Subacute Thyroiditis in Denmark: A Nationwide Study of 1763 Cases

ABSTRACT

Objective

Subacute thyroiditis (SAT) is a differential diagnosis in the clinical care of hypo‐ and hyperthyroid patients, but large‐scale population data on the incidence, patient characteristics, and outcome are limited. Thus, the aim was to determine the incidence of SAT in the Danish population and to characterise patient demographics and the risk of permanent hypothyroidism.

Design

Retrospective cohort study based on Danish nationwide health registers.

Patients

Individuals diagnosed with SAT at Danish hospitals from 1995 to 2016, covering the years before and after the implementation of mandatory iodine fortification in Denmark as of July 2000.

Measurements

Annual incidence rates of SAT were calculated overall and by geographical region (East and West Denmark). The development of permanent hypothyroidism following SAT was evaluated based on redeemed prescriptions of levothyroxine (L‐T4) within a 4‐year follow‐up period.

Results

A total of 1763 individuals were identified with a diagnosis of SAT among 1,747,168 individuals at risk. The overall nationwide incidence increased from 3.3/100,000 individuals (95% confidence interval (CI): 2.5–4.4) in 1995 to 6.2/100,000 individuals (95% CI: 5.1–7.6) in 2016, with no regional difference (p = 0.18). Follow‐up showed that 459 of 1763 patients with SAT (26.0%) started treatment with L‐T4, and in 224 of these patients (48.7%), the duration of L‐T4 treatment exceeded 36 months.

Conclusions

In a Danish nationwide cohort, the incidence of SAT increased during a 22‐year period without geographical difference. Notably, the frequency of long‐term L‐T4 treatment was high, emphasising a focus on the post‐diagnosis management of patients with SAT.

1. Introduction

Subacute thyroiditis (SAT) is an inflammatory disorder of the thyroid gland. The aetiology is not completely clear, but the condition has been linked to an exaggerated immune response after a viral infection in individuals who are genetically predisposed [1, 2]. A distinctive clinical course characterises SAT due to the destructive inflammation of the thyroid gland [3], and the clinical hallmarks are anterior neck pain and thyroid dysfunction. The damage of follicular cells leads to the extravasation of stored thyroid hormones, which causes a transient thyrotoxic phase. As hormone stores eventually deplete, and the synthesis of new hormones is impaired, a subsequent hypothyroid phase may occur [3]. Most patients will recover to euthyroidism, but some develop permanent hypothyroidism [4]. Patients often present with SAT in the initial thyrotoxic phase, and the specific diagnosis of SAT, as opposed to other subtypes of thyrotoxicosis, may pose a challenge at the time of clinical presentation. Correct diagnosis, however, is important to ensure proper management, the choice of treatment and accurate patient information.

Previous studies on SAT have included a relatively small number of patients [4, 5, 6, 7, 8, 9], and evidence from large population‐based cohorts is encouraged to provide accurate figures on the occurrence of SAT and to substantiate patient characteristics and treatment trends. We established a nationwide cohort of patients diagnosed with SAT in Danish hospitals using the Danish nationwide health registers. The study period was from 1995 to 2016, encompassing the introduction of the Danish mandatory iodine fortification of salt as of July 1, 2000 [10]. Denmark has a history of long‐term iodine deficiency (ID) with regional differences, as indicated by mild ID in East Denmark and moderate ID in West Denmark when divided by the Great Belt [11]. Thus, the nationwide study cohort enabled us to provide figures on the incidence of SAT overall and to study the development over time and in each geographical region. Furthermore, the large cohort of patients with SAT provided an opportunity to evaluate patient characteristics and to follow up on each patient for subsequent management and treatment of the disease.

2. Materials and Methods

2.1. Study Cohort

This is a retrospective register‐based cohort study of individuals diagnosed with SAT in Denmark from January 1, 1995, to December 31, 2016. All citizens of Denmark are assigned a unique 10‐digit personal identification number as part of the Danish Civil Registration System [12]. This system enables the linking of health‐related registrations at an individual level. The study population was identified using the Danish Medical Birth Register (MBR) [13], which includes all births in Denmark and information on the mother (from 1973 onwards) and the father (from 1990 onwards). Thus, we included women in these records from January 1, 1973, and their male partner from January 1, 1990, and up until December 31, 2016 (n = 2,277,910). Everyone was followed from January 1, 1995, until December 31, 2016, for a diagnosis of SAT, and we defined age of entry into the cohort as minimum 18 years to ensure that follow‐up was during adulthood in all study participants. Thus, individuals below the age of 18 as of January 1, 1995, were excluded (n = 530,742), leaving a study population at risk of 1,747,168 unique individuals for follow‐up.

The study was registered in accordance with the General Data Protection Regulation (GDPR) in the North Denmark Region (2016‐76). Ethical permission is not required for register‐based studies in Denmark.

2.2. Diagnosis of SAT

The Danish National Hospital Registry (DNHR) was used to identify individuals diagnosed with SAT, and each study participant was under observation for SAT diagnoses from January 1, 1995, to December 31, 2016. The DNHR contains nationwide registrations of diagnoses for all in‐ and outpatient contacts in a Danish hospital [14], and the diagnoses have been registered according to the 10th version of the International Classification of Diseases and Related Health Problems (ICD‐10) since 1995. In the DNHR, all diagnoses of SAT (DE061) were identified in the cohort period (1995–2016). Other diagnoses were identified to support the definition of SAT, including postpartum thyroiditis (DO905) (1995–2017), thyroid surgery (KBAA20‐KBAA60) (1996–2017) and radioiodine therapy (WEETTJLXX) (1999–2017). The Danish National Prescription Registry (DNPR) contains nationwide data on all redeemed prescriptions of medication distributed at Danish pharmacies since 1994, registered according to the Anatomical Therapeutic Chemical (ATC) classification system [15]. We identified the following medications: L‐T4 (H03A) (1995–2020), amiodarone (C01BD01) (1995–2016) and lithium (N05AN) (1995‐2016).

Overall, we identified 1887 individuals who had a SAT diagnosis registered in the DNHR between 1995 and 2016. Individuals who underwent thyroid surgery or received radioiodine therapy within 6 months of the SAT diagnosis (n = 28), women diagnosed with postpartum thyroiditis within 3 months (n = 8), individuals treated with amiodarone or lithium within 3 months prior to diagnosis (n = 14), and individuals receiving long‐term thyroid hormone replacement therapy (minimum two redeemed L‐T4 prescriptions), indicating preexisting autoimmune thyroiditis (n = 74), were excluded. In total, we excluded 124 individuals (6.6%), leaving a final study cohort of 1763 unique individuals diagnosed with SAT. For each SAT case, initiation and duration of L‐T4 treatment were assessed for a 4‐year follow‐up period from the diagnosis date. Therapeutic regimens with non‐steroidal anti‐inflammatory drugs (NSAIDs) (M01A) (1995–2020) and corticosteroids (H02A) (1995–2020) were also identified using the DNPR. Treatment occurring within the acute phase of the disease, defined as 3 months before to 3 months after the SAT diagnosis, was considered related to SAT. Individuals with ongoing treatment in the months prior to this acute phase were not considered in the assessment, as SAT was unlikely to be the indication for treatment.

2.3. Biochemical Assessment

To support the validity of the SAT case identification, a biochemical assessment was conducted in a subgroup of individuals diagnosed in the North Denmark Region, which is geographically part of western Denmark. The biochemical assessment was available for study purposes in the region, and data covered 82 of the 1763 SAT cases, corresponding to the region's relative share based on the national population. We utilised the clinical laboratory information system (LABKA II) database [16], which contains biochemical test results collected as part of routine patient care from patients in general practices and hospital departments in the North Denmark Region from 2006 to 2020. The biochemical results were assessed in each SAT case for a 4‐year follow‐up period from the diagnosis date and an additional 3 months prior to the date of diagnosis, to ensure that the assessment included the early phase of the disease. The assessment included thyroid function parameters (thyroid‐stimulating hormone (TSH), total triiodothyronine (TT3) and total thyroxine (TT4)) as well as thyroid autoantibodies (TSH‐receptor antibodies (TRAb), thyroid peroxidase antibodies (TPO‐Ab) and thyroglobulin antibodies (Tg‐Ab)). Additionally, inflammatory parameters, C‐reactive protein (CRP) and erythrocyte sedimentation rate (ESR), were included in the assessment. The following reference ranges were applied in the routine laboratory during the study period: TSH: 0.3–4.5 mIU/L; TT3: 1.1–2.5 nmol/L; TT4: 60–140 nmol/L; CRP: < 8 mg/L; ESR: females, 0–50 years < 20 mm/h, 50–200 years < 30 mm/h; males, 0–50 years < 15 mm/h, 50–200 years < 20 mm/h; TRAb: < 1.5 UI/L; TPO‐Ab: < 60 U/mL; Tg‐Ab: < 60 U/mL. Automated immunoassays were used for measurement of TSH, TT3 and TT4 (Roche Diagnostics and Siemens Healthineers) and TPO‐Ab and Tg‐Ab (Thermo Fischer Diagnostics Aps). CPR was measured using an automated immunoturbidimetric assay (Roche Diagnostics and Siemens Healthineers), and TRAb was measured using a radioimmunoassay (BRAHMS Diagnostica).

2.4. Statistical Analyses

The annual incidence rate of SAT was calculated as the number of individuals with SAT divided by the population at risk and reported as events per 100,000 individuals. The population at risk was updated annually by excluding individuals who had died or who had already been diagnosed with SAT. We used Wald test to assess whether the incidence rates differed across regions. Poisson regression was applied to determine if the incidence rates of SAT increased during the study cohort period when adjusting for age and sex, and to assess seasonal variation in SAT occurrence.

A control group was established in a 5:1 ratio matched to patients diagnosed with SAT by age, sex and region, which was possible for 1753 of the 1763 SAT cases. Characteristics were assessed at the date of diagnosis for cases and at the corresponding date for controls. The chi‐square test was used to compare differences in demographics and treatment trends across periods.

The biochemical measurements were assessed in defined periods around and after the time of diagnosis of SAT, and the first result was assessed if multiple measurements were performed within a time window. Biochemical test results below the detection limit for the specific analyte were replaced by half the concentration of the lowest measurement range [17]. Statistical analyses were performed using STATA version 19 (StataCorp LLC, College Station, Texas, USA).

3. Results

We included a total of 1763 unique individuals diagnosed with SAT in a Danish hospital during a 22‐year period. Most cases were female (86.1%), and the median age at the start of follow‐up in 1995 was 33 years (range: 18–57) and 45 years (range: 19–76) at the time of being diagnosed with SAT. Cases were followed in the hospital for a median of 76 days (IQR: 4–211 days). We stratified by geographical region in Denmark, and 53.3% of the SAT cases were diagnosed in West Denmark. These cases were slightly older, more often of Danish origin, and had a lower socioeconomic status (SES) compared to cases diagnosed in East Denmark (Table 1). Compared with matched controls, SAT cases were more often of non‐Danish origin, had higher SES as reflected by educational attainment, occupational status and income levels, and they were more often married (Table 2).

Table 1.

Baseline characteristics of individuals diagnosed with subacute thyroiditis at the time of diagnosis.

	Nationwide cohort		East cohorta		West cohorta
	n	%	n	%	n	%
Individuals	1763	100	821	46.7	938	53.3
Sex
Female	1517	86.1	696	84.8	819	87.3
Male	246	13.9	125	15.2	119	12.7
Age start follow‐up (1995)
18–29 years	642	36.4	318	38.7	321	34.2
30–40 years	756	42.9	347	42.3	409	43.6
> 40 years	365	20.7	156	19.0	208	22.2
Age when diagnosed
18–29 years	75	4.3	41	5.0	33	3.5
30–40 years	480	27.2	234	28.5	244	26.0
> 40 years	1208	68.5	546	66.5	661	70.5
Paritya, b
Nulliparous	217	14.9	119	17.9	98	12.4
Multiparous	1236	85.1	544	82.1	692	87.6
Origin
Danish	1565	88.8	700	85.3	862	91.9
Non‐Danish	198	11.2	121	14.7	76	8.1
Marital statusa
Married	1229	70.0	565	69.0	664	70.9
Not married	527	30.0	254	31.0	272	29.1
Educational attainmenta, c
Basic	336	19.3	147	18.2	189	20.4
Low	131	7.6	66	8.1	65	7.0
Middle	1147	66.0	522	64.4	625	67.4
High	124	7.1	75	9.3	48	5.2
Occupational status
Employed	1465	83.1	693	84.4	770	82.1
Not employed	298	16.9	128	15.6	168	17.9
Income level
1st quartile (lowest)	449	25.5	191	23.3	254	27.1
2nd quartile	437	24.8	198	24.1	239	25.5
3rd quartile	446	25.3	192	23.4	254	27.1
4th quartile	431	24.4	240	29.2	191	20.3

^a Missing values: region: n = 4; parity: n = 64; marital status: n = 7; educational attainment: n = 25.

^b Males not included.

^c Educational attainment was characterised as: ‘basic’ (primary/secondary education only), ‘low’ (basic education and vocational education), ‘middle’ (short‐ or medium‐cycle higher education) and ‘high’ (long‐cycle higher education).

Table 2.

Baseline characteristics of individuals diagnosed with subacute thyroiditis (SAT) and matched controls.

	Cases of SATa		Matched controls
	n	%	n	%	p‐value
Individuals	1753	100	8788	100
Sex					0.958
Female	1509	86.1	7569	86.1
Male	244	13.9	1219	13.9
Age start follow‐up (1995)					0.993
18–29 years	639	36.5	3193	36.3
30–40 years	751	42.8	3778	43.0
> 40 years	363	20.7	1817	20.7
Region					0.976
East	818	46.6	4102	46.7
West	935	53.4	4687	53.3
Paritya, b					0.879
Nulliparous	216	14.9	1087	15.1
Multiparous	1229	85.1	6109	84.9
Origin					0.001
Danish	1557	88.8	8023	91.4
Non‐Danish	196	11.2	760	8.6
Marital statusa					0.002
Married	1225	70.0	5552	66.3
Not married	524	30.0	2825	33.7
Educational attainmenta, c					< 0.001
Basic	334	19.3	1994	24.0
Low	130	7.5	481	5.8
Middle	1145	66.1	5333	64.4
High	123	7.1	472	5.7
Occupational status a					0.020
Employed	1457	83.1	6856	80.7
Not employed	296	16.9	1636	19.3
Income levela					0.001
1st quartile (lowest)	379	21.6	2192	25.8
2nd quartile	439	25.0	2121	25.0
3rd quartile	459	26.2	2103	24.8
4th quartile	476	27.2	2076	24.4

^a Missing values: cases of SAT: n = 10; parity: n = 437; origin: n = 5; marital status: n = 415; educational attainment: n = 529; occupational status: n = 296; income level: n = 296.

^b Males not included.

^c Educational attainment was characterised as: ‘basic’ (primary/secondary education only), ‘low’ (basic education and vocational education), ‘middle’ (short‐ or medium‐cycle higher education) and ‘high’ (long‐cycle higher education).

The overall incidence of SAT doubled throughout the study period (Figure 1A). The nationwide incidence increased from 3.3/100,000 individuals (95% CI: 2.5–4.4) in 1995 (pre‐fortification) to 6.2/100,000 individuals (95% CI: 5.1–7.6) in 2016 (post‐fortification) (Figure 1A). When adjusting for sex and age, the incidence of SAT annually increased by an estimated 3.2% (95% CI: 2.2–4.2; p < 0.001) during the study period. Incidence rates were similar across regions (1995: East 2.8 (95% CI; 1.7–3.8) and West 3.8 (95% CI; 2.3–5.3)/100,000 individuals; 2016: East 5.3 (95% CI; 3.6–6.9) and West 7.2 (95% CI; 5.3–9.1)/100,000 individuals), a difference not statistically significant (p = 0.18) (Figure 1B). Based on the date of referral, SAT occurred more frequently during summer (incidence rate ratio (IRR): 1.21 (95% CI; 1.07–1.40), p = 0.004) and autumn (IRR: 1.31 (95% CI; 1.15–1.50), p < 0.001), but not during winter (IRR: 1.02 (95% CI; 0.89–1.18), p = 0.761), compared with spring (Figure 2).

Figure 1.

Annual incidence of subacute thyroiditis in Denmark per 100,000 individuals when evaluated overall (A) and stratified by geographical region (B). The grey areas indicate 95% confidence intervals, and in the stratified analyses (B), the dashed line with light grey area is East Denmark, while the solid line with dark grey area is West Denmark.

Figure 2.

Seasonality of subacute thyroiditis. Asterisks indicate p < 0.05 with spring as the reference season.

When evaluating treatment trends, 459 individuals diagnosed with SAT (26.0%) were treated with L‐T4 within 4 years of follow‐up from the date of diagnosis (Table 3). Thyroid hormone replacement therapy was initiated quickly, as half of the treated individuals received treatment within 1.6 months following the SAT diagnosis (IQR: 0.8–3.2 months). Long‐term treatment was frequent, as 224 of the treated individuals (48.7%) remained on L‐T4 beyond 36 months past the diagnosis date. There was no significant difference in the share of individuals treated when comparing the different periods (p = 0.88) (Table 3). Treatment with anti‐inflammatory drugs was also assessed. 723 individuals (41.0%) received anti‐inflammatory treatment. Of these, 382 (21.7%) received corticosteroid monotherapy, 203 (21.7%) NSAID monotherapy and 138 (7.8%) combined treatment with both corticosteroids and NSAIDs. Corticosteroid therapy was most often initiated after the SAT diagnosis (n = 406, 78.1%), most frequently within the first week (n = 291, 71.7%). In contrast, NSAIDs were more commonly initiated prior to the diagnosis (n = 205, 60.1%). Long‐term treatment with L‐T4 (> 36 months) occurred most frequently among individuals receiving NSAID monotherapy (n = 33, 61.1%), compared to those receiving corticosteroid monotherapy (n = 40, 48.8%), combined therapy (n = 16, 50.0%), or no anti‐inflammatory treatment (n = 135, 46.4%).

Table 3.

Levothyroxine (L‐T4) treatment during 4‐year follow‐up following subacute thyroiditis (SAT) stratified by diagnosis period.

	Total	1995–2000	2001–2006	2007–2011	2012–2016
SAT diagnoses, n	1763	360	506	378	519
L‐T4 treated, n (%)a	459 (26.0)	76 (21.1)	93 (18.4)	148 (39.2)	142 (27.3)
Treatment duration, n (%)
0–3 months	83 (18.1)	11 (14.5)	21 (22.6)	29 (19.6)	22 (15.5)
3–12 months	84 (18.3)	8 (10.5)	17 (18.3)	34 (23.0)	25 (17.6)
12–24 months	35 (7.6)	6 (7.9)	6 (6.4)	12 (8.1)	11 (7.8)
24–36 months	33 (7.2)	4 (5.3)	7 (7.5)	13 (8.8)	9 (6.3)
36–48 months	224 (48.8)	47 (61.8)	42 (45.2)	60 (40.5)	75 (52.8)

^a L‐T4 treatment was initiated later than 4 years post‐diagnosis in 54 individuals, thus not presented in the table.

The longitudinal biochemical results from cases of SAT in the North Denmark Region highlighted a biphasic response in thyroid function (Table 4). The thyrotoxic phase was most prominent in the first month following the date of diagnosis, whereas hypofunction of the thyroid gland was most prominent after 2–4 months upon normalisation of thyroid function shortly after (Table 4). Inflammatory parameters were elevated only during the early thyrotoxic phase and until around 1‐month post‐diagnosis (Table 4). The frequency of having elevated thyroid autoantibodies during follow‐up differed by antibody type, with TRAb being elevated in 2.9% of cases, TPO‐Ab in 23.6% and Tg‐Ab in 44.8%.

Table 4.

Biochemical assessment of thyroid function and inflammatory parameters at selected time points relative to the date of diagnosis of subacute thyroiditis (SAT) for individuals residing in the North Denmark Region.

	TSH (mIU/L)		TT3 (nmol/L)		TT4 (nmol/L)		CRP (mg/L)		ESR (mm/h)
Months from SAT	n	Median [IQR]	n	Median [IQR]	n	Median [IQR]	n	Median [IQR]	n	Median [IQR]
Before diagnosis
> 3	55	1.7 [1.2–2.5]	18	1.8 [1.4–2.1]	17	106 [93–116]	53	5.0 [5.0–5.0]	6	5.5 [5.0–9.0]
0–3	30	0.1 [0.01–1.5]	25	2.8 [2.3–3.6]	25	163 [123–205]	30	23 [5.0–44]	4	53 [28–60]
After diagnosis
0–1	56	0.02 [0.01–0.7]	59	2.8 [2.0–3.6]	60	170 [119–215]	46	25 [5.0–77]	26	67 [38–86]
1–2	52	1.7 [0.01–8.0]	57	1.6 [1.2–2.1]	57	87 [61–120]	24	5.0 [4.0–6.2]	16	19 [7.5–28]
2–4	51	5.1 [2.6–8.5]	57	1.5 [1.4–1.8]	56	76 [63–88]	19	5.0 [1.5–5.0]	10	6.5 [5.0–36]
4–6	46	3.5 [2.4–5.7]	48	1.6 [1.4–1.8]	48	84 [75–97]	10	5.0 [1.4–5.0]	5	7.0 [7.0–8.0]
6–12	53	2.9 [2.1–4.1]	47	1.6 [1.4–1.8]	46	89 [77–101]	29	5.0 [1.1–5.0]	5	7.0 [7.0–11]
12–24	41	2.6 [1.8–4.2]	33	1.7 [1.4–1.8]	32	89 [84–104]	25	2.6 [1.0–5.0]	5	6.0 [5.0–10]
24–36	43	2.5 [1.7–4.2]	33	1.5 [1.4–1.8]	31	86 [77–99]	26	1.5 [0.7–5.0]	NA	NA
36–48	31	2.9 [1.9–3.9]	12	1.6 [1.4–1.9]	14	86 [78–96]	21	2.6 [0.8–5.0]	6	12 [8.0–14]

Note: Biochemical data were available for 82 individuals. Data were excluded if obtained following the initiation of thyroid hormone replacement therapy. All values ≥ 10 are reported without decimals.

Abbreviations: CRP, C‐reactive protein; ESR, erythrocyte sedimentation rate; IQR, interquartile range; NA, not available; TSH, thyroid‐stimulating hormone; TT3, total triiodothyronine; TT4, total thyroxine.

4. Discussion

In this nationwide cohort study, we identified cases of SAT diagnosed in Danish hospitals over a 22‐year period within a population of 1,747,168 individuals. The register‐based design ensured a large cohort for studying a relatively rare disease and allowed us to examine disease occurrence, patient characteristics and treatment trends across decades. The incidence of SAT doubled during the study period, but no geographical difference was observed according to regional differences in population iodine intake. A marked proportion of patients with SAT initiated L‐T4 therapy during follow‐up, and half of these remained treated beyond 36 months.

SAT often presents with non‐specific symptoms in the early stage, making it a diagnostic challenge [8]. This can result in misdiagnosis of the subtype of thyroid disease and potentially delay relevant treatment [14]. Previous work showed that the positive predictive value (PPV) of SAT in the DNHR was 62% [9]. To extend the cohort in a nationwide register‐based design, we applied the diagnostic characteristics identified in the previous validation study [9] to define exclusion criteria and thereby increase the specificity of the identified SAT cases. Access to individual patient records was not available; therefore, information on clinical presentation and diagnostic imaging, such as thyroid ultrasound or scintigraphy, could not be assessed, limiting further refinement of the specificity of our case identification.

The autoimmune thyroid disorders of Graves’ disease and Hashimoto's thyroiditis may also be misdiagnosed as SAT [9]. To support the validity of the SAT case identification, we assessed biochemical parameters in the subgroup of cases diagnosed in the North Denmark Region to evaluate if the identified cases demonstrated the expected biochemical characteristics of SAT. The subgroup size was therefore deemed sufficient. We observed the characteristic biphasic response of thyroid function as well as detectable TRAb, TPO‐Ab and Tg‐Ab in ~3%, ~24% and ~45% of patients with SAT, respectively. These figures are overall consistent with previous reports on patients with SAT [2, 18], although a recent meta‐analysis reported lower prevalences [19]. The comparison of figures on antibody‐positivity across cohorts is subject to some uncertainty because differences may exist in the applied assay cut‐off, the assay sensitivity, or the time of antibody measurement according to the diagnosis of SAT. The frequency of TPO‐Ab observed in this study is comparable to that of the general population of Denmark [20], whereas the frequency of elevated Tg‐Ab seemed higher, even when considering assay differences. Overall, the biochemical findings in our subgroup, support the validity of the diagnosis of SAT in our register‐based study. In addition, we observed the characteristic seasonal pattern of SAT [21], which further supports the validity of our case identification.

Monitoring of thyroid diseases is crucial following the implementation of iodine fortification programmes. The DanThyr study reported an increase in SAT incidence from 1.8/100,000/year in 1997–1999, at a time prior to the introduction of mandatory iodine fortification in Denmark, which doubled to 3.7/100,000/year in 2014–2016 [22]. The cases of SAT in the DanThyr study were identified from biochemical test results; thus, this design would also capture cases of SAT managed in general practice without referral to the endocrine hospital department. In contrast, another Danish study assessed the occurrence of hospital‐diagnosed SAT in the North Denmark Region, reporting an incidence of 2.4/100,000/year from 2016 to 2018 [9]. In our cohort, the incidence of SAT showed an increase over time parallel to findings in the DanThyr study, but the incidence of SAT was higher in our cohort than previously reported. This disparity likely reflects differences in the selection of the study cohort in DanThyr and in our study. DanThyr relied on a consecutive laboratory‐based identification of all new cases with abnormal thyroid function in two regions of Denmark, followed by a review of the cases for subtyping the cause of thyroid disease. Our study was register‐based and nationwide; however, the study population was identified via the Danish Medical Birth Register and was conditioned on women who had given birth in Denmark and their male partners. The distributions of sex and age were skewed compared with the general Danish population. The sex imbalance reflects the design of the cohort, where women contributed to the population at risk from 1973 to 2016, whereas men contributed only from 1990 to 2016. Thus, the overall incidence of SAT in our report should be interpreted with caution. However, we applied the same selection criteria throughout the study period, and the trend over time is a key finding that was significant, even after adjusting for age and sex. Finally, the slight and non‐significant differences in incidence by geographical region suggest that the introduction of iodine fortification did not influence the occurrence of SAT based on previous iodine nutritional status, in contrast to findings for other subtypes of thyroid disease in Denmark [23, 24, 25, 26].

Lower SES is associated with higher rates of morbidity and mortality [27], and a link between SES and thyroid disease has been proposed [28, 29]. We found that SAT was more common among individuals with higher SES. SES in patients with SAT has not been previously reported, and we can only speculate about the underlying mechanisms. It should be noted that all cases of SAT were referred from general practice to the hospital for management. In Denmark, universal health care coverage exists, and health care services are equally utilised across socioeconomic groups [27, 30]. Still, referral bias is possible, and it may be speculated that individuals with higher SES are more likely to be referred for hospital management. Thus, follow‐up on these findings is recommended.

Another focus of our study was the treatment with L‐T4 in patients with SAT and the risk of permanent hypothyroidism. It is a characteristic of SAT that thyroid function often recovers and that attempts should be made to withdraw L‐T4 treatment during follow‐up [31]. In the previous smaller and regional Danish study of patients with SAT, a concern was raised that such treatment withdrawal is not always attempted in routine clinical care [9]. In line with the previous regional report, we found that L‐T4 treatment was used in one of four patients and that long‐term treatment was commonly seen. Only one in three discontinued the treatment within 12 months, while almost half of the individuals were still being treated beyond 36 months, corresponding to long‐term treatment in ~13% of all patients with SAT. This aligns with general guideline figures stating that 5%–15% develop permanent hypothyroidism [31], and a recent meta‐analysis reporting a prevalence of ~12% [19]. Our study was register‐based and retrospective in design, relying on secondary research data and reflecting clinical practice. Thus, the figures we report on permanent hypothyroidism were assessed from the timing of redeemed prescriptions of L‐T4 and did not include biochemical assessment of the recovery of thyroid function in each patient. The cases of SAT that we identified were followed in hospital by the endocrinologist for a median of 76 days, meaning that the subsequent management during follow‐up was performed in general practice. Thus, our findings raise a focus on the transition of patients across medical sectors and the level of information provided to general practitioners regarding the need to determine whether L‐T4 can be withdrawn. Finally, it should be emphasised that the conditioning on hospital‐diagnosed patients in our cohort may indicate that these patients had more severe disease, which in turn led to hospital referral. It has been suggested that an initial severe inflammatory and/or thyrotoxic response may be a predictor of permanent hypothyroidism [5, 7, 32]. The prognostic role of thyroid antibodies remains uncertain [19]. We found that long‐term L‐T4 treatment occurred more frequently among individuals receiving NSAID monotherapy compared with the other treatment groups, which is in line with previous findings [33]. However, differences in disease severity across treatment groups is important to consider when interpreting these results, as individuals who did not receive anti‐inflammatory therapy may represent cases with milder SAT and therefore have a lower risk of receiving long‐term L‐T4 treatment. Furthermore, the identification of treatment was from registrations of redeemed prescriptions which may underestimate the use of NSAID which is also available as an over‐the‐counter drug.

5. Conclusion

This large, nationwide study of 1763 patients with SAT, covering two decades, including the implementation of iodine fortification of salt in Denmark, found a doubling in the incidence of SAT, with no geographical differences observed according to population iodine intake. SAT patients were characterised by higher SES, and long‐term L‐T4 treatment was often administered. Further studies across populations are needed to extend findings on patient characteristics, including SES, and to evaluate the long‐term management of patients with SAT.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Research data cannot be shared due to regulatory restrictions that apply to the availability of data generated and analysed during this study to preserve patient confidentiality and according to the GDPR regulations.