Persisting symptoms in patients with Hashimoto’s disease despite normal thyroid hormone levels: Does thyroid autoimmunity play a role? A systematic review

Objective Patients with hypothyroidism due to Hashimoto’s disease (HD) may experience persisting symptoms despite normal serum thyroid hormone (TH) levels. Several hypotheses have been postulated to explain these persisting symptoms. We hypothesized that thyroid autoimmunity may play a role. Design A systematic literature review. Methods A PubMed search was performed to find studies investigating the relation between the presence of thyroid autoimmunity and (persisting) symptoms. Included studies were critically appraised by the Newcastle – Ottawa Scale (NOS) and then subdivided into (A) disease-based studies, comparing biochemically euthyroid patients with HD, and euthyroid patients with non-autoimmune hypothyroidism or euthyroid benign goitre, and (B) (general) population-based studies. Due to different outcome measures among all studies, meta-analysis of data could not be performed. Results Thirty out of 1259 articles found in the PubMed search were included in this systematic review. Five out of seven disease-based studies found an association between thyroid autoimmunity and symptoms or lower quality of life (QoL). Sixteen of 23 population-based studies found a comparable positive association. In total, the majority of included studies reported an association between thyroid autoimmunity and persisting symptoms or lower QoL in biochemically euthyroid patients. Conclusion (Thyroid) autoimmunity seems to be associated with persisting symptoms or lower QoL in biochemically euthyroid HD patients. As outcome measures differed among the included studies, we propose the use of similar outcome measures in future studies. To prove causality, a necessary next step is to design and conduct intervention studies, for example immunomodulation vs. placebo preferably in the form of a randomized controlled trial, with symptoms and QoL as main outcomes.


Introduction
Hypothyroidism is defined as lower than optimal thyroid hormone (TH) production by the thyroid gland, resulting in too low or suboptimal plasma TH concentrations [1]. The most frequent cause of hypothyroidism is thyroid dysfunction, also known as primary hypothyroidism [2]. This type of hypothyroidism is characterized by a low or (low-) normal serum free thyroxine (FT4) concentration in combination with a (very) high or elevated thyrotropin (TSH) concentration, and can be a congenital or acquired problem. Worldwide, the most frequent causes of acquired primary hypothyroidism are iodine deficiency and (chronic) autoimmune thyroiditis [3,4].
Autoimmune thyroiditis or Hashimoto's disease (HD) is an autoimmune disorder, in which T-and B cells (slowly) destruct the thyroid gland. A key role in this process seems to be reserved for cytotoxic T cells that are activated by excessively stimulated CD4 positive T cells. Serological markers are anti-thyroid antibodies: anti-thyroid peroxidase and anti-thyroglobulin antibodies (TPO-and Tg-abs, respectively), produced by B cells [5,6].
Despite normalized TSH and FT4 levels by LT4 treatment, approximately five to ten percent of HD patients experience persisting symptoms [12][13][14]. With respect to these persisting symptoms, several hypotheses have been postulated and discussed: 1) TSH is not a perfect marker; consequently, there standard LT4 treatment may not result in a truly biochemically euthyroid state [4]; therefore, some experts suggest to treat with a supraphysiological LT4 dose, which would result in a suppressed TSH, but fewer complaints; however, not all studies show the same results, and higher LT4 doses may increase the risk of cardiovascular disease; 2) a healthy thyroid gland produces approximately 80-90% T4 and 10-20% T3 [8,11,15]; since not all administered LT4 will be converted into active T3, combination therapy of LT4 and LT3 may result in less persisting symptoms; however, until now, it has not been shown that adding LT3 is better than LT4 alone [8]; 3) since deiodinase type 2 (DIO2) facilitates peripheral deiodination of T4 into active T3, patients with DIO2 gene polymorphisms may have variable peripheral T3 availability; in such cases LT4 treatment alone may not be enough [16,17]; with the Thr92Ala DIO2 polymorphism being present in 12-36% of the population [18], this might explain persisting symptoms in a considerable part of affected patients. Yet, none of these three hypotheses about the cause of persisting symptoms in treated patients with HD has been definitely proven. Therefore, according to the American Thyroid Association guideline from 2014, currently LT4-monotherapy is the best choice of treatment in hypothyroidism [8].
In the past years results of several studies have suggested that persisting symptoms in HD patients may be related to autoimmunity [19][20][21]; for example, in a systematic review Siegmann et al. reported a possible correlation between depression and anxiety disorders, and thyroid autoimmunity [22]. While hypothyroidism in HD patients is treated with TH, the autoimmune process affecting the thyroid gland is left untreated. Although, it has been shown that serum TPO-Ab levels decline in most patients with HD who are taking LT4 after a mean of 50 months, TPO-Ab levels became negative in only 16% of the studied patients, illustrating that the majority of patients have persisting elevated TPO-Ab levels [23]. We therefore hypothesized that persisting symptoms in treated patients with HD may be related to autoimmunity. Already in the 1960s [24], it has been recognized that, regardless of thyroid function, thyroid autoimmunity may cause neurological or psychiatric symptoms; in the absence of another obvious cause this clinical picture was called Hashimoto's encephalopathy. The idea that thyroid autoimmunity causes the encephalitis has been abandoned, and is replaced by the hypothesis that these patients suffer from autoimmunity that not only affects the thyroid, but also the brain. Hence the name "Steroid-Responsive Encephalopathy with Autoimmune Thyroiditis" (SREAT). With this in mind, we hypothesized that persisting symptoms encountered in TH treated HD patients also results from autoimmunity affecting the brain. Besides thyroid autoimmunity other latent autoimmune diseases could hypothetically play a role in persisting symptoms in treated HD patients. A recent meta-analysis showed that (latent) poly-autoimmunity is common in patients with an autoimmune thyroid disorder. However, its effect on the course of the persisting symptoms is still unclear [25].
The main objective of this systematic review was to find out whether or not the presence of thyroid autoimmunity is associated with persisting symptoms in HD patients. We performed a literature search in PubMed for original studies investigating the relation between the presence of thyroid autoimmunity and symptoms performed in (LT4 treated) euthyroid patients with hypothyroidism due to HD compared with euthyroid patients with non-autoimmune hypothyroidism or euthyroid benign goitre screened for persisting symptoms, or in general or specific non-HD populations (persons positive or negative for anti-thyroid antibodies, screened for symptoms with specific questionnaires). The "general populations" consisted of either healthy persons, or of patients prone for autoimmune thyroid disease because of already existing other autoimmune disease.

Methods
This systematic review was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 guidelines [26].

Information sources and literature search
For this systematic review the PubMed database was searched for relevant articles. The search was conducted with Mesh and TIAB key terms, using the components Population and Outcome of the PICOstrategy by Glasziou et al.: 'autoimmune hypothyroidism' and 'persisting symptoms', respectively [27]. The following equivalents of these key terms were used: 1) autoantibodies, autoimmunity, autoantigens, autoantibody, antibody, 2) hypothyroidism, 3) thyroglobulin, iodide peroxidase, thyrotropin receptors, TSH receptor, TPO, peroxidase, 4) Hashimoto disease, autoimmune thyroiditis, Hashimoto, Hashimoto's thyroiditis, 5) brain diseases, behavioural symptoms, mental disorders, signs and symptoms, quality of life, brain, fatigue, depression. The equivalent terms were combined as follow: ((1 AND (2 OR 3)) OR 4) AND 5.

Study selection and quality assessment
Title and abstract screening were performed independently by two of the authors (KLG and ASPvT). Full text screening was performed together. Inclusion criteria were original studies investigating the relation between the presence of thyroid autoimmunity and symptoms performed in (LT4 treated) euthyroid patients with hypothyroidism due to HD compared with euthyroid patients with non-autoimmune hypothyroidism or euthyroid benign goitre who were screened for persisting symptoms, or in well-defined (general) populations. Exclusion criteria were: review articles, case reports or series, articles in other languages than English, articles about thyroid antibodies without any relation to thyroid disorders, and articles about HD in relation to other diseases than persisting symptoms or quality of life. Part of the full text screening was a critical appraisal to evaluate the quality of each study following the Newcastle-Ottawa Scale (NOS), a quality assessment form for cohort studies. In this eight-item checklist, we considered thyroid autoimmunity as the 'exposure', and -as explained earlier -the various persisting symptoms as the 'outcome'. Due to the cross-sectional design of many of the included studies, these could not be fully scored on all aspects (e.g., follow-up) of the outcome domain. The original thresholds for good or fair studies were therefore not always applicable. Nevertheless, the NOS scale gave a good indication of the quality of the included studies.
As already mentioned in the introduction, included articles were categorized in either 1) disease-based studies: groups of (LT4 treated) euthyroid patients with hypothyroidism due to HD compared with euthyroid non-autoimmune hypothyroidism patients or euthyroid patients with benign goitre, or 2) (general) population-based studies. In the population-based studies the results of different well-being questionnaires were analysed in relation to the presence or absence of thyroid auto-antibodies in well-defined groups. These studies were subsequently subdivided into 2A) studies in healthy persons, and 2B) studies performed in patients prone for thyroid autoimmunity, and thus prone for poly-autoimmunity, because of already existing other autoimmune disease, e.g., rheumatoid arthritis or celiac disease. The main reasons for making this subdivision were that patients suffering from other autoimmune disease might have had knowledge of having a higher chance of also developing autoimmune thyroid disease, and that symptoms of "other" autoimmune diseases may resemble those of autoimmune thyroid disease/HD. This makes results from these studies related to the research question of this systematic review somewhat more difficult to interpret. Nonetheless, these studies were included if there was independent evaluation of a possible relation between thyroid autoimmunity and (persisting) symptoms or QoL.

Data presentation
Disease-based studies are presented according to the way patients were assessed or to used outcome measure, in the following order: 1) well-being (including all neuropsychological tests, quality of life, fatigue and other mood-parameters) and 2) brain-function (including functional imaging of the brain). Results of the population-based studies are presented in the following order: 1) (truly) healthy persons, 2) individuals recruited from primary care facilities, 3) postpartum women, 4) pregnant women, 5) perimenopausal women, 6) individuals with an already existing other autoimmune disease.
Data are presented as four evidence tables, two for the populationbased studies, and two for the disease-based studies: study and patient characteristics (Tables 1 and 3), and study results (Tables 2 and 4).

Statistical analysis
Due to different outcome measures and presentation of results in all studies, data could not be aggregated. Therefore, meta-analysis was not performed. Instead, a qualitative synthesis of the included data was performed.

Results
The PubMed search was performed on January 10th, 2020 in PubMed and yielded 1259 articles. We excluded 1229 articles through title, abstract or full text screening (n ¼ 1094, n ¼ 105, and n ¼ 30, respectively). A total of 30 articles was included in this systematic review (Fig. 1). Seven articles could be classified as disease-based studies, twenty-three as population-based studies. The results of the critical appraisal are shown in supplementary table 1 and the used PRISMA checklist is shown in supplementary table 2.

Study characteristics
For study characteristics see Table 1 and Table 3. In the disease-based studies sample size ranged from n ¼ 21 to n ¼ 379, and mean age from 35.1 to 54.6 years. In the population-based studies sample size ranged from n ¼ 36 to n ¼ 7,634, and mean age from 17 to 98 years. The included studies could be classified as cross-sectional (n ¼ 17), prospective cohort (n ¼ 8), retrospective cohort (n ¼ 4) or case-control (n ¼ 1).

Disease-based studies
The results of the individual included disease-based studies are summarized in Table 2. Five disease-based studies evaluated well-being in LT4 treated patients with HD compared with patients with nonautoimmune hypothyroidism, using the following questionnaires: DSM-IV mood and anxiety disorders, FIQ, FM comorbidity, General Symptom Questionnaire, MFI-20, QoL, SCID, ThyPRO, VAS (Table 4) [19,[28][29][30][31]. Three studies reported a significant relation between persisting symptoms and thyroid autoimmunity [19,30,31]. In two studies neurocognitive function was investigated with patient and control groups as  described above, using the d2 attention test. [20,21] In one of these studies the results were also related to grey matter density of the left inferior frontal gyrus determined by brain magnetic resonance imaging [21]. Both studies showed a significant association between symptoms and thyroid autoimmunity. Thyroid function was specified in six of the seven disease-based studies. TSH, free T4 and free T3 levels were similar in HD patients and controls (non-autoimmune hypothyroidism and euthyroid benign goitre patients) in four studies [19][20][21]31]. In two studies TSH values were significantly lower in the control groups [28,30], that consisted of patients who underwent thyroidectomy because of differentiated thyroid cancer and who were subsequently treated with LT4 aiming at a suppressed TSH [30]. In one study all HD patients and controls were classified as euthyroid, but thyroid function was not specified [29].
Overall, five of the seven disease-based studies found a statistically significant association between persisting symptoms and the presence of thyroid autoimmunity.
No, except for a significantly worse QoL score on the domain "sex life" QoL pre-operative in ThyPRO domains: Postoperatively, goiter symptoms were significantly worse in the HT patients than in the BG patients (¼0.042). This result is probably due to the fact that an inflammatory process develops around the thyroid gland in HT patients, but not in BG patients. HT N ¼51 Goiter N¼45 P-value A higher percentage of major depression, OCD, PD, GAD, an anxiety disorder, any depressive disorder and any psychiatric disorder was found in the HT group compared with the goiter group. However, differences were not statistically significant. No significant differences in the total score of the d2 attention test were detected between groups (P¼0.9).
Performance in attention testing is associated with GM density LIFG in patients with HT, but not in patients with other thyroid diseases. Particularly low achievement was associated with reduced GM density of this brain region suggesting an influence of autoimmune processes on the frontal cortex in this disease. This could be due to not yet known antibodies affecting brain morphology or an influence of thyroid antibodies themselves.  20 Neurocognitive function assessed by the d2 attention test.

Control group N¼25
Hashimoto's thyroiditis N¼26 P-value No significant differences between groups were detected comparing the main values of the performances in the neuropsychological tests. However, significantly more patients with HT than in controls were found with z-scores below the normal range (less than -1.5) in de d2 attention test regarding total scores.

Yes
Number of patients below the normal range (z-scores).
Overall, 16 of the 23 population-based studies reported a statistically significant association between symptoms and thyroid autoimmunity. However, the total number of people studied in the seven studies showing no association between symptoms and thyroid autoimmunity was much higher (n ¼ 20,769; study sample size range 147-7634) than the number of people in the 16 studies that did show an association (n ¼ 8038; study sample size range 36-1644).

Discussion
In this systematic review we have tried to answer the question whether or not the presence of thyroid autoimmunity is associated with persisting symptoms in HD patients. Twenty-one out of 30 well-designed studies included in the review (70%) reported a probable relation between the presence of thyroid autoimmunity, and (persisting) symptoms or lower QoL. Validity of the studies was evaluated through critical appraisal following the pillars of the NOS. The included studies were divided into studies evaluating LT4 treated patients with hypothyroidism due to HD versus patients with non-autoimmune hypothyroidism (disease-based studies), and (mostly healthy general) population-based studies. An association between the presence of thyroid autoimmunity and (persisting) symptoms was found in five of the seven disease-based studies, and in 16 of the 23 population-based studies. Due to great variety in tests, questionnaires and outcome measures among the studies, data could not be combined nor could a meta-analysis be performed. Yet, to our best knowledge this is the first systematic review on this topic.
In the population-based studies, most participants with and without markers of thyroid autoimmunity -mostly TPO-abs -had a normal thyroid function. Yet, in many of these studies a number of patients had (subclinical) hypothyroidism or hyperthyroidism. Although some studies reported a relation between thyroid function and symptoms [32,34,37,[46][47][48]50], most studies that reported an association between thyroid autoimmunity and symptoms did so after correction for thyroid function. In the disease-based studies, biochemically euthyroid auto-immune hypothyroid patients reported more symptoms than biochemically euthyroid non-autoimmune hypothyroid patients or euthyroid patients with a benign goitre. In most of these studies thyroid function was similar in the HD patients and controls. Although the factor suboptimal thyroid hormone treatment cannot be ruled out, we feel that ongoing (thyroid) autoimmunity may at least play an additional role in the persisting symptoms or lower QoL of these HD patients.
Thyroid autoimmunity as cause of persisting symptoms in treated HD patients has been suggested before. Leyhe et al. reported an association between cognitive and affective disorders, and (euthyroid) autoimmune thyroid disease in their review [55]. The well-designed study by Ott et al., included in this review, showed strong evidence for a relation between thyroid auto-immunity and persisting symptoms [19]. In this study, 426 consecutive euthyroid female patients who underwent thyroid surgery for benign thyroid disease (goitre) were evaluated. Removed thyroid glands were examined for lymphocytic infiltration, and based on histology results and pre-surgery anti-TPO levels an anti-TPO concentration cut-off for "true" inflammation (121.0 IU/mL) was calculated. Subsequently, anti-TPO negative and positive patients were compared with respect to pre-operatively symptoms and QoL. A significant association between anti-TPO levels and chronic fatigue, chronic irritability, chronic nervousness, and lower QoL levels was found.
Results of several other studies are also in support of the hypothesis that there is a relation between thyroid auto-immunity and (persisting) symptoms in euthyroid HD patients. In a cross-sectional study, Watt et al.    31 Percentage of all euthyroid subjects with: Anti-TPO -Anti-TPO + P-value No significant differences were found in well-being or depression between euthyroid TPO-abs positive and TPO-abs negative individuals.
Depression assessed with MDI questionnaire on depression categories: Well-being raw score !50% 85 86 0.4 ISEME 2015 34 Change in CES-D from baseline to followup (5yr)  Higher Tg-ab levels were positively correlated with more errors on: -Trail Making Test Part B (r ¼ 0.470; P ¼ 0.000) -Word Fluency Test (r ¼ 0.284; P ¼ 0.023) -Design Fluency (r ¼ 0.28; P ¼ 0.045) test. The demographic, mood, and neuropsychological test data of all participants with Tg-Ab levels lower than 20mU/L were merged into a low (<20mU/L) Tg-ab group (N ¼ 96) and compared to that of the women whose Tg-ab levels were >20mU/L (high Tg-ab group; N ¼ 29). There were no significant differences between the groups on any of the demographic or mood scores. However, the women in the high Tg-ab group made significantly more perseverative errors on the Design Fluency test (P ¼ 0.003) compared to women in the log Tg-ab group.
The hypothesis that higher levels of Tg-ab would be associated with worse performance on all of the neuropsychological tests was partially supported. Only on the Trails Making Test-Part B, the Design Fluency and the Word Fluency tests, higher levels of Tg-abs were associated with more errors. These findings suggest that higher levels of Tg-abs antibodies are related to poorer performance on tasks of executive functions.

MANOVA
There is some preliminary evidence, that AIT, even without pathologic changes in thyroid hormones, could alter mental wellbeing at least in females. Therefore, AIT could be associated with negative wellbeing independently from the current thyroid function.  33 Association between anti-TPO+, mood and anxiety disorders: OR anti-TPO + vs anti-TPO -P-value (95% CI) Anti-TPO positivity is associated with a higher lifetime risk of a diagnosis of one mood or anxiety disorder.   evaluated health-related QoL, using the ThyPRO, in 199 patients with auto-immune hypothyroidism with TPO-abs levels >60IU/l [56]. No association between QoL scores and thyroid function tests was seen. However, in a multivariate model the TPO-abs level was related to goitre symptoms, depression and anxiety. Half of the studied patients were euthyroid at the time of the study and only 2% were overtly hypothyroid, which may explain that no relation between thyroid function test and QoL was found. Watt el al. concluded that the health-related QoL in patients with auto-immune hypothyroidism was related to TPO-abs level, but not thyroid function. A recent Norwegian randomized trial evaluated the effect of thyroidectomy versus no thyroid surgery on persisting symptoms in adequately LT4 treated HD patients. One hundred and fifty adult euthyroid HD patients with anti-TPO concentrations greater than 1000 IU/mL and persisting symptoms were studied. SF-36 general health score, fatigue score and chronic fatigue frequency improved only in the patients who underwent thyroidectomy [57]. The median serum TPO-abs concentration decreased sharply after thyroidectomy (2232-152 IU/mL). The combination of the decrease in anti-TPO concentrations and the observed clinical improvement suggests that thyroid autoimmunity plays a role in persisting symptoms in euthyroid HD patients. Yet, an important limitation of this study is that it was not blinded by performing a sham operation instead of "no thyroid surgery". Therefore, a placebo effect of thyroid surgery cannot be ruled out. Because these two studies lacked control groups as defined in our methods, they were not included in this review.

Yes
With respect to the population-based studies included in this systematic review, two other studies showed a relation between presence of TPO-abs, and mood and anxiety disorders [35,58]. However, these two studies were not included in this review because the investigated populations consisted of psychiatric patients which were considered to be non-representative study groups. In all included general population studies -also in the seven studies that showed no significant difference between the groups with and without thyroid autoimmunity -, a proportion of the TPO-abs positive participants experienced symptoms, while another part did not. This fits within our hypothesis that thyroid autoantibodies and thus (low-grade) thyroiditis may be present and cause symptoms, even though thyroid function is not yet compromised.
A relation between the presence of thyroid autoimmunity and neurological or psychiatric symptoms has been recognized previously in SREAT, formerly known as "Hashimoto's encephalopathy". In this condition, affected persons have encephalopathy that was attributed to the presence of thyroid autoimmunity [59]. For a long time, TPO-abs crossing the blood-brain barrier were held responsible, but a causative role for these abs has never been proven. It is suggested that both thyroid and brain are targets of autoimmunity, hence the name SREAT, but how is still unknown [60]. SREAT is a rare condition with an estimated prevalence of 2.1/100,000 [61]. Yet, if there is a causal relation between the (thyroid) autoimmunity in persons with TPO-abs or patients with HD on the one hand, and (persisting) symptoms on the other hand -many symptoms may be traced back to (sensing by) the brain, and might be viewed as mild brain dysfunction.
In addition to SREAT or HD patients with (persisting) symptoms, brain autoimmunity may also play a role in the pathogenesis of (late) neurological symptoms in other disorders. For example, it was shown that herpes simplex virus encephalitis may promote the development of neuronal autoantibodies targeting mostly the NMDA receptor causing subsequent autoimmune encephalitis [62]. During the SARS-CoV-2 pandemic neurological symptoms in COVID-19 patients (disabling fatigue, anosmia, Guillain-Barr e syndrome and encephalopathy) can persist or reemerge after clearance of SARS-CoV-2, and are hypothesized to be caused by cross-reactive antibodies generated in response to the primary viral infection [63]. These examples illustrate that a secondary brain autoimmunity may cause neurological symptoms, which is in line with our postulated hypothesis.
A strength of this systematic review is the long period from which articles are included. Although the scientific literature database was also  searched for relevant articles published before 1980, none were found. The most important limitation of this systematic review is that we were not able to compare or aggregate results of studies due to different outcome measurements, and different cut-off values used for TPO-abs. For future research we therefore suggest uniformity in measurements. For example, within the field of rheumatology OMERACT (Outcome Measurements in Rheumatology) is an initiative to increase the standardization of outcomes, with successful results [64]. Since a meta-analysis could not be performed, we could also not correct for differences in sample size; the association between thyroid autoimmunity and symptoms that we found in the population-based studies may therefore be less pronounced, since the studies that showed no association were performed in considerably larger populations. Another limitation of the population-based studies is that thyroid function was not always specified, and that a possible relation between TSH or thyroid hormone levels, and symptoms was not always evaluated or corrected for. Therefore, in these studies minor differences in thyroid function may have contributed to symptoms that were attributed to thyroid auto-immunity alone. Furthermore, studies in populations with another diagnosed autoimmune disease (such as celiac disease and rheumatoid arthritis) are prone for selection bias due the chance of overt and latent poly-autoimmunity. In cases of (latent) poly-autoimmunity, it will be difficult to ascertain a possible relation between thyroid autoimmunity alone and persisting symptoms. However, the latencies for other autoantibodies may hypothetically also play a role in the clinical course of patients with HD. Therefore, it is interesting to evaluate the presence of other autoantibodies in relation to persisting symptoms of HD. Only two of the herein included studies measured other autoantibodies, but did not analyze any relationship between the autoantibodies [31,36]. Six studies reported diabetes and/or rheumatoid arthritis in their population characteristics [52], of which two studies stated that the presence of diabetes mellitus was not a confounder [19,46], and three adjusted their results for the presence of diabetes and/or rheumatoid arthritis [36,38,39]. Future studies are needed to evaluate an additional role of latencies for other autoantibodies in treated HD patients with persisting symptoms. Finally, the results reported in this review may be influenced by publication bias, although we think the chance of publication bias is small as studies both in favour and against our hypothesis were published and subsequently included in this review.

Conclusions
In summary, the majority of the included studies in this systematic review reported an association between thyroid autoimmunity and persisting symptoms or low QoL in patients with HD. Meta-analysis of data was not possible due to the wide variety of used outcome measures. Several possible causes of persisting symptoms in hypothyroid patients have been proposed previously, like LT4 not being the optimal drug for treatment of hypothyroidism and DIO2 gene polymorphisms. However, given the overall results of this systematic review (thyroid) autoimmunity per se may also play a role in persisting symptoms in a part of HD patients. This however, needs further investigation as the found association does not prove a causality. Conducting a randomized placebocontrolled trial, evaluating the effect of additional immunomodulating treatment, for example intravenous immunoglobulins versus placebo, may further elucidate the role of thyroid autoimmunity in persisting symptoms in patients with HD.

Registration and protocols
This review was not registered and no protocol was not prepared.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.