Introduction

Subacute thyroiditis (SAT), or De Quervain thyroiditis, is a transient inflammatory disorder of the thyroid gland caused by a viral infection or a post-viral inflammatory response. It accounts for up to 5% of all thyroid abnormalities and mostly affects females in their fourth and fifth decades of life [1, 2]. During the early stages of SAT, most patients experience anterior neck pain, fever, malaise, and mild thyrotoxic symptoms. These symptoms, along with some laboratory results such as elevated erythrocyte sedimentation rate (ESR), low thyroid-stimulating hormone (TSH), and suppressed radioactive iodine uptake (RAIU), are sufficient for the diagnosis [3]. The usual course of SAT involves a thyrotoxic phase lasting 3 to 6 weeks due to the destruction of thyroid follicles, resulting in the excessive release of thyroid hormones into the bloodstream, which is then followed by a hypothyroid phase, typically resolving within six months [4]. Although in most patients, thyroid functions return to normal within twelve months, 4% to 26% of patients might develop permanent hypothyroidism [2, 5]. Recurrence might also happen in some patients. A systematic review and meta-analysis of 18 cohort studies revealed that 12% of individuals with SAT experience episodes of recurrence [6].

Thyroid autoantibodies have a central role in autoimmune thyroid diseases (AITD). They include thyroid peroxidase antibody (TPOAb) in Hashimoto’s thyroiditis (HT) and TSH receptor antibodies (TRAB) in Grave’s disease (GD). These antibodies are well-established for diagnosing AITD. Moreover, they play a crucial role in assessing treatment response and determining the appropriate time to stop treatment [7, 8]. The literature is scarce on the possible role of these antibodies in SAT.

A wide range of thyroglobulin antibody (TgAb) and TPOAb positivity rates are reported among SAT patients. Several studies showed that TgAb was positive in less than 10% of SAT patients [5, 9, 10]. Conversely, some studies found TgAb positivity in more than 45% of patients. For instance, Omori et al. reported its positivity in as high as 66% of SAT patients [11,12,13]. The same controversy exists for TPOAb. While some studies reported its positivity in around 5% of SAT cases [2, 9, 14], several studies state that TPOAb is positive in more than 25% of the patients. Among them is a study conducted by Sato et al., reporting TPOAb to be positive in 38% of the cases [12, 15, 16]. Therefore, we conducted this systematic review and meta-analysis to find the pooled prevalence of thyroid autoantibodies in patients with SAT. We also investigated the recurrence rate and the occurrence of permanent hypothyroidism in SAT patients as our secondary objectives.

Methods

This systematic review and meta-analysis was conducted and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [17]. The study was conducted based on a predefined protocol registered in PROSPERO (CRD42022366629).

Search Strategy

A comprehensive literature search was performed in Pubmed, Embase, Scopus, and Web of Science from their inception until March 25th, 2023, with the restriction of the English language. We used MeSH, Emtree, the free text method, and expert opinion to select the most appropriate search terms. The full search strategies of each electronic database are available in the supplementary material (Table S1).

Study selection and eligibility criteria

We imported the retrieved studies into Endnote software and removed duplicate studies. First, two reviewers (AS and AY) independently screened studies by title and abstract. Then they examined the full texts of the remaining potentially eligible studies against inclusion and exclusion criteria. Any disagreements were resolved through discussion, and if any disagreements remained unresolved, we sought a third expert reviewer’s opinion (MH). Additionally, the reference lists of included studies were hand-searched to find any undetected eligible studies.

Observational studies published in English that reported the prevalence of thyroid autoantibodies’ positivity in more than 10 SAT patients were eligible to be included in our study.

The exclusion criteria were: (1) Studies that included (i) patients with a history of prior autoimmune or thyroid diseases before the development of SAT; (ii) pregnant or lactating patients; (iii) patients with a history of thyroid surgery or radiotherapy to the head and neck; (iv) patients receiving certain drugs, including lithium, amiodarone, methimazole, carbimazole, radioactive iodine, interferon alpha, immune checkpoint inhibitors, and tyrosine kinase inhibitors. (2) studies that used the broad term of thyroid antibodies and did not specify the name of the antibody. (3) studies with inclusion or exclusion criteria of positive or negative thyroid autoantibodies. (4) Studies that included patients with prior use of corticosteroids before the development of SAT were excluded. However, since corticosteroids are one of the main treatment options for SAT, excluding the studies that used corticosteroids to treat SAT would have resulted in the loss of many valuable studies from our systematic review. Furthermore, in most cases, corticosteroid treatment was initiated after the measurement of thyroid autoantibodies. Thus, we decided not to exclude these studies.

Data extraction

The following data were extracted from the studies by two reviewers (SM-T and AS) independently: first author, publication year, country, study design, the total number of SAT patients, the number of men and women, the mean age of patients, the number of SAT patients with antibody data, the number of SAT patients with positive thyroid autoantibodies, the assays used for thyroid autoantibody detection and their cut-off values, the number of patients that developed recurrence, the number of patients that developed permanent hypothyroidism, and the mean and standard deviation (SD) of laboratory tests including TSH, T3, T4, and ESR.

Quality assessment

The Joanna Briggs Institute’s (JBI) critical appraisal checklist for studies reporting prevalence data was used for the quality assessment of the included studies [18]. Two reviewers (MF and AS) independently assessed the quality of the included studies. Disagreements were resolved by discussion. In cases where disagreements persisted, a third reviewer’s opinion was asked (MH). The JBI critical appraisal checklist consists of nine questions. The checklist does not provide a specific cut-off score. We categorized studies with scores of 7 or higher as having a low risk of bias, those with scores of 5 or 6 as having a moderate risk of bias, and those with scores of 4 or lower as having a high risk of bias.

Statistical analysis

Stata software version 14.0 was used for meta-analysis. The metaprop command was used to calculate the estimated pooled prevalence with a 95% confidence interval (CI). The inconsistency index (I2) was used to assess the degree of heterogeneity across studies. I2 < 25%, 25 ≤ I2 < 50, and I2 ≥ 50 were considered low, moderate, and high heterogeneity, respectively. Due to substantial methodological and statistical heterogeneity between studies, the random effect model was chosen.

We performed subgroup analysis based on publication year, study design, risk of bias, mean ESR level, mean TSH level, and mean age of patients to find the potential sources of the heterogeneity for the prevalence of both TPOAb and TgAb.

The funnel plot, Begg, and Egger tests were used to assess the existence of publication bias. If the funnel plot was asymmetrical or any of the Begg or Egger tests indicated the potential presence of publication bias, the Duval and Tweedie trim and fill method was used to explore the influence of publication bias [19,20,21]. The leave-one-out sensitivity analysis was implemented to assess the robustness of our findings by removing one study at a time and estimating the pooled prevalence with the rest of the studies. The significance level was set at a p value of < 0.05 for all the analyses.

Results

Study selection and study characteristics

Our literature search results and study selection process are presented in the PRISMA flow diagram (Fig. 1). The literature search in the four above-mentioned electronic databases yielded 2061 results, of which 688 were duplicates. After title and abstract screening, 127 studies were eligible for full-text examination. The full text of nine studies could not be retrieved. Eighty-seven studies were excluded due to (1) lack of prevalence data for more than 10 SAT patients (n = 50); (2) meeting one of the exclusion criteria (n = 18); (3) conference papers, case reports, and review studies (n = 11); (4) unclear SAT diagnosis or unclear antibody data (n = 8). One additional study was found in the reference lists of the included studies. Finally, 32 observational studies were included in our systematic review, and 24 were included in the meta-analysis.

Fig. 1
figure 1

Flow diagram of the study selection process

Full size image

Table 1 represents the main characteristics of the included studies. In terms of study design, there were 18 cross-sectional studies, 10 retrospective cohort studies, and 4 prospective cohort studies. Geographically, 22 studies were carried out in Asia, 6 in Europe, 2 in North America, 1 in South America, and 1 international study. The included studies had a total of 2348 SAT patients with a female preponderance (81.57%) and a mean age of 44.26 ± 4.5 years. The prevalence of TPOAb, TgAb, TRAB, thyrotropin binding inhibiting immunoglobulins (TBII), and thyroid-stimulating immunoglobulins (TSI) positivity was reported in 22, 22, 12, 4, and 5 studies, respectively. Seven cohort studies investigated the number of patients who developed permanent hypothyroidism. Eleven cohort studies reported the number of patients who developed recurrence.

Table 1 Main characteristics of the included studies
Full size table

Quality assessment

The results of the quality assessment using the JBI checklist are presented in Table S2. Out of the 32 included studies, 16 had a high risk of bias, 10 had a moderate risk of bias, and 6 had a low risk of bias. The mean JBI score of the studies was 4.62 ± 1.56.

Data synthesis (meta-analysis)

The overall pooled prevalence of TgAb positivity in 22 studies was 22.8% (95% CI: 16.3–29.9%). With a high level of heterogeneity (I2 = 89%). With the highest prevalence of 66.7% reported by Omori et al. and the lowest prevalence of 0% reported by Bogazzi et al. [9, 11] (Fig. 2).

Fig. 2
figure 2

Forest plot showing the pooled prevalence of thyroglobulin antibody

Full size image

The overall pooled prevalence of TPOAb positivity in 22 studies was 12.2% (95% CI: 9.4–15.3%). With a high level of heterogeneity (I2 = 63%), with the highest prevalence of 38.1% reported by Sato et al. and the lowest prevalence of 0% reported by Bogazzi et al. [9, 12] (Fig. 3).

Fig. 3
figure 3

Forest plot showing the pooled prevalence of thyroid peroxidase antibody

Full size image

The pooled recurrence rate of SAT patients reported in 11 cohort studies was 14.7% (95% CI: 12.2–17.3%) with moderate heterogeneity (I2 = 33%) (Fig. 4).

Fig. 4
figure 4

Forest plot showing the pooled recurrence rate

Full size image

The pooled estimate of developing permanent hypothyroidism reported in 7 cohort studies was 11.6% (95% CI: 6.5–18%) with high heterogeneity (I2 = 84%) (Fig. 5).

Fig. 5
figure 5

Forest plot showing the pooled occurrence of permanent hypothyroidism

Full size image

Subgroup analysis

The results of subgroup analyses are presented in Table 2, Figs. S1 and S2. Subgroup analysis was performed based on the following variables: publication year, study design, risk of bias, mean ESR levels, mean TSH levels, and mean age of patients. For the estimated pooled prevalence of TPOAb, the mean ESR level was identified as a source of heterogeneity (the overall I2 of 53.8% was reduced to 34.4% in the mean ESR ≥ 60 subgroup). The mean TSH level was another source of heterogeneity (the overall I2 of 61% was reduced to 41.1% in the mean TSH ≥ 0.04 subgroup). The study design was another source of heterogeneity (the overall I2 of 63.4% was reduced to 0% in the cross-sectional subgroup). None of the mentioned variables was identified as a source of heterogeneity for the pooled prevalence of TgAb.

Table 2 Subgroup analysis
Full size table

Publication bias and sensitivity analysis

Based on the symmetrical look of the logit-transformed prevalence funnel plot of TPOAb (Fig. S3) and the results of the Begg (p value = 0.31) and Egger tests (p value = 0.06), publication bias was not suspected. The logit-transformed prevalence funnel plot of TgAb seemed mildly asymmetric (Fig. S3), indicating a risk of potential publication bias. Begg (p value = 0.21) and Egger tests (p value = 0.11) revealed no risk of potential publication bias. Duval and Tweedie’s trim and fill technique showed no missing studies. The leave-one-out sensitivity analysis showed that omitting any of the included studies does not change the pooled estimate of prevalence substantially (Fig. S4).

Discussion

Although SAT is not classified as an AITD, a wide range of thyroid autoantibody positivity rates have been reported in the literature.

Based on our findings, TPOAb is positive in 12.2% of SAT patients, which is similar to the general population. Based on the National Health and Nutrition Examination Survey III (NHANES), TPOAb is positive in 13% of the general population [22]. Another population-based cohort study demonstrated a TPOAb positivity rate of 12.8% in the general population [23]. ESR and CRP are almost always elevated in SAT, and their absence might question the diagnosis of SAT [24]. SAT severity is defined based on clinical judgment and the levels of inflammatory markers [25]. Bahadir et al. used a combination of clinical features and an ESR cut-off value of 60 mm/h to divide SAT patients into mild and severe cases [14]. We used the same ESR cut-off value in one of our subgroup analyses and found the ESR levels to be a source of heterogeneity between studies. Nevertheless, the difference in TPOAb positivity between the two groups (12.1% vs. 9.7%) was not significant (interaction test p value = 0.335), Which is in contrast with Luotola et al.’s finding of a negative correlation between ESR and TPOAb levels in SAT patients [26]. Subgroup analysis revealed that the mean TSH level was another source of heterogeneity between studies. However, the prevalence of positive TPOAb was not significantly different between the two groups (interaction test p value = 0.671).

The estimated pooled prevalence of TgAb positivity among SAT patients was 22.8%, which is higher than the general population. Several population-based studies have reported TgAb to be present in 10% to 12% of the general population [22, 27, 28]. The observed difference in TgAb positivity between SAT patients and the general population could be explained by the destruction of thyroid follicles during the thyrotoxic phase of SAT, which leads to the subsequent release of thyroglobulin into the bloodstream, thereby triggering the production of TgAb [14]. In a recent study by Ricci et al., TgAb-IgM was positive in 10 out of 16 patients. However, none of the patients developed an AITD during the follow-up period. They stated that SAT is a non-autoimmune thyroid disease associated with the transient positivity of thyroid autoantibodies [29]. In line with this, Nishihara et al. observed that TgAb decreased or disappeared in all six patients initially positive for TgAb who had follow-up data [13]. Similarly, Volpe et al. reported that thyroid autoantibodies disappeared in 21 out of 30 SAT patients during the follow-up period [30]. Ricci et al. suggested that TgAb-IgM plays a protective role against developing a persistent thyroid autoimmune response by increasing the clearance of thyroglobulin from the bloodstream. Further prospective research is warranted to validate this hypothesis [29].

TSH receptor antibodies (TRAB) are a group of heterogeneous antibodies against TSH receptors and are positive in over 90% of patients with Graves’ disease [31]. Two main assays are commonly used for measuring TRAB. These include competition-based immunoassays that measure thyrotropin binding inhibiting immunoglobulins (TBII) and cell-based bioassays that detect thyroid-stimulating immunoglobulins (TSI). TBII assays cannot differentiate stimulating from non-stimulating TRAB. On the other hand, TSI assays detect only stimulating TRAB by measuring cyclic adenosine monophosphate (cAMP) production [32, 33]. We decided not to perform a meta-analysis for these antibodies for various reasons. Firstly, most studies reported their prevalence to be less than 3%, with many reporting their prevalence to be 0% in SAT patients. Secondly, the number of studies reporting each measurement method was insufficient to perform subgroup analysis and look for sources of heterogeneity. Thirdly, studies reporting a prevalence of 0% would pose a challenge to logit transformation, which is necessary for publication bias assessment. And most importantly, we could not combine the data of TBII, TSI, and TRAB (not specifying TBII or TSI) since some studies reported both TBII and TSI measurement methods and would be included twice in the overall estimate.

Recurrence is a common phenomenon among SAT patients. Eleven cohort studies included in our study investigated the recurrence rate of SAT. Their findings ranged from 8.8% to 22.5% [14, 30]. As one of our secondary endpoints, we found a recurrence rate of 14.7% in SAT patients, which is in line with a systematic review and meta-analysis published in 2021 that pooled the recurrence rate of 18 cohort studies and found a recurrence rate of 12% [6]. The literature regarding the association between thyroid antibodies and SAT recurrence is scarce and controversial. Stasiak et al. found TPOAb to be a strong protective factor against recurrence. Conversely, Erdem et al. found no significant difference in recurrence rates between patients with or without thyroid autoantibodies [2, 34].

Permanent hypothyroidism was another secondary endpoint of our study. Our meta-analysis of seven cohort studies showed that 11.6% of SAT patients develop permanent hypothyroidism. Our finding aligns with the American Thyroid Association Guidelines, which report permanent hypothyroidism in 5% to 15% of SAT cases [35]. All the included studies reported permanent hypothyroidism in less than 20% of cases. The only exception was a study by Gorges et al., which found permanent hypothyroidism in 26.8% of patients. They attributed their finding to their cases’ severity and longer follow-up period [5]. Sencar et al. found TPOAb positivity as a risk factor for permanent hypothyroidism. In contrast, Saklamaz found no correlation between thyroid autoantibodies and permanent hypothyroidism [10, 36].

Coronavirus disease 2019 (COVID-19), primarily acknowledged as a respiratory infection, is known for its multi-organ involvement, including the thyroid gland [37]. The expression of angiotensin-converting enzyme 2 (ACE2) receptors by thyroid cells serves as an entry point for SARS-CoV-2, making the thyroid gland susceptible to the virus [38]. Notably, SAT has emerged as one of the most prevalent forms of COVID-19-related thyroiditis [39]. A recent systematic review comprising 24 articles summarizing data from 69 SAT cases caused by COVID-19 revealed an average onset time of 32 days for SAT since the initiation of the COVID-19 illness. Most patients presented with typical signs and symptoms, and glucocorticoids were the primary mode of treatment. Similar to cases before the COVID-19 era, the symptomatic response to treatments proved to be excellent. TRAB was negative in all 28 tested patients. TgAb and TPOAb levels were available in only a few patients and were positive at low levels [40]. Another systematic review encompassing 43 articles with 100 SAT cases caused by COVID-19 reported that among patients with available data, TPOAb and TgAb were positive in 17% and 18% of patients, respectively. Consistent with the previously mentioned systematic review, TRAB was negative in all 24 tested patients [41]. A study by Brancatella et al. compared the clinical and laboratory findings of SAT cases with positive COVID-19 tests within a 45-day window before SAT onset with pre-pandemic SAT cases. Their findings indicated a more severe clinical picture in SAT cases associated with COVID-19. Notably, laboratory results revealed significantly higher ESR, CRP, and FT4 in SAT cases induced by COVID-19. Additionally, SAT cases associated with COVID-19 exhibited a higher chance of developing hypothyroidism [25].

To the best of our knowledge, this is the first systematic review and meta-analysis to estimate the pooled prevalence of thyroid autoantibodies’ positivity in SAT patients. We implemented a comprehensive search strategy in the four main electronic databases of the medical field. Our extensive data extraction sheet enabled us to perform subgroup analysis for six variables, revealing mean ESR, mean TSH, and study design as sources of heterogeneity for the prevalence of TPOAb positivity among SAT patients. Furthermore, our result was not influenced by publication bias. Lastly, the leave-one-out sensitivity analysis method showed the robustness of our results. Our study has several limitations as well. Non-English-language studies are not included. We could not estimate the pooled prevalence of thyroid autoantibodies in men and women separately since only a few studies reported that data. Despite our efforts, we could not find sources of heterogeneity for the prevalence of TgAb positivity. This high heterogeneity can be explained by the fact that different assays and cut-off values were used to measure TgAb in different studies and that antibody measurements were done at different times in different studies. Moreover, 26 of the 32 included studies are categorized as having a high or moderate risk of bias. Since SAT is not a common disorder, most of the included studies had a sample size of less than 100 SAT cases. Additionally, the non-random sampling methods employed in the studies may have introduced selection bias by including more severe cases that sought medical care. Moreover, measuring antibodies at different times from the initial symptoms across different studies could introduce additional bias.

Although we tried our best to avoid, minimize, and evaluate the effects of various biases on the results of our study by utilizing both methodological and statistical approaches, it is worth noting that the fact that autoantibodies of any single patient in original studies are measured during a specific phase of SAT could have introduced bias and influenced the pooled findings of our study. Although we performed a subgroup analysis based on the mean TSH level, since this bias lies within the core of each original study, a direct assessment of its impact on our pooled results was not feasible.

In conclusion, our study revealed a low TPOAb positivity rate in SAT patients, consistent with its non-autoimmune origin. The prevalence of TgAb positivity among SAT patients was higher than that of the general population, possibly resulting from the transient release of thyroglobulin into the bloodstream during the thyrotoxic phase, triggering subsequent TgAb production. Moreover, the results of this systematic review and meta-analysis demonstrated a notable recurrence rate and the development of permanent hypothyroidism among SAT patients, emphasizing the crucial need for ongoing follow-up care. Further prospective studies are warranted to elucidate the relationships between thyroid autoantibodies and disease severity, recurrence, and permanent hypothyroidism.