Establishment of clinical diagnosis model of Graves’ disease and Hashimoto’s thyroiditis

Graves’ disease (GD) is a common organ-specific autoimmune endocrine disease, with an occurrence of 1.2% in the Chinese population. Autoimmune thyroid disease (AITD) includes two major clinical manifestations: GD and Hashimoto’s thyroiditis HT [1]. The prevalence of AITD is approximately 5% [1, 2], and the clinical features of this disease are mainly hyperthyroidism and hypothyroidism. Although the pathogenesis is not very clear, there is evidence that environmental factors (infection, drugs, smoking, iodine, etc.) could trigger AITD in susceptible individuals [3‐6]. Case reports have shown that HT can progress to GD [7‐10]. In addition, other studies have reported that about 10–15% of patients with GD may have hypothyroidism after anti-thyroid treatment [11], suggesting that GD and HT may exist at the same time. In some cases, GD patients may have lymphocytic thyroiditis, and treatment with antithyroid drugs will lead to drug-induced hypothyroidism. Therefore, for patients with early hyperthyroidism, it is crucial to identify the cause of the disease.

At present, the diagnosis of GD is mainly based on the typical clinical manifestations of hyperthyroidism, diffuse enlargement (lesion) of thyroid B in ultrasound, and positive expression of thyrotropin receptor antibody (TRAb), thyroglobulin antibody (TGAb), and thyroid peroxidase antibody (TPOAb). However, in the early stage of HT, there may also be clinical manifestation of hyperthyroidism, positive TGAb and TPOAb. For example, 70% of GD patients have positive TPOAb and TGAb. Similarly, TRAb is also positive in a few HT patients [12]. Therefore, it is difficult for clinicians to distinguish GD and HT with atypical clinical symptoms and positive antibodies.

Currently, the most valuable laboratory test is the determination of serum thyroid stimulating antibody (TSAb) and thyroid stimulating blocking antibody (TSBAb). It is generally believed that TSAb is dominant in GD patients. When the dominant antibody is TSBAb, the incidence of hypothyroidism is increased [13, 14]. However, the detection methods of TSBAb and TSAb are mainly used for scientific research, and it is not applicable in clinical detection for the diagnosis and subsequent treatment of the GD and HT.

So far, there is still no effective differentiation method. Therefore, for patients with early hyperthyroidism, we hypothesize that a clinical differentiation model could be established based on thyroid function (FT3, FT4, TSH), thyroid antibodies (TGAb, TPOAb, TRAb), RAIU (2 h, 6 h, 24 h) and cytological pathology of fine needle aspiration. Model could help clinicians to quantify the indicators while making an accurate diagnosis, thus providing a reference for the clinical diagnosis and identification of hyperthyroidism to effectively save existing medical resources and reduce the economic burden on patients.

AITD is caused by dysregulation of the thyroid immune system. AITD is an organ-specific autoimmune disease mediated by T lymphocytes [1, 16]. AITD includes two major clinical manifestations: GD and HT, both of which are characterized by infiltration of thyroid parenchymal lymphocytes and elevation of thyroid antibodies. The clinical features of GD and HT are mainly thyrotoxicosis and hypothyroidism, respectively. Epidemiological data suggest that the interaction between genetic susceptibility and the environment is a key factor for the occurrences of GD and HT. However, the mechanism of pathogenesis is still not clear.

It is still difficult to make differential diagnosis of GD and HT. The main clinical manifestation of GD is hyperthyroidism, while the clinical features in the early stage of HT also manifested by hyperthyroidism. At present, clinicians distinguish GD and HT mainly based on clinical characteristics, thyroid antibodies, and RAIU.

Generally, TPOAb and TGAb are mainly used to diagnose HT, while TRAb is mainly used for the diagnosis of GD. About 70% of GD patients have positive TPOAb and TGAb. However, TRAb can also be significantly increased in HT patients [9]. TPOAb, TGAb, TRAb can only provide a reference for the clinical diagnosis of GD and HT. To confirm of the diagnoses, the measurement of TSAb and TSBAb is not clinically feasible despite of their accuracy. Similarly, the RAIU is increased in GD patients. Such increase in HT patients is not as high as GD patients, but there is no clearly defined range and clinicians can only judge based on clinical experience.

Therefore, for the initial onset of GD and HT, it may be difficult to differentiate, and so far there is no clear and objective diagnostic criterion in clinical practice. Without properly and effectively guide clinical treatment, patients might accept the wrong treatment plan, leading to serious adverse consequences. Therefore, in this study, the pathological results were used as the basis for the diagnosis of GD and HT.

After screening, these two groups enter model 1 and 2 respectively. The larger value of each variable in the two models indicated the higher probability to diagnose GD. This is the first time to establish a differential diagnosis model using the clinical features of GD and HT and related laboratory results.

From the perspective of area under the ROC curve of the two models, model 1 and model 2 have high diagnostic value for the identification of GD and HT. It is recommended to use − 0.596 as the cutoff point for diagnosing GD and HT when using Model 1. Similarly, it is recommended to use − 0.113 as the cutoff point to diagnose GD and HT when using model 2. If the AUC is greater than cutoff point, GD should be diagnosed. Conversely, if the AUC is less than the cutoff point, HT should be diagnosed. When the patient’s etiological analysis is not clear, our model can help clinicians make diagnosis based on readily available clinical data. Furthermore, the two models can be used to choose laboratory tests that can be performed (local hospitals usually have the ability to perform detection of thyroid antibodies).

However, there are also some shortcomings in our model. For example, model 1 is currently only applicable to hospitals that can carry out radionuclide scanning, and there are limitations for some local hospitals in China. In addition, due to the limited sample size, we have not yet verified the results for large sample sizes. Future studies will further optimize and verify the model.

In the present study, clinical indexes of GD and HT including FT3, FT4, TGAb, TPOAb, TRAb, 2-h RAIU, 24-h RAIU and gender were used to establish a clinical diagnostic model, which can quantitative assist for the diagnosis of GD and HT. Therefore, it helps clinicians differentiate GD and HT, making it easier and timelier to provide patients with optimal treatments.