut-off values of the proposed markers can be recommended for general population, such surrogate tools might be helpful in the initial differential diagnosis of SAT in resource-limited settings.

It is well known that during SAT, both thyroid lobes or only one lobe may be affected in US. In the latter case, the inflammatory infiltration often migrates from one lobe to the other. However, at the time of diagnosis, the US lesions can usually be described as unilateral or bilateral. Sencar et al. analyzed the association between the degree of thyroid involvement in US (unilateral or bilateral) and the clinical course of SAT [44]. No correlation between the initial thyroid involvement in US and either the risk of recurrence or the risk of permanent hypothyroidism were demonstrated [44]. These findings probably result from the fact that the degree of thyroid involvement in US usually changes with time, and the initial US examination may not reflect the most severe stage of the disease.

Signs and symptoms of SAT are non-characteristic. Recent changes in the clinical presentation make the diagnosis even more difficult despite the progress in diagnostic methods. The symptoms of SAT may be assigned to the infection, as neck pain (if present) may be referred to throat or lymph nodes involvement, while fever, fatigue, and increased inflammatory markers (CRP, WBC) may further confirm the false negative diagnosis.

Despite the high availability of diagnostic tools, including laboratory markers and US examination, the diagnosis of SAT is still frequently delayed. The patient often visits many physicians, starting from GPs, through laryngologists and other specialists, before being finally diagnosed with SAT. The period of this delay was reported by our research team as ranging from two weeks even to six months [59]. Due to misdiagnosis of infection, antibiotics were unnecessarily administered in nearly 50% of SAT patients. The ineffectiveness of one antibiotic resulted frequently in the use of another one, covering wider spectrum of bacteria. Characteristic features, occurring more frequently in the antibiotic treated patients included fever, preceding infection, elevated CRP and WBC [59]. These observations are alarming because antibiotic resistance, caused by antibiotic misuse or overuse, is one of the most serious threats to public health and a very important challenge for the current medicine. The WHO, European Commission and health organizations in many countries introduced programs and provided guidelines to prevent antibiotic resistance [60‐63]. Despite these recommendations, wide spectrum antibiotics were often administered to SAT patients. Thus, the educational programs, especially directed for GPs should be provided.

For the last year, this issue has been appearing as more important than ever due to COVID-19 pandemic. As it has been described above, the proper SAT diagnosis is even more difficult in patients with COVID-19.

False negative SAT diagnosis results in a delayed treatment and disturbs patient’s quality of life, although it is not a life-threatening situation. On the other hand, false positive SAT diagnosis can refer to patients with thyroid primary and metastatic malignant tumors of poor prognosis. In 2019, we presented several cases of such malignancies, underlining that the clinical presentation mimicking SAT led to a delayed diagnosis which shorted the patients’ life [64]. The proper diagnosis was made in our Department after—not rarely—many weeks of previous incorrect treatment with NSAIDs and/or steroids administered by GPs or various specialists. Those cases gave us a significant impulse to re-analyze the usefulness of the SAT diagnostic criteria and to modify them adequately to the current state of knowledge. For several decades, SAT diagnosis had been made mostly on the basis of the presence of neck pain and significantly accelerated ESR, which were considered the main diagnostic criteria. These two main criteria plus one of the additional criteria (significantly reduced iodine uptake, transient thyrotoxicosis, fever, typical US pattern, typical FNAB result, low concentration of anti-thyroid antibodies) had been considered enough to diagnose SAT. All of the patients with advanced malignancies, presented in our report, met both the main criteria and one of the additional criteria [64]. Malignant tumors of poor prognosis, such as anaplastic thyroid cancer or progressive neoplasms with thyroid metastases may present with neck pain due to thyroid capsule extension, and elevated ESR due to the advanced progression of cancer. Additionally, thyrotoxicosis can be observed due to the gland destruction, iodine uptake can be decreased and even fever can occur because of the presence of abscess in the necrotic tumor masses. We have reported 5 cases, including 4 anaplastic thyroid carcinomas and one non-small cell lung cancer thyroid metastasis [64]. All of them suffered from neck pain and all had accelerated ESR. The proper diagnosis was made on the basis of atypical US pattern and FNAB results. Thus, we concluded that no SAT diagnosis can be made without US examination [64]. However, one should remember that some US features of malignant lesions, such as blurred lesion margins or hypoechogenity, may be similar to those of SAT. Pan et al. analyzed sonographic features of 165 lesions diagnosed as SAT or malignancies and concluded that useful sonographic predictors of SAT are poorly defined margin, centripetal reduction of echogenicity and the absence of internal vascularity [65]. Nonetheless, a significant overlap of features of SAT and malignant nodules was demonstrated [65]. The possible difficulties in distinguishing SAT and malignancies were also described by other authors [66]. Therefore, if US pattern is doubtful, FNAB should be considered the decisive test [64].

The burning issue of false positive diagnosis of SAT on the basis of neck pain and clinical and biochemical features of inflammation has been raised for many years. Meier and Nagle reported cases of laryngeal carcinoma, Hodgkin’s disease and metastatic lung carcinoma which were initially diagnosed and treated as SAT [67]. Laboratory markers of thyrotoxicosis were present in most of the cases. The final diagnosis of malignancy was made only after the US and FNAB were finally performed due to lack of the treatment effectiveness [67]. Similar situations were reported in the cases of thyroid lymphoma [68], thyroid primary carcinomas, including even follicular carcinoma [69], and thyroid metastases of numerous different malignancies including mainly lung carcinomas [70] and adenocarcinomas of different origins [71, 72]. In all of these cases, the proper diagnosis was possible only on the basis of FNAB result.

Jonklaas has recently analyzed cases of thyrotoxicosis caused by thyroid infiltration by non-thyroid malignancies [73]. This kind of thyroiditis-like presentation was observed mostly in metastases of breast and lung cancers, as well as in hematologic malignancies. The author underlined that the malignancy-associated hyperthyroidism resulted from a destruction of thyroid tissue – a mechanism similar to the one causing thyrotoxicosis in SAT [4].

Therefore, taking into account the large number of malignancies which can mimic SAT, the differential diagnosis based on US should always be introduced, and FNAB should be performed in all doubtful cases.

Recent reports have proven that SAT diagnosis is actually not as easy as it was believed before. Malignancies can be erroneously diagnosed as SAT and vice versa. Moreover, one should always remember about rare – but possible – co-presence of SAT and differentiated thyroid carcinoma (DTC). A typical SAT US pattern, being a manifestation of the inflammatory process, does not warrant lack of malignancy. The simultaneous presence of SAT and DTC was believed to be very uncommon and used to be presented mostly as case reports [74, 75]. However, the frequency of co-occurrence of SAT and DTC has recently been proved as higher than it was previously thought. Nishihara et al. conducted a study involving 710 patients with SAT who underwent US examinations at initial diagnosis and during follow-up, with subsequent FNAB evaluation of suspicious nodules [76]. The US examination at the initial screening of thyroid nodules in SAT patients showed a sensitivity of 72.4%, specificity of 89.0%, positive predictive value of 80.4%, and negative predictive value of 83.8%. Co-presence of papillary thyroid carcinoma (PTC) was found in 3.1% of SAT patients. As many as 30% of PTC cases were not identified during the initial scan and were only found during the follow-up US. Interestingly, all tumors in this group were latently localized in the SAT-related bilateral hypoechoic areas of the thyroid and showed no calcifications [76]. On the contrary, out of the 15 tumors detected during the initial examination, 7 presented calcifications and 5 were located in areas unaffected by SAT. The authors underlined that SAT can significantly obscure coexisting PTC when inflammatory hypoechoic regions are present [76]. Gül et al. confirmed the diagnosis of PTC in even higher percentage of SAT patients, reaching 4.4% [77]. Thus, US re-examination after resolution of SAT-related thyroid lesions should always be performed. Any suspicious thyroid lesions – nodules or areas – which do not resolve or which appear after resolution of SAT should undergo FNAB.