Impact of Diabetic Ketoacidosis on Thyroid Function in Patients with Diabetes Mellitus

Background Changes in thyroid function in diabetes patients who developed diabetic ketoacidosis (DKA) still need to be fully elucidated. The aim of this study was to systematically review available data on the relationship between thyroid function and DKA in diabetes patients who developed DKA. Methods Electronic databases (PubMed, EMBASE, Cochrane Library, and China Academic Journal Full-text Database (CNKI)) were searched systematically to search relevant literature before December 2020. The mean ± standard deviation and 95% confidence interval (95% CI) were used for evaluation, and sensitivity analysis was performed. Publication bias was estimated by funnel plot, Egger's test, and Begger's test. Results 29 studies were included in the meta-analysis, and the indicators (T4, T3, FT3, FT4, TSH, T3RU, and rT3) of patients with DKA were compared and analyzed. The results of this study showed that the levels of T4, T3, FT3, FT4, and TSH were decreased and the level of rT3 was increased in patients with DKA. Compared with after treatment, the levels of T4, T3, FT3, and FT4 in patients with DKA were decreased before treatment, while the levels of rT3 were increased, and there was no significant difference in changes of TSH. With the aggravation of DKA, the levels of T4, T3, FT3, and FT4 will further decrease, while the changes of TSH have no statistical difference. Conclusion Thyroid function changed in diabetic patients with DKA. It changed with the severity of DKA. This condition may be transient, preceding further recovery of DKA.


Introduction
Diabetic ketoacidosis (DKA) is an acute life-threatening complication of diabetes. It is not only a sign of acute absolute insulin deficiency in type 1 diabetes mellitus (T1DM) but also increasingly seen in patients with type 2 diabetes mellitus. In patients with diabetes, ketoacidosis is caused by an acute decrease in insulin secretion and action in a severe insulin resistant state [1]. From 2002 to 2010 in the United States, about 30% of adolescents newly diagnosed with T1DM developed DKA [2]. e prevalence of DKA estimated at the onset of type 2 diabetes is quite different. African-American youth in Cincinnati and Arkansas was 41.4% [3] and 16% [4]. Statistics showed that thyroid dysfunction in people with diabetes is 2-3 times higher than people without diabetes [5]. e effect of nonthyroid diseases on thyroid function has been studied in anorexia nervosa, liver disease, kidney disease, and many other diseases [6]. Since the 1970s, it has been reported that acute disease can cause a variety of changes in the levels of thyroid hormones in patients who were not previously diagnosed with intrinsic thyroid disease. ese changes are nonspecific and are related to the severity of the disease [7]. Diabetes can have a definite effect on thyroid function in various ways, leading to changes in the levels of thyroid hormones, including immunological mechanisms, cytokine pathways, and regulatory pathways of the hypothalamic-pituitarythyroid axis [8]. When DKA occurred in patients with diabetes, the changes in thyroid function has received a great deal of attention from researchers. At present, there are limited studies on the changes in levels of thyroid hormone in patients with DKA. DKA and its implication in the thyroid function has not been adequately reviewed. e study aimed to analyze the changes in the levels of thyroid hormones in patients with DKA and the relationship between the changes and the severity of DKA.

Inclusion Criteria.
(1) e article related to patients with DKA; (2) involving the changes of thyroid function indicators in patients with DKA before and after treatment or between the diabetic patients with and without DKA and providing the exact sample size and data on various indicators of thyroid function; and (3) the diagnosis of diabetic ketoacidosis is clear [9]. (1) e data of literature are incomplete and the information is not enough to calculate the statistics of this study; (2) case reports; (3) repeated articles; and (4) studies limited to animals.

Literature Screening.
Two researchers independently screened the literature, extracted data, and cross-checked. If there is a disagreement on the results, they would discuss it together or resolve it by a third senior researcher. In the study, data were extracted from the literature finally included in the meta-analysis using a premade data extraction table. e extracted content included the first author, year of publication, study area, sample size, mean ± standard deviation of thyroid function indicators, inclusion criteria, exclusion criteria, DKA diagnostic cutoff point, the determination method of thyroid hormone, therapeutic approach, and duration of treatment of DKA (Table 1).

Statistical Analysis.
According to the requirements of meta-analysis, the data were sorted out, the database was established, the data were carefully checked, and the standardized mean difference (SMD) and 95% CI were used to quantitatively analyze the measurement data. I2 was used to quantitatively test the heterogeneity among the studies. If I2 ≤ 50%, it was considered that the heterogeneity was not statistically significant, and the fixed effect model was used to analyze; on the contrary, if I2 > 50%, the heterogeneity was considered to be statistically significant, and the random effect model was used to analyze. Sensitivity analysis was performed to ensure the stability of the meta-analysis results. Funnel plot and Egger's test were used to evaluate publication bias, and p < 0.05 was considered as statistically significant, indicating that publication bias was not excluded. e trim-and-fill method was used to estimate the effect of publication bias on the interpretation of the results.

Result
3.1. Literature Search Results. 314 related studies were initially retrieved based on keywords and subject terms, and finally, 29 studies met the predetermined inclusion and exclusion criteria (Figure 1). 17 studies evaluated the changes of thyroid function before and after treatment in patients with DKA, 17 studies evaluated the difference of thyroid function between patients with diabetes with and without DKA, and 3 studies related to the changes of thyroid function with different severities of DKA. e relevant literature was published from 1978 to 2018 (Tables 1-3).  Figure 2). ere was no statistical difference in TSH between patients with diabetes with and without DKA. After sensitivity analysis, the result showed that TSH was significantly different (I2 � 42.6%, p < 0.05, Z � 2.01, p < 0.05, SMD � −0.138, 95% CI: −0.273 to −0.003 Figure 3). erefore, patients with DKA have lower levels of T4, T3, FT4, FT3, and TSH and higher level of rT3. Egger's test (T4, p � 0.861; FT4, p � 0.504; rT3, p � 0.445) showed that there was no obvious publication bias. Further analysis by the cutand-fill method showed that the publication bias (T3, p � 0.043; FT3, p � 0.003; TSH, 0.003) did not affect the      International Journal of Endocrinology estimator. It is more certain that the effect estimates obtained in the meta-analysis are effective. e funnel plot is shown in Figure 4.

Comparison of yroid Function before and after Treatment in Patients with Diabetes and DKA.
14 studies involved the comparison of T4 before and after treatment in patients with DKA, including a total of 640 patients with DkA; 13 studies involved the comparison of T3 before and after treatment, including a total of 623 patients with DkA; 13 studies involved the comparison of FT4 before and after treatment, including a total of 755 patients with DkA; 10 studies involved the comparison of FT3 before and after treatment, including a total of 599 patients with DkA; 15 studies involved the comparison of TSH before and after treatment, including a total of 832 patients with DkA; 5 studies involved the comparison of T3RU before and after treatment, including a total of 148 patients with DkA; and 3 studies involved the comparison of rT3 before and after treatment, including a total of 114 patients with DkA. e results showed that patients with DKA had lower levels of T4, T3, FT4, and FT3 and higher level of rT3 compared with after treatment. e difference  showed that there was no obvious publication bias, further analysis by the trim-and-fill method showed that the publication bias (T3, p � 0.006) did not affect the estimator, and it was more certain that the effect estimation obtained in the meta-analysis was effective. e funnel plot is shown in Figure 6.

Comparison of Severity of DKA and yroid Function in Patients with Diabetes and DKA.
ree studies involved the comparison of the severity of DKA with thyroid function. e results showed that as the degree of DKA aggravated, the levels of T4, T3, FT4, and FT3 further decreased. e level of TSH increased with the aggravation of DKA, but it was not statistically significant (Figure 7).

Discussion
is meta-analysis study showed that the levels of T4, T3, FT3, FT4, and TSH were lower and the level of rT3 was higher in patients with DKA compared with patients with diabetes but not DKA. e levels of T4, T3, FT3, and FT4 were lower and the level of rT3 was higher compared with after treatment in patients with diabetes and DKA. As the aggravation of DKA, the levels of T4, T3, FT3, and FT4 would further decrease, but there was no statistical difference in the change of TSH.
DKA can affect the function of the hypothalamus-pituitary-thyroid axis directly or indirectly due to various factors such as relatively insufficient insulin secretion and metabolic disorders, thus affecting thyroid function [38]. Piconi et al. found that large blood glucose fluctuations trigger the production of nitrotyrosine and induce the expression of adhesion molecules and IL-6 [39]. e release of a large number of cytokines acted on the hypothalamuspituitary-thyroid axis through a variety of ways, which can also affect the synthesis, secretion, metabolism, and feedback of thyroid hormones [40]. An increase in cytokines such as IL-6 synchronizing with a low T3 level is often observed which may cause hypothalamus involvement [41]. e body's caloric intake is seriously insufficient in patients with DKA, leading to hypoxia in the cells, which reduced the biological activity of 5′-deiodinase, resulting in a significant    International Journal of Endocrinology reduction in the conversion of T4 to T3, and a significant reduction in the levels and activity of thyroid hormones [42]. Studies have shown that T1DM and thyroid diseases have a common genetic basis [43]. ere is a significant positive correlation between serum TSH and antithyroid antibodies (TRAb, TPOAb, and TGAb) in patients with T2DM, suggesting that abnormal thyroid function in patients with T2DM is autoimmune-mediated pathogenesis [44]. Studies also found that the severity of impaired hypothalamus-hypophysial-thyroid regulation seems to be related to the degree of metabolic disorders regardless of the presence of antithyroid antibodies [45]. Previous studies have shown that the levels of serum T3 and T4 are related to the severity of the disease [46,47]. Similarly, Balsamo et al. showed that changes in hormone levels are usually related to the severity of metabolic disorders, among which thyroid function is one of the most serious disorders. e hypothalamus-pituitary-thyroid axis showed variable damage, which was defined as nonthyroid disease syndrome (NTIS) [45]. e relationship between the degree of NTIS and the severity of metabolic disorders has previously been reported in adults and children [48][49][50][51]. NTIS is now more commonly used to describe a typical change in the serum levels of thyroid-related hormones that may occur after an acute or chronic disease not caused by intrinsic abnormalities in thyroid function. Changes in the hypothalamic-pituitary-thyroid axis also occur in diseases, usually associated with low levels of T3, which gave rise to the term "low T3 syndrome" [52].
It was now well known that most circulating T3 and almost all rT3 came from the peripheral deiodination of T4 [53,54]. Pittman et al. found that DKA played a certain role in the peripheral transformation of T4 [55]. e moderate decrease in serum T4 observed in patients with DKA has been described previously, which was corrected after treatment, and it seemed to be due to acquired deficiency of T4 binding to serum protein [56]. e factors of dietary, especially carbohydrates, played an important role in the regulation of T3 [57,58]. e presence of carbohydrate deprivation in DKA seemed to rapidly inhibit the deiodination of T4 by type 1 iodothyronine-deiodinase in the liver, thereby inhibiting the production of T3 and preventing the metabolism of rT3 [59]. Carbohydrate deprivation will lead to a decrease in basal metabolic rate. e decrease in thyroid hormones is represented the body's remaining adaptive response to calories and protein by inducing hypothyroidism theoretically [60]. It was reported that the average level of rT3 was increased in patients with insulin-dependent diabetes, and the average metabolic clearance rate of rT3 is decreased [55,61]. e result of Pittman CS et al. suggested that T4 monodeiodination of both phenyl rings was significantly impaired in uncontrolled diabetes, and they      believed that long-term insulin insufficiency could lead to a more severe and extensive damage of T4 deiodination [55]. Type 2 deiodinase (Dio2) is an intracellular enzyme that catalyzes the conversion of T4 to T3 [62]. A meta-analysis showed that the polymorphism of Dio2 r92Ala is associated with poor blood glucose control in patients with T2DM [63]. e limitation of this study is that meta-analysis is a secondary literature analysis based on previous research evidence, so there are limitations and bias in the analysis.
e study lacked data for long-term follow-up. e methods used to measure thyroid hormones were much less sensitive than those used in the last decade.

Conclusion
yroid function changed in patients with DKA. It changed with the severity of DKA. is condition may be transient, preceding further recovery of DKA.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.