Introduction
Excess weight and obesity are defined as an abnormal or excessive fat accumulation representing a risk to health [1]. They are classified by calculating body mass index (BMI), with overweight classified as a BMI between 25.0 and 29.9 kg/m2 and obese as a BMI ≥ 30 kg/m2. The obese category is subdivided into class I (BMI 30.0 to 34.9 kg/m2), class II (BMI 35.0 to 39.9 kg/m2), and class III (BMI ≥ 40 kg/m2), the latter also referred to as severe or massive obesity [1].
The issue has grown to epidemic proportions both in adults and in children. In the last 40 years, the prevalence of overweight or obese subjects has increased from 4% to 18% worldwide [1]. Many comorbidities are related to this condition, such as type 2 diabetes mellitus, cardiovascular and immune-mediated diseases, and some malignancies [2]. Moreover, more than 4 million people die from obesity complications every year [3].
Thyroid dysfunctions are quite common in the general population. In community surveys, subclinical and overt hypo- and hyperthyroidism have a prevalence ranging from 0.1% to 12.4% and from 0.2% to 10% in adults, respectively [4–9]. The most frequent cause is autoimmune thyroid diseases (AITDs) that are T cell-mediated, organ-specific disorders [10, 11]. Graves’ disease and Hashimoto thyroiditis are the most frequent, both characterised by lymphocytic infiltration, although clinically different [12]. The former is characterized by the presence of thyroid-stimulating hormone receptor antibodies (TRAb) that activate the follicular cell receptor, thereby stimulating thyroid hormone synthesis and secretion [13]. The latter shows antibodies to thyroglobulin (TgAb) and thyroid peroxidase (TPOAb) that are correlated with the active phase of disease leading to hypothyroidism [14–17].
The high prevalence of obesity and thyroid diseases worldwide justifies di per se their simultaneous coexistence. Indeed, in recent decades, there has been a parallel and significant rise in obesity and autoimmune disorders, including thyroid diseases, in industrialised countries, although the underlying mechanisms are complex and not well known [18–20].
For these reasons, we performed a literature review with the aim of assessing the prevalence of thyroid dysfunction, focusing on thyroid autoimmunity with or without hypo- and hyperthyroidism, both in obese children and in adults.
Material and methods
The review was conducted according to the PRISMA statement, and the checklist is reported in Supplementary File.
A PubMed/MEDLINE, Web of Science, and Scopus search was performed for free-text words and terms related to “obesity”, “obese”, “overweight”, “thyroid autoimmunity”, “Hashimoto’s thyroiditis”, “Graves’ disease”, “Graves hyperthyroidism”, “hyperthyroidism”, “hypothyroidism”, “thyroid peroxidase antibody”, “TPOAb”, “thyroglobulin antibody”, “thyroglobulin antibodies”, “TgAb”, “thyroiditis”, and “thyrotoxicosis”. Original studies and reviews in English published online up to 30 April 2023 were selected and reviewed. The final reference list was defined based on the relevance of each paper to the scope of this review.
Three authors extracted data from all the included studies in the full text, tables, and figures concerning general study information (authors, publication year, country, study design, and funding sources) and patient characteristics (i.e. sample size, age, sex ratio, and clinical setting).
The selected method used for assessing the risk of bias in individual studies and the applicability to the review question was QUADAS-2, a tool for evaluating quality in diagnostic test accuracy studies [21]. Three reviewers assessed the studies’ grades in the systematic review in four domains (patient selection, index test, reference standard, and flow and timing) concerning the risk of bias and in three fields regarding the applicability (patient selection, index test, and reference standard).
Results
In the preliminary search, 9742 articles were identified, and 6901 remained after removing duplicates. A total of 458 full-text publications were retrieved after further careful review of abstracts. A total of 143 studies were eligible for full-text and 79 articles were included in the present analysis. A PRISMA flow diagram of the screening and selection process can be found in Figure 1.
Taking advantage of the data reported in each study, the Authors assessed the risk of bias and concerns about the applicability of the included papers based on the QUADAS-2 instruments. The results of the quality assessment are reported in Figure 2.
Post hoc, the studies were divided by the age of the subjects. In detail, 24 studies involved children (10 cross-sectional studies, 7 case controls, and 7 cohort studies) and 55 involved adults (42 cross-sectional studies, 4 case controls, and 9 cohort studies) (Tab. 1 and 2).
Author, date |
Study design |
Total obese patients |
Thyropathy |
Patients affected (%), male/female* |
Outcome |
Stichel et al., 2000 [22] |
Retrospective case-control |
290 |
AITD |
10 (3.4%) |
The prevalence of AITD increased in obese children, mostly in those with elevated TSH |
Subclinical hypothyroidism |
22 (7.6%), 14/8 |
||||
Hypothyroidism |
1 (0.3%) |
||||
Eliakim et al., 2006 [29] |
Prospective cohort |
196 |
Subclinical hypothyroidism |
41 (20.9%), 20/21 |
No beneficial effects on body weight, body mass index, linear growth, and body lipids were found in treated obese subjects, suggesting that thyroid substitution is not necessary in most cases |
Bhowmick et al., 2007 [23] |
Retrospective case-control |
308 |
AITD |
5 (1.6%), 1/4 |
A higher prevalence of TSH elevation was observed in the obese group |
Subclinical hypothyroidism |
36 (11.7%), 14/22 |
||||
Dekelbab et al., 2010 [35] |
Retrospective case-control |
191 |
AITD |
6 (3.1%) |
Mild elevation of TSH values in the absence of AITD is not uncommon in obese children, but no special characteristics was found |
Subclinical hypothyroidism |
20 (10.8%) |
||||
Grandone et al., 2010 [28] |
Prospective cohort |
938 |
Subclinical hypothyroidism |
120 (20.8%), 58/62 |
A moderate elevation of TSH concentrations is frequent in obese children, it is not associated to metabolic risk factors, it is reversible after weight loss, and it should not be treated |
Marras et al., 2010 [27] |
Prospective cohort |
468 |
Subclinical hypothyroidism |
15 (3.2%), 6/9 |
An increased fT3 concentration is the most frequent abnormality. Serum fT3 and TSH correlate with BMI. Moderate weight loss frequently restores abnormalities |
Ong et al., 2012 [134] |
Prospective cohort |
264 |
AITD/ hypothyroidism |
NA |
Childhood weight gain and childhood overweight conferred an increased susceptibility to later hypothyroidism and AITD, particularly in females |
Ittermann et al., 2013 [44] |
Retrospective case-control |
563 |
Hypothyroidism |
29 (5.1%) |
Active and passive smoking may mediate the association between thyroid function and BMI in adolescents |
Hyperthyroidism |
4 (0.7%) |
||||
Marwaha et al., 2013 [24] |
Retrospective cross-sectional |
488 |
Subclinical hypothyroidism |
48 (9.8%), 26/22 |
Serum fT3 and TSH were positively while fT4 was negatively associated with BMI in apparently healthy euthyroid children |
Bouglè et al., 2014 [34] |
Prospective cohort |
528 |
Subclinical hypothyroidism |
69 (13.1%) |
Increased TSH may be predictive of a decrease in insulin resistance, fT4 was associated with a low metabolic risk. Changes in thyroid function could protect against the occurrence of obesity-associated metabolic diseases |
Ghergherehchi et al., 2015 [25] |
Retrospective case-control |
190 |
AITD |
28 (14.7%), 11/17 |
TSH and fT4 levels are increased in obese children, but the incidence of AITD is lower |
Subclinical hypothyroidism |
20 (10.7%) |
||||
Krause et al., 2015 [42] |
Retrospective case-control |
572 |
Subclinical hypothyroidism |
28 (4.9%) |
Paediatric obesity is associated with higher TSH and lower FT4 concentrations and with a greater prevalence of abnormally high TSH |
Matusik et al., 2015 [41] |
Prospective cohort |
NA |
Subclinical hypothyroidism |
51, 20/31 |
In obese children with sHT dietary-behavioural management intervention contributed to reduction of body mass index, irrespective of levothyroxine use. Moderately elevated levels of TSH are a consequence rather than cause of overweight, and pharmacological treatment should be avoided |
Garcia-Garcia et al., 2016 [38] |
Retrospective cross-sectional |
129 |
AITD |
4 (5.6%) |
Obese children and adolescents had higher levels of TSH, and among them the prevalence of AITD is higher |
Subclinical hypothyroidism |
11 (14.4%) |
||||
Dahl et al., 2018 [32] |
Retrospective case-control |
1796 |
Subclinical hypothyroidism |
186 (10.4%), 121/142 |
The prevalence of SH was higher among overweight/obese study participants. TSH and fT4 are positively correlated with WHtR |
Jin, 2018 [40] |
Retrospective cross-sectional |
111 |
Subclinical hypothyroidism |
27 (24.3%) |
SCH is common in obese children; TSH levels were linked with the lipid profile |
Dursun et al., 2019 [31] |
Retrospective cross-sectional |
218 |
AITD |
16 (7.3%), 13/3 |
Obese adolescents with non-autoimmune thyroiditis had a higher incidence of insulin resistance |
Kumar et al., 2019 [45] |
Prospective cohort |
162 |
Hypothyroidism |
51 (31.4%), 17/34 |
The addition of levothyroxine to the weight reduction program did not show beneficial effects on body weight, BMI, and thyroid profiles in obese children with isolated hyperthyrotropinaemia |
Ruszała et al., 2019 [33] |
Retrospective cross-sectional |
100 |
Subclinical hypothyroidism |
20 (20.9%), 7/13 |
Isolated increased TSH level is common in obese adolescents, there is no correlation between TSH, fT3, and fT4 levels and BMI SDS value. Autoimmune thyroiditis in obese adolescents is more common than in the general population. |
Thiagarajan et al., 2019 [39] |
Prospective cross-sectional |
102 |
Hypothyroidism |
15 (14.7%) |
No correlation was found between fT4, TSH, and BMI |
Özcabı et al., 2021 [37] |
Retrospective cross-sectional |
56 |
AITD, TPOAb |
13 (23.2%) |
An association between central adiposity and TgAb levels was found |
AITD, TgAb |
17 (30.4%) |
||||
Patel et al., 2021 [36] |
Retrospective cross-sectional |
404 |
Subclinical hypothyroidism |
122 (30.2%) |
Obesity-related subclinical hypothyroidism predisposes to increased non-alcoholic steatohepatitis independently of severity adiposity |
Dündar et al., 2022 [26] |
Retrospective cross-sectional |
1130 |
Subclinical hypothyroidism |
59 (5.2%), 25/34 |
Subclinical hypothyroidism can negatively affect the lipid and glucose profile in obese children |
Di Sessa et al., 2023 [30] |
Retrospective cross-sectional |
844 |
Subclinical hypothyroidism |
85 (10.1%), 27/58 |
Children with obesity and NAFLD presented increased risk of SCH, and vice versa |
Author, date |
Study design |
Total obese patients |
Thyropathy |
Patients affected (%), male/female* |
Outcome |
Rimm et al., 1975 [63] |
Retrospective cross-sectional |
73532 |
Hypothyroidism |
35950 (48.9%), 0/35950 |
Obesity is related with a higher risk of hypothyroidism and a decreased risk of Graves’ hyperthyroidism |
Hyperthyroidism |
35651 (48.5%), 0/35651 |
||||
Szomstein et al., 2002 [94] |
Retrospective cross-sectional |
195 |
Hypothyroidism |
14 (7.2%) |
Article included for epidemiological data |
Raftopoulos [90] et al., 2004 |
Retrospective cross-sectional |
86 |
Subclinical hypothyroidism |
9 (10.5%) |
Article included for epidemiological data |
Holm et al., 2005 [135] |
Retrospective cohort |
NA |
NA |
68, 0/68 |
Obesity was associated with a decreased risk of Graves’ hyperthyroidism |
Knudsen et al., 2005 [136] |
Retrospective cross-sectional |
NA |
NA |
NA |
Even slightly elevated serum TSH levels were associated with an increase in the occurrence of obesity |
Moulin de Moraes et al., 2005 [50] |
Retrospective cross-sectional |
72 |
Subclinical hypothyroidism |
18 (25.0%), 2/16 |
Article included for epidemiological data |
Chikunguwo et al., 2007 [77] |
Retrospective cross-sectional |
86 |
Subclinical hypothyroidism |
9 (10.5%) |
Article included for epidemiological data |
Asvold et al., 2009 [57] |
Retrospective cross-sectional |
27097 |
Subclinical hypothyroidism |
1804 (6.7%), 451/1353 |
Hypothyroidism is associated with high BMI, both in current smokers and in never-smokers |
Hypothyroidism |
158 (0.6%), 21/137 |
||||
Hyperthyroidism |
574 (2.1%), 141/433 |
||||
Gniuli et al., 2009 [80] |
Retrospective cross-sectional |
45 |
Subclinical hypothyroidism |
11 (24.4%) |
Article included for epidemiological data |
Gopinath et al., 2010 [137] |
Prospective cohort |
NA |
Subclinical hypothyroidism |
1.9% |
Obesity is a risk factor for SCH and overt hypothyroidism |
Hypothyroidism |
6.2% |
||||
Marzullo et al., 2010 [47] |
Retrospective cross-sectional |
165 |
AITD |
18 (10.9%), 3/15 |
Obese patients had lower fT3 levels and fT4 levels, greater prevalence of hypothyroidism, and higher prevalence of TPOAb positivity |
Hypothyroidism |
17 (10.3%), 4/13 |
||||
Somwaru et al., 2011 [138] |
Retrospective cohort |
NA |
Hypothyroidism |
1027 |
Higher BMI was independently associated with greater risk for overt hypothyroidism |
Hemminki et al., 2012 [87] |
Prospective cohort |
29665 |
Hypothyroidism |
83 (0.3%) |
The risk of Hashimoto’s disease/hypothyroidism was significantly increased; a small but significant increase was also noted for the Graves’ disease/hyperthyroidism |
Hyperthyroidism |
54 (0.2%) |
||||
Sundaram et al., 2013 [92] |
Retrospective cross-sectional |
1221 |
Hypothyroidism |
231 (18.9%) |
Article included for epidemiological data |
Agbaht et al., 2014 [70] |
Retrospective cross-sectional |
548 |
AITD |
96 (17.5%) |
Article included for epidemiological data |
Hypothyroidism |
123 (22.4%) |
Article included for epidemiological data |
|||
Ruiz-Tovar et al., 2014 [81] |
Retrospective cross-sectional |
60 |
Subclinical hypothyroidism |
10 (16.7%) |
Article included for epidemiological data |
Han et al., 2015 [58] |
Retrospective cross-sectional |
90 pregnant women |
AITD |
21 (23.3%), 0/21 |
Obesity was associated with increases in the odds of hypothyroidism and TPOAb positivity |
Subclinical hypothyroidism |
7 (7.8%), 0/7 |
||||
Hypothyroidism |
3 (3.3%), 0/3 |
||||
Janssen et al., 2015 [52] |
Retrospective cross-sectional |
503 |
Subclinical hypothyroidism |
61 (12.1%) |
Article included for epidemiological data |
Amouzegar et al., 2016 [139] |
Prospective cohort |
NA |
NA |
NA |
Early diagnosis of SCH and hypothyroidism was significantly associated with obesity |
Bedaiwy et al., 2017 [61] |
Retrospective cross-sectional |
105 |
Subclinical hypothyroidism |
25 (23.8%), 0/25 |
Article included for epidemiological data |
Răcătăianu et al., 2017 [72] |
Retrospective cross-sectional |
82 |
Hypothyroidism |
12 (14.6%) |
Article included for epidemiological data |
Hyperthyroidism |
1 (1.2%) |
Article included for epidemiological data |
|||
AITD |
27 (32.9%) |
Article included for epidemiological data |
|||
Valdes et al., 2017 [75] |
Retrospective cross-sectional |
1193 |
Subclinical hypothyroidism |
41 (3.4%) |
SCH is more prevalent in obese and morbid obese population |
Zendel et al., 2017 [84] |
Retrospective cross-sectional |
1823 |
Hypothyroidism |
93 (5.1%), 8/85 |
Article included for epidemiological data |
Milla Matute et al., 2018 [89] |
Retrospective cross-sectional |
1581 |
Hypothyroidism |
35 (2.2%) |
Article included for epidemiological data |
Mousa et al., 2018 [69] |
Retrospective cross-sectional |
102 |
AITD |
26 (25.5%) |
Article included for epidemiological data |
Ornaghi et al., 2018 [88] |
Retrospective cohort |
98 pregnant women |
Hypothyroidism |
17 (17.3%), 0/17 |
Obese patients with chronic hypertension showed a 2.4-fold increased risk of developing hypothyroidism during pregnancy versus normal BMI women |
Pedro et al., 2018 [83] |
Retrospective cross-sectional |
1449 |
Hypothyroidism |
106 (7.3%) |
Article included for epidemiological data |
Răcătăianu et al., 2018 [72] |
Retrospective cohort study |
158 |
Hypothyroidism |
28 (17.7%) |
Article included for epidemiological data |
Hyperthyroidism |
1 (0.6%) |
Article included for epidemiological data |
|||
AITD |
39 (24.7%) |
Article included for epidemiological data |
|||
Rudnicki et al., 2018 [93] |
Retrospective cross-sectional |
1756 |
Hypothyroidism |
90 (5.1%), 17/73 |
Article included for epidemiological data |
Sami et al., 2018 [56] |
Retrospective cross-sectional |
1193 |
Subclinical hypothyroidism |
19 (15.0%), 6/13 |
Subclinical hypothyroidism is highly prevalent in obese patients |
Wang et al., 2018 [59] |
Retrospective cross-sectional |
310 |
AITD |
33 (10.6%), 15/18 |
Obese females had higher risk of SCH than non-obese females. No association between obesity and hypothyroidism was observed in male participants |
Subclinical hypothyroidism |
37 (11.9%), 14/23 |
||||
Hypothyroidism |
39 (10.6%), 15/24 |
||||
Zhang et al., 2018 [46] |
Retrospective cross-sectional |
534 |
Subclinical hypothyroidism |
108 (20.2%), 0/108 |
Article included for epidemiological data |
Abdelbaki et al., 2019 [53] |
Retrospective cross-sectional |
554 |
Subclinical hypothyroidism |
72 (13.0%), 9/63 |
Article included for epidemiological data |
Almunif et al., 2019 [79] |
Retrospective cross-sectional |
1480 |
Subclinical hypothyroidism |
106 (7.2%) |
Article included for epidemiological data |
Hypothyroidism |
160 (10.8%) |
Article included for epidemiological data |
|||
Dambros Granzotto et al., 2019 [49] |
Retrospective cross-sectional |
215 |
Subclinical hypothyroidism |
20 (9.3%), 6/14 |
Article included for epidemiological data |
Nayak et al., 2019 [62] |
Retrospective cross-sectional |
175 |
Hypothyroidism |
37 (21.1%), 0/37 |
Article included for epidemiological data |
Neves et al, 2019 [78] |
Retrospective cross-sectional |
641
|
Subclinical hypothyroidism |
11 (1.7%) |
Article included for epidemiological data |
SubHyper |
4 (0.6%) |
Article included for epidemiological data |
|||
Xia et al, 2019 [51] |
Retrospective cross-sectional |
101 |
AITD |
12 (11.9%), 7/5 |
Article included for epidemiological data |
Zhu et al., 2019 [48] |
Retrospective cross-sectional |
88 |
Subclinical hypothyroidism |
28 (31.8%), 11/17 |
Article included for epidemiological data |
Khan et al., 2020 [91] |
Retrospective cross-sectional |
883 |
Hypothyroidism |
93 (10.5%), 12/81 |
Article included for epidemiological data |
Makwane et al., 2020 [86] |
Retrospective cross-sectional |
31 |
Hypothyroidism |
6 (19.0%) |
No significant relationship between thyroid hormones and BMI was found in normal or in obese groups |
Hyperthyroidism |
1 (3.0%) |
||||
Wu et al., 2020 [74] |
Retrospective cross-sectional |
NA |
AITD |
1041 |
Obesity was positively correlated with the prevalence of positive thyroid autoantibodies in euthyroid subjects |
Mahdavi et al., 2021 [55] |
Retrospective case-control |
1144 |
AITD |
198 (17.3%), 33/165 |
Higher prevalence of hypothyroidism and TPOAb positivity among obese patients |
Subclinical hypothyroidism |
87 (7.6%), 9/78 |
||||
Hypothyroidism |
46 (4.0%), 5/41 |
||||
Subclinical hyperthyroidism |
31 (2.7%), 11/20 |
||||
Hyperthyroidism |
17 (1.5%), 4/13 |
||||
Yin et al., 2021 [68] |
Retrospective case-control |
280 |
AITD |
80 (28.5%) |
No significant associations between obesity and TGAb levels. Prevalence of SCH was higher in obese patients, rather than overweight and normal ones, whereas there were no differences in prevalence of hypothyroidism, hyperthyroidism, and subclinical hyperthyroidism between the three groups |
Subclinical hypothyroidism |
56 (20,0%) |
||||
Hypothyroidism |
3 (1.0%) |
||||
Subclinical hyperthyroidism |
1 (0.3%) |
||||
Hyperthyroidism |
1 (0.3%) |
||||
Agbaht et al., 2022 [70] |
Retrospective cohort study |
285 |
Hypothyroidism |
109 (38.2%) |
Article included for epidemiological data |
AITD |
64 (22.5%) |
Article included for epidemiological data |
|||
Dewantoro et al., 2022 [82] |
Retrospective cross-sectional |
887 |
Hypothyroidism |
49 (5.5%) |
Article included for epidemiological data |
Fierabracci et al., 2022 [64] |
Retrospective cross-sectional |
749 |
AITD |
135 (18.0%) |
Autoimmune thyroiditis is highly prevalent in patients with obesity. TgAb may be associated with hypothyroidism in the absence of TPOAb |
Subclinical hypothyroidism |
153 (20.4%) |
||||
Hypothyroidism |
104 (13.8%) |
||||
Hyperthyroidism |
10 (1.3%) |
||||
Walczak et al., 2022 [54] |
Retrospective cross-sectional |
181 |
AITD |
57 (31.4%), 6/51 |
BMI is significantly higher in patients with high normal TSH |
Zhao et al., 2022 [140] |
Prospective cross sectional |
764 |
Subclinical hypothyroidism |
66 (8.6%) |
Article included for epidemiological data |
Yan et al., 2022 [66] |
Retrospective cross-sectional |
289 |
AITD |
58 (20.7%) |
No significant associations between obesity and TPOAb levels. Among patients with positive TgAb and TPOAb, SCH prevalence was significantly higher in obese subjects |
Subclinical hypothyroidism |
27 (9.3%) |
||||
Yang et al., 2022 [67] |
Retrospective case-control |
3697 |
AITD |
870 (23.0%), 358/512 |
Obesity was a significant independent risk factor for hypothyroidism in males, whereas in females it was not |
Sharma et al., 2022 [60] |
Retrospective case-control |
15 |
AITD |
7 (46.7%), 0/7 |
Article included for epidemiological data |
Alourfi et al., 2023 [141] |
Retrospective cross sectional |
292 |
Subclinical hypothyroidism |
13 (4.5%) |
Article included for epidemiological data |
Kang et al., 2023 [76] |
Retrospective cross-sectional |
51 |
Subclinical hypothyroidism |
42 (82.4%) |
Article included for epidemiological data |
Thyroid disease in obese children
A total of 9784 obese children were evaluated and enrolled in this review. Among them, 1174 (12%) children (mean age 10.9 ± 1.4 years) showed thyroid autoantibodies and/or hypo-hyperthyroidism (Tab. 1 and 3). Based on the available data, no difference in sex distribution was observed among obese children with (415/343 F/M) or without thyroid dysfunctions (3452/3288, F/M) (p = 0.065) [22, 23], whereas a significant difference in age was identified (10.9 ± 1.2 vs. 11.0 ± 0.4 years, respectively; p < 0.001) [26, 28, 29, 31, 34–36].
Thyropathy |
Patients affected (%) |
Male/female (OR F/M)* |
Children (n = 9784) |
||
Autoimmune thyroiditis |
86 (0.9%) |
1/4 (2.5) |
Subclinical hypothyroidism |
988 (10.1%) |
64/81 (1.4) |
Overt hypothyroidism |
96 (1.0%) |
17/34 |
Hyperthyroidism |
4 (0.1%) |
NA |
Adults (n = 125,958) |
||
Autoimmune thyroiditis |
1741 (1.4%) |
415/794 (1.5) |
Subclinical hypothyroidism |
2752 (2.2%) |
480/1474 (1.5) |
Overt hypothyroidism |
37600 (29.9%) |
45/232 (2.5) |
Subclinical hyperthyroidism |
36 (0.03%) |
11/20 (0.8) |
Overt hyperthyroidism |
36256 (28.8%) |
145/36097 (39.3) |
Autoimmune thyroiditis was found in 86/9784 children (0.9%) [22, 23, 25, 31, 35, 37, 38], whereas thyroid disfunction was found in 1088 (11%) [22–44]. In more detail, 988 (10.1%) children showed subclinical hypothyroidism [22–30, 32–36, 38, 39], 96 (1%) overt hypothyroidism [22, 39, 44, 45], and 4 (0.04%) hyperthyroidism [44].
Data on sex distribution within each thyroid disease category were only available for a few articles, mainly regarding subclinical hypothyroidism [22–30, 32, 33, 41, 45], which showed that 439/788 (55.7%) obese children suffering from this condition were female [22–30, 32, 33, 41]. In addition, the onset of this condition was observed during puberty for 179/788 (23%) children: 113 (63%) experienced it in prepubertal age [26, 28].
Conversely, Dursun et al. showed, in a small number of patients, the presence of anti-peroxidase and/or thyroglobulin antibodies mainly post puberty (81%) rather than in prepubertal age [31].
Regarding the possible differences in the prevalence of thyroid antibodies between obese and non-obese patients, four studies evaluated 2317 subjects [22, 23, 37, 38]. The data revealed a significantly higher prevalence among obese compared to non-obese children (7% vs. 3%, respectively, p < 0.001). Nine studies, for a total of 9667 subjects, investigated the prevalence of subclinical hypothyroidism; again, the prevalence was higher in obese as compared to non-obese children (10% vs. 6%, p < 0.001) [22–25, 27, 32, 35, 38, 40, 42]. Only one study, conducted in a large cohort of children (6165 subjects), investigated the prevalence of overt hypothyroidism and hyperthyroidism. An equal distribution was found between obese and non-obese children (5% vs. 7%, respectively, p = 0.101), whereas hyperthyroidism was associated more with non-obese subjects (1% vs. 3%, respectively, p < 0.001) [44].
Thyroid disease in obese adults
A total of 125,953 obese adults were evaluated and included in this review. Among them, 78,385 (62.2%) patients (43.2 ± 9.1 years old) showed AITDs and/or hypo-hyperthyroidism (Tab. 2 and 3). The available data revealed significant differences in age, BMI, and sex distribution between obese patients with and without thyroid diseases. In detail, those with thyroid diseases were younger (35.3 ± 6.8 vs. 41.0 ± 1.9 years, p < 0.001) [46–53], heavier (BMI: 39.4 ± 6.3 vs. 36.1 ± 2.3 kg/m2, p < 0.001) [46–49, 51, 52], and more frequently female (13% vs. 8%, p < 0.001) [46–51, 53–62]. These last data were obtained excluding the study by Rimm et al., which enrolled only female patients [63].
Autoimmune thyroiditis was found in 1741/125,953 (1.4%) patients [47, 51, 55, 58–60, 64–47] whereas thyroid dysfunction was found in 76,644 (60.9%) patients. In detail, 2752 (2.2%) subjects showed subclinical hypothyroidism [46, 48–50, 52, 53, 55–59, 61, 64, 66, 69, 75–81], 37,600 (29.9%) overt hypothyroidism [47, 55, 57–59, 62–64, 68, 70–73, 79, 82–94], 36 (0.03%) subclinical hyperthyroidism [55, 68, 78], and 36,256 (28.8%) overt hyperthyroidism [55, 57, 63, 64, 68, 72, 73, 86, 87]. All these conditions were significantly associated with females (Fig. 3). The BMI values according to the patients’ thyroid status are reported in Figure 4. In detail, patients with AITDs showed a BMI of 34.6 ± 7.0 kg/m2 [47, 51, 55, 60, 64–68], those with subclinical hypothyroidism had a BMI of 39.3 ± 7.0 kg/m2 [46, 48–50, 52, 53, 55, 56, 66, 68, 80], and those with overt hypothyroidism had a BMI of 40.9 ± 8.2 kg/m2 [47, 55, 64, 68, 71, 82, 84, 86, 88, 90, 91, 93]. Patients affected by subclinical hyperthyroidism showed a BMI of 31.1 ± 0.6 kg/m2 [55, 68] and those affected by overt hyperthyroidism a BMI of 34.1 ± 4.2 kg/m2 [55, 64, 68, 86].
Nine studies, including 74,662 patients, compared the prevalence of thyroid autoantibodies in obese compared to non-obese subjects [47, 55, 58–60, 66, 68, 69, 74]. The prevalence was superimposable: 15% of obese and 14% of non-obese patients showed thyroid autoimmunity (p = 0.178). Eight studies, including a total of 23,032 subjects, investigated the prevalence of subclinical hypothyroidism, again comparing obese and non-obese subjects [55, 58, 59, 61, 66, 66, 68, 75, 76], and again no difference was observed among the two groups (10% vs. 9%, respectively, p = 0.296). By contrast, among 16,605 patients, overt hypothyroidism was significantly more frequent in obese patients (7% vs. 3%, respectively, p < 0.005) [47, 55, 58, 59, 62, 68, 86, 88]. Sub-clinical or overt hyperthyroidism was evaluated in 6162 patients: no difference was observed between obese and non-obese subjects (2% vs. 3%, respectively, p = 0.067) and (1% vs. 2%, respectively, p = 0.150) [55, 68, 86].
Discussion
Thyroid impairment and obesity are among the most frequent conditions in the general population [1, 4–9]. Although available data have uncovered an intriguing relationship between these two conditions, the chicken-egg conundrum has not yet been completely solved. Considering the large set of data published on obese adults and obese children, we performed a systematic review with the aim of improving knowledge in this field.
With reference to obese children, thyroid diseases were found to be equally distributed among female and male patients. The large set of data included subjects with a mean age of 10.9 ± 1.4 years, thus mostly prepubertal, and significantly younger than those without thyroid impairments (11.0 ± 0.4 years, p < 0.001). In addition, the most frequent finding was subclinical hypothyroidism, identified in 988/9784 (10.1%) children. Grandone et al. reported that the mean age is lower for obese children with higher thyroid-stimulating hormone (TSH) levels [28]. Moreover, some authors described a spontaneous reduction in TSH with increasing age both in obese and in non-obese children [95–97]. In addition, cross-sectional and longitudinal studies demonstrated a positive correlation between obesity and weight status [22, 23, 27, 28, 41, 98–102]. Isolated hyperthyrotropinaemia is commonly reported in obese children. In the paediatric population, risk groups for this condition include children with diabetes mellitus and those with Down syndrome [29]. Elevated TSH levels in childhood obesity may be a direct effect of nutrition or an effect of leptin (which is usually elevated in obesity) on the production of hypothalamic thyrotropin-releasing hormone (TRH), which stimulates pituitary TSH secretion [29, 103–105]. In addition, Burman et al. suggest the presence of thyroid hormone resistance at the pituitary level [106], whereas D’Adamo et al. believed it may be caused by an increase in oxidative stress [107]. Even without any certain cause, it is reasonable to assume that during puberty there is a sexually dimorphic effect with subtle alterations in the hypothalamic-pituitary-thyroid axis [96]. To treat or not to treat patients with hyperthyrotropinaemia with levothyroxine is still debated, even though it is common and accepted practice to treat patients who test negative for thyroid antibodies and display clinical signs or symptoms of hypothyroidism. In obese children, many authors suggest that pharmacological treatment should be avoided, showing that changes in diet, lifestyle, and physical activity lead to a spontaneous restoration of normal TSH values and reduction of body mass index [27–29, 41]. In particular, Eliakim et al. concluded that there are no beneficial effects on body weight, body mass index, linear growth, and body lipids in treated subjects [29].
Thyroid antibody tests were carried out in 211 children with subclinical hypothyroidism, with just 13% testing positive [22, 23, 25, 29, 33, 34], which would appear to support the notion of isolated hyperthyrotropinaemia.
Dursun et al. compared the positivity for thyroid autoantibodies between prepubertal and pubertal age, finding a prevalence in pubertal age (81%) [31]. These data were recently confirmed by Calcaterra et al. [108]. These results support the idea that sexual hormonal changes during puberty could play a fundamental role in immune function [109–111]. In obese subjects, this seems to amplify the obesity-related, chronic, low-grade inflammation process, which could initiate the autoimmune cascade and consequently affect the autoimmunological response [112]. It is widely accepted that white adipose tissue is a significant source of cytokines, chemokines, and adipokines, such as interleukin 6 (IL-6), tumour necrosis factor-alpha (TNF-a) [113], and leptin, a hormone well-known for its role in inflammatory processes and autoimmunity [112]. Indeed, leptin, a 16-kDa polypeptide hormone, appears to exercise several functions in metabolism [114], but it is also implicated in innate and acquired immunity by stimulating the proliferation of monocytes and the production of pro-inflammatory cytokines, such as IL-6, IL-12, and TNF-a [115]. In addition, it promotes an increase in naive T cells, a decrease in T regulatory cells (Tregs), and the differentiation of memory T cells toward T helper 1 (Th1) suppressing T helper 2 (Th2), resulting in an increase of pro-inflammatory cytokines [115]. Therefore, leptin regulates immune response by promoting a pro-inflammatory profile and facilitating the onset of autoimmune disorders, such as AITDs, in the context of obesity [20].
By contrast, greater attention should be paid to obese children with AITDs, insofar as there is the non-negligible risk of progression from subclinical to overt hypothyroidism, indicating that levothyroxine replacement therapy should be started early [116–118].
In terms of obese adults, thyroid diseases are more frequently associated with women, as in the general population [4, 8]. Thyroid impairments occur five times more frequently in obese adults than in obese children. This could be due, firstly, to sexual hormones, which play an unquestionable role in autoimmune regulation. It is well known that male sex protects against the development of autoimmune disease, and it is likely that this is due to sexual hormones and the Y chromosome [119, 120]. On the other hand, women have a marked preponderance of many autoimmune diseases, and this is likely due, at least in part, to the stronger female system. Oestrogen has numerous effects on the regulation of autoimmune cascade, in particular enhancing Th1 responses [121]. Second, it is believed that longer exposure to obesity triggers thyroid diseases di per se [122–128]. In agreement, we showed that adults affected by thyroid diseases had a higher BMI, with one-third showing overt-hypothyroidism. We cannot, however, state that the obesity is caused by hypothyroidism; we are back to the chicken and the egg. Sami et al. give us a Giano Bifronte (two-sided) explanation: “on one side, lack of thyroid hormone leads to weight gain culminating in overt obesity, which in turn predisposes patients to develop autoimmune hypothyroidism. On the other side, raised TSH in obese patients does not show hypothyroidism but it is the result of weight gain rather than its cause” [56].
However, this review clearly showed a direct correlation between BMI and thyroid function (Fig. 4): AITD patients in spontaneous euthyroidism showed a BMI that was midway between hyperthyroid and hypothyroid, as widely expected. Indeed, as Giano teaches us, among a large series of patients, hypothyroidism was more frequent in obese than in non-obese subjects (p < 0.005), even though the prevalence of autoimmunity was superimposable (p = 0.178). According to the data collected, we can assume that the main cause of the rise in TSH levels in adults is progressive thyroid impairment. This is not surprising. On one side, it is reported that pro-inflammatory cytokines can inhibit sodium/iodide symporter (NIS) mRNA expression and iodide uptake [116, 129, 130], leading to a reduction in thyroid function. On the other side, the long-standing inflammation can lead to organ damage. However, these data are in contrast with a recent review by Gajda et al. The Authors described a two-fold higher prevalence of hyperthyrotropinaemia among obese patients, even if the data were gathered from a small number of articles and patients [131].
Differently from that reported for obese children, in adult patients, treatment with levothyroxine should also always be considered for mild hypothyroidism, in line with current guidelines [132]. The Endocrine Work-up in Obesity recommended in any case that hyperthyrotropinaemia should not be treated with the aim of reducing body weight [133].
The major limitation of the present review is the lack of some data which are, unfortunately, often not provided by the authors when requested. However, the large set of patients, the careful selection procedure, and detailed analysis of data strengthen these results.
In conclusion, we summarized the available studies on the association between obesity and thyroid disorders in children and in adults. There is an undeniable relationship between obesity and thyroid impairments, mainly caused by the state of obesity. Isolated hyperthyrotropinaemia is frequently seen in obese children, often followed by spontaneous resolution. BMI appears to be correlated with TSH levels in obese adults. Hyperthyrotropinaemia should never be treated in children and in adults with the aim of reducing body weight.
Authors’ contributions
F.B.: review of literature, writing — original manuscript; E.G.: review of literature, formal analysis, writing — original manuscript; R.D.: review of literature, writing — review and editing; F.D.: review of literature, formal analysis, writing — review and editing; G.P.: conceptualization and methodology, writing — review and editing; I.P.: formal analysis, supervision, validation, and visualization; F.B.: conceptualization and methodology, supervision, validation, and visualization; C.C.: project administration, conceptualization and methodology, writing — review and editing.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Funding
This review did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.