Authors: Xueqi Zhang, Guofeng Zhang, Songwen Wang, Jing Jin, Shimiao Zhang, Xiaochun Teng
Categories: Systematic Review, Meta-analysis, Prospective study, Subclinical hypothyroidism, TSH, Thyroid function
Source: BMC Endocrine Disorders
Subclinical hypothyroidism (SCH) is characterized by elevated levels of thyroid hormone (TSH) and normal levels of free thyroxine (FT4). The outcomes of SCH patients are crucial for determining treatment plans; therefore, our aim is to summarize the existing prospective studies to understand the changes in thyroid function over time in SCH patients and the factors influencing these changes, providing references for clinical diagnosis and treatment.
We searched PubMed, Embase, Cochrane Library, and Web of Science for prospective follow-up studies on natural outcomes of SCH published until September 2024. Results are presented as the overall risk ratio (RR) and 95% confidence intervals (CI).
We reviewed 8 prospective follow-up studies involving 1,859 individuals and extracted data from them for a meta-analysis. We found that when TSH levels are ≥ 10 mU/L, patients with SCH are more likely to progress to overt hypothyroidism (OH) (RR11.38, 95%CI 4.98–26.03, P<0.001) and were less likely to return to normal TSH levels (RR 0.20, 95%CI 0.09–0.42, P<0.001) compared to patients with TSH between 4.5 and 9.9 mU/L. In addition, patients who test positive for thyroid peroxidase antibodies (TPOAb) are more likely to progress to OH (RR 2.53, 95% CI 1.86–3.44, P < 0.001) and less likely to return to euthyroidism (RR 0.68, 95% CI 0.60–0.76, P < 0.001) compared to TPOAb-negative patients.
The results indicated that a large proportion of patients diagnosed with SCH will return to normal TSH levels or maintain SCH. Additionally, patients with TSH levels ≥ 10 mU/L or positive for TPOAb are more likely to experience progression and should be closely monitored. However, we did not find any gender differences in the natural outcome of SCH.
The online version contains supplementary material available at 10.1186/s12902-024-01754-7.
Keywords: Subclinical hypothyroidism, TSH, Thyroid function, Meta-analysis, Prospective study
Subclinical hypothyroidism (SCH) is a form of thyroid dysfunction where the serum levels of thyroxine (T4) and/or triiodothyronine (T3) are within the reference range, while the level of thyroid-stimulating hormone (TSH) is slightly elevated [1]. The etiology of SCH is complex and can be induced by autoimmune diseases, surgery, radioactive iodine therapy, iodine excess, and medication [2]. Since SCH is often asymptomatic, it is frequently discovered incidentally during routine blood tests or other disease screenings. However, due to significant fluctuations in TSH levels, it usually requires 2–3 months to repeat measurements of TSH and free thyroxine (FT4) to confirm the diagnosis [3]. Due to the lack of specific clinical manifestations of this disease, a considerable number of patients can maintain good physical condition even without treatment, and only a small number of SCH patients may progress to overt hypothyroidism (OH). Even if it does not progress to OH, many research suggested that the disease is also related to cardiovascular symptoms [4–6], dyslipidemia [7], cognitive impairment [8] and depression [9, 10]. For most patients, we cannot currently determine the outcome of their disease. Therefore, it is essential to identify risk factors for disease progression and implement specific management strategies for patients with different levels of risk [11, 12].
According to previous studies, patients with SCH mainly exhibit three return to euthyroidism (EU), persistent SCH, and progression to OH or experiencing significant hypothyroidism-related symptoms [13–15]. Therefore, we attempt to elucidate the occurrence of three different outcomes in SCH patients who have not been treated by L-T4 and summarize the research data in the original literature to stratify the patients as much as possible to clarify the impact of age, TSH level, thyroid peroxidase antibody (TPOAb), and gender on the prognosis of patients with SCH.
We conducted the comprehensive literature search in PubMed, Embase, the Cochrane Library, and Web of Science databases up to September 2024. The search utilized terms such as “subclinical hypothyroidism”, “outcomes”, and “follow-up studies”. We initially screened article titles, followed by reviewing abstracts, and subsequently performed a detailed examination of the full texts. In the end, we selected relevant articles and restricted the included studies to those published in English. Two researchers independently chose the articles and meticulously read and analyzed both the abstracts and the full texts.
We searched PubMed, Embase, Cochrane Library, and Web of Science, removed duplicate articles, and further excluded meta-analyses, review articles, and conference papers. After screening the titles and abstracts, we thoroughly read the articles related to the topic and performed a meta-analysis on the research studies that were ultimately included. The literature screening process is presented in Fig. 1.
Fig. 1 Flow chart of the inclusion or exclusion procedure
Studies were considered eligible if they met the following (1) The study subjects were individuals diagnosed with SCH who had FT3 and FT4 levels within the normal range but TSH levels above the upper limit of the normal reference range. (2) A prospective follow-up study was conducted. (3) None of the patients participating in the follow-up have taken levothyroxine for treatment. (4) The number of people with different outcomes can be found in the articles. In contrast, studies were excluded if they met one of the following (1) The article was a case report, review, editorial, letter, or animal model subjects. (2) Missing the number of patients with each different outcome and basic information on study characteristics. (3) Duplicate reports of the same original data. The selection process was completed by two authors independently, and differences were resolved through consultation or discussion with the third author.
The following data were (1) basic information (first authors, year of publication, study period, and location); (2) participant characteristics (sample size, age, and the number of patients with different outcomes). Taking into consideration various aspects such as research design, sample size, statistical methods, result consistency, publication bias, health status, and quality assessment tools, we ensured that the obtained results of the meta-analysis are highly reliable and effective. We evaluated the quality of the included studies based on the Newcastle-Ottawa Quality Assessment Scale, which includes criteria for the selection of case and control groups, comparability, and exposure. The total score ranges from 7 to 9, with high and low scores determining the overall quality of the study. Any discrepancies were resolved through consultation with the third author.
The Review Manager 5 software (Copenhagen: Nordic Cochrane Centre, Cochrane Collaboration) was employed for all analytical procedures. Rate ratios (RR) and their corresponding 95% confidence intervals (95% CI) were calculated using this software. The significance of the combined odds ratio (OR) was determined through Z-tests, where a P-value < 0.05 was considered significant. We utilized the Cochran Q test (P < 0.05 indicating statistical significance) and I2 statistic to assess heterogeneity. I2 values of 25%, 50%, and 75% correspond to low, moderate, and high heterogeneity, respectively. When heterogeneity is high (I2 > 50%), a random-effects model was employed to pool the results. Additionally, if heterogeneity is low, a fixed-effects model was used. To assess the stability of the results, sensitivity analysis was conducted by systematically omitting one study at a time and recalculating the summary estimates for the remaining studies.
By searching PubMed (n = 41), Embase (n = 77), Cochrane Library (n = 309), and Web of Science (n = 480), we screened a total of 907 original studies. After the initial literature screening, we reviewed the titles and abstracts of 605 articles, excluding 566 irrelevant studies, and thoroughly read the full text of 39 articles. The flowchart illustrates the entire screening process (Fig. 1). After all these exclusion steps, we used 8 articles to summarize the characteristics of SCH, extracted data from them, and conducted a meta-analysis. Two of them were grouped based on TSH levels at the time of initial prevention, to determine the predictive significance for prognosis [16, 17]. Eight of them examined the impact of TPOAb on prognosis [16–23]. In addition, six of these articles explored the impact of gender on prognosis [16–18, 20, 22, 23] (Table 1).
We selected two prospective studies that provided patient TSH levels and investigated their impact on prognosis [16, 17]. The results suggested that patients with TSH ≥ 10mU/L were more likely to progress to OH (RR 11.38, 95%CI 4.98–26.03, P<0.001) and less likely to return to euthyroidism (RR 0.20, 95%CI 0.09–0.42, P<0.001) while the group of persistent SCH is not significantly different from patients with TSH between 4.5 and 9.9 mU/L (RR 0.92, 95%CI 0.36–2.36, P = 0.870)(Fig. 2).
Fig. 2 Natural outcomes of SCH patients with different TSH levels
Eight studies investigated the impact of TPOAb on the progression of SCH to various outcomes [16–23]. This indicates that patients who are TPOAb-positive are more likely to progress to OH (RR 2.53, 95%CI 1.86–3.44, P<0.001). Such patients are also more likely to maintain SCH status (RR 1.16, 95%CI 1.04–1.28, P = 0.006) and less likely to reclassified to EU (RR 0.68, 95%CI 0.60–0.76, P<0.001)(Fig. 3). This result suggests that SCH patients who are TPOAb-positive are less likely to achieve recovery.
Fig. 3 Natural outcomes of SCH patients with positive and negative TPOAb
Six studies analyzed the impact of gender on the outcomes of SCH [16–18, 20, 22, 23]. We found that gender had no significant impact on the three outcomes of the progression to OH (RR 0.98, 95%CI 0.64–1.50, P = 0.920), persistent SCH (RR 1.03, 95%CI 0.91–1.16, P = 0.620), and return to EU (RR 0.97, 95%CI 0.85–1.11, P = 0.650) (Fig. 4).
Fig. 4 Natural outcomes of SCH patients of different genders
To assess the stability of the results, we performed a sensitivity analysis. Each article was successively excluded and a meta-analysis was conducted on the remaining literature, and we observed that the results did not change significantly, which means that the results were stable and reasonable. We created a funnel plot for the 8 included studies, which can be found in the supplementary materials (Supplementary Fig. 1).
Overall, in this systematic review of 8 studies involving 1,859 participants, many patients maintained persistent SCH (45.4%) or returned to EU (44.4%) during the study period, while 10.1% of patients progressed to OH. This indicates that the natural outcome of SCH in the general population is relatively favorable and further underscores the importance of distinguishing high-risk individuals for targeted treatment.
In 1995, the Whickham survey recorded the prevalence of thyroid diseases in adult populations in the community, suggesting that elevated TSH and positive antibodies may be risk factors for the progression of hypothyroidism [24]. Another 13-year longitudinal study also confirmed that TSH levels combined with thyroid antibodies can provide a clinically useful assessment of the long-term risk of hypothyroidism [25]. However, considering the substantial variability in TSH levels, the controversy surrounding the use of TSH measurements for diagnosing SCH remains a significant concern. A five-year network follow-up study was conducted in 2007 and found that over 50% of patients with elevated or decreased serum TSH levels exhibited normal levels during repeated measurements [26]. Our results verified that patients with TSH ≥ 10mU/L were more prone to disease progression. In addition, compared to TPOAb-negative patients, TPOAb-positive patients were more likely to progress to OH. Therefore, in TPOAb-negative patients with TSH ranging from 4.5 to 9.9mU/L, monitoring may not be as frequent due to the low rate of progression to OH. However, we found no association between gender and SCH outcomes, which is consistent with some studies, but further validation is needed in future research [18]. To enhance the precision of patient management, it is recommended to engage in repeated monitoring of TSH levels before initiating treatment. Additionally, since TSH levels are influenced by diurnal rhythms, peaking between 11 PM and 4 AM and subsequently reaching low and stable serum levels after 11 in the morning, future follow-up studies should strive to ensure consistency in TSH measurement times as much as possible [27]. Furthermore, to enhance the accuracy of research in this field, it is crucial to emphasize the changes in FT4 levels during disease progression. Among the studies we included, only a few mentioned this aspect, and due to data inconsistencies, we were unable to perform a categorized analysis. This stands as a significant limitation of our article. We eagerly anticipate that future research endeavors will prioritize the assessment of FT4 variations alongside TSH, thereby contributing new evidence to this area.
Various previous studies have implied that the incidence rate of SCH increased with age. According to the cross-sectional National Health and Nutrition Examination Survey (NHANES) III, the average TSH of adults aged 80 is 69% higher than that of adults aged 20 [28, 29]. Data from Scotland, and Australia also confirm that TSH level increases with age [11, 30, 31]. Our previous epidemiological survey in ten cities indicated that the prevalence of SCH in people over 65 years old was 19.87% according to the current TSH reference interval, but when the age-specific TSH reference range was used, the prevalence decreased to 3.3% [32]. Although the mechanism is unclear, the increase in TSH in the elderly may be a normal phenomenon associated with aging. However, the data provided by the original studies in this meta-analysis is limited, and these studies have significant differences in age grouping, making it difficult to determine the differences in natural outcomes between elderly and young patients, which is one of the shortcomings of our meta-analysis. However, the relationship between age and the natural outcome of SCH is a very meaningful clinical issue, and more reliable prospective studies are still needed to provide evidence in the future.
Due to the current scarcity of prospective studies and limited data, we were unable to perform a meta-analysis of more articles on the impact of TSH levels on the prognosis of SCH. Furthermore, due to the inconsistency in follow-up periods of the literature we reviewed, we are unable to summarize the outcomes of SCH patients at different time points. Additionally, a shorter follow-up period better highlights the significance of repeated measurements for this condition. We hope that more prospective studies will be conducted in the future to yield additional conclusions. The other limitations of this study are that due to the limitations of the original study, we are unable to summarize the relevant symptoms exhibited during the natural progression of SCH, such as fatigue, depression [33], cardiovascular disease [34, 35], and dyslipidemia [36]. Therefore, future research needs to make efforts in these fields.
In the overall population, the vast majority of SCH patients can maintain their disease status or return to the normal range of TSH during follow-up. TSH ≥ 10mU/L and TPOAb-positive were risk factors for disease progression in SCH patients, while gender was not related to the natural outcome of SCH patients. When it comes to understanding the change in thyroid function of SCH patients in different age groups, it’s imperative for future research to categorize participants based on their age in order to attain accurate answers.
Below is the link to the electronic supplementary material.
The authors hereby confirm that neither the manuscript nor any part of it has been published or is being considered for publication elsewhere. We acknowledge that all authors participated sufficiently in the work and take public responsibility for its content.
XZ and GZ are the co-first author of this study. XT is the corresponding author supervising this work. XZ and GZ drafted the manuscript. SW , JJ and SZ assisted in the literature review. All authors contributed to the article and approved the submitted version.
This study was supported by the National Key Technologies R&D Program provided by Ministry of Science and Technology of the People’s Republic of China (Project Grant # 2022YFC3602300, Sub-project Grant # 2022YFC3602303).
All the data and materials generated and/or analyzed during this study are included in this published article.
Not applicable.
Not applicable.
Not applicable.
The authors declare that no conflict of interest could be perceived as prejudicing the impartiality of the research reported.
All the data and materials generated and/or analyzed during this study are included in this published article.