Investigation of thermal changes in the thyroid gland region of individuals with hypothyroidism and fibromyalgia by analyzing the temperature of brown adipose tissue

This exploratory retrospective study aims to investigate the thermal changes in the thyroid gland region of patients with hypothyroidism and fibromyalgia by analyzing the temperature of the brown adipose tissue (BAT). A total of 166 individuals from 1000 thermographic electronic medical records were classified into four groups: Group HP + FM-50 individuals with hypothyroidism and fibromyalgia; Group FM-56 individuals with fibromyalgia only; Group HP-30 individuals with hypothyroidism only, and Group Control-30 healthy individuals. The thermal images from the electronic medical records were acquired by a FLIR T650SC infrared camera (used for thermometry) and the temperature data for each group were statistically analyzed. Group HP + FM showed r = 0, meaning that the average temperatures of the thyroid and BAT are independent of each other. Groups FM, HP and Control showed r = 1, meaning that the average temperatures of the thyroid and BAT were directly related. Our findings showed that the average temperatures of the thyroid and BAT regions are similar. Also, there was no correlation between thyroid gland temperature and the presence of hypothyroidism or fibromyalgia using thermometry.

Medical infrared thermography (MIT) is a noninvasive and nonradioactive analysis method capable of analyzing physiological functions related to the control of skin temperature, an important organ for controlling body temperature 1,2 . This thermography´s technique allows the evaluation of physiological changes 2,3 , with applications in the field of medicine to identify neurological, rheumatological and dermatological disorders, vascular diseases, urologic, and gynecological and orthopedic pathologies [3][4][5][6][7] , and this method can provide support for sports medicine 3 .
Fibromyalgia (FM) is a rheumatologic disorder, of undefined cause, but its development is associated with the central nervous system's regulation of pain 8,9 , neurosensory, neuroendocrine and neurotransmitter-related disorders, as well as a genetic predisposition 10,11 . FM is characterized by musculoskeletal pain in diffuse to chronic intensity and is associated with symptoms such as fatigue, sleep disturbances, palpebral venous congestion, morning stiffness, diffuse paresthesia, subjective sensation of edema, cognitive disorders, depression and anxiety 10,12,13 . When a thermography examination is performed in a patient with FM, the patient presents a characteristic image pattern of the mantle sign that means a wide and diffuse hyper-radiation on the cervicothoracic region, low level As the research have been conducted using a database, without any interview with patients, the need for informed consent was waived by the Ethics Committee of the Federal University of Technology-Parana (UTFPR).

Materials and methods
The research work was carried out from January 2017 to December 2019 and, relating to the objectives, the study design was exploratory retrospective and the approach was quantitative transversal. The data acquisition process was based on access to a database of a thermography clinic, where the records of patients diagnosed with fibromyalgia, hypothyroidism and without both diseases were selected. The electronic records of 166 individuals from a total of 1000 records were selected using the following inclusion criteria:  13,38 , and HP diagnosed by laboratory examination; and • Preliminary questionnaire for FM criteria that assists the medical professionals in the diagnosis (ACR/2010) 13,40 . The results obtained within this questionnaire for the generalized index of pain (WPI), and scale of gravity of symptoms (SS) should remain between WPI ≥ 7/19 + SS ≥ 5 or WPI 3-6 + SS ≥ 9.
The individuals' records were selected from the database and classified into four groups: The database thermal images were acquired retrospectively using the FLIR T650SC infrared Camera (Flir Systems Inc. Nashua, NH, USA) with the technical specifications shown in Table 1. The AP thermography's images with cervical extension and the AP upper orthostatic images were obtained with individuals positioned at 1 m from the camera in an orthostatic position. The images were processed on a computer using the FLIR Report program (FLIR Tools, version 4.1.140661001) and the temperature data were obtained and analyzed. The block diagram of the setup used to acquire the thermal images is shown in Fig. 1.
The AP thermography's images with cervical extension were analyzed using the average temperatures of three thermal points (Fig. 2a) selected by the researchers. For the AP upper orthostatic images, the thermal points located bilaterally (SP3 and SP10) in the medial supraclavicular region were used, that is, in the BAT region (Fig. 2b).
With the temperatures obtained for each individual, the data were organized into a table with the following specifications: data were organized with a numerical identification for each individual, with an associating age, gender, and FM diagnosis according to the criteria of the ACR/2010 and Preliminary Questionnaire for FM Table 1. Technical specifications of the FLIR T650SC infrared camera used to acquire the patients' thermal images. Temperatures obtained for each group were statistically analyzed with the Microcal Origin 6.0 software (OriginLab Corporation, Northampton, Massachusetts, USA), which obtained the average age and standard deviation of the groups, the average temperature (Tm) and standard deviation for each group (thyroid and BAT), comparisons between the groups, and the average temperature difference (Δ1) between each region studied (Fig. 3). The software also applied a variance test (ANOVA) to evaluate if the average temperatures of the two regions were equal or different. Student's t-test was used to compare the average temperature differences between the groups.

Camera parameters Values
The Action Stat application was used to obtain the Principal Component Analysis (PCA). Individuals were grouped according to their variances, that is, according to their behavior within the population, represented by the variation of their set of characteristics (Thyroid and BAT temperatures). The analysis allowed for a reduction in the number of variables to provide a view of the data set, and thereby helped to identify the most important variables in the space of the main components.
The last statistical analysis was used to obtain the Pearson's linear correlation coefficient or "Pearson's r" to assess the degree of correlation and the direction of this correlation. Thus, if the correlation the variables is equal to 1, it is a positive correlation, meaning the variables are directly proportional to each other. If the correlation is negative (r = − 1), the variables are indirectly proportional to each other. Finally, if r = 0, the two variables are not linearly dependent.
Ethics approval. This

Results
From the 166 individuals selected for this research, 106 were diagnosed with FM according to the ACR criteria 13,40 , 80 patients were diagnosed with HP, and 30 showed no evidence of the pathology. The temperature difference found in the thyroid gland region of the individuals in Group HP + FM was + 1.1 °C. The average temperatures (Tm) of the three selected points in the thyroid gland were 31.7 ± 1.6 °C for Group HP + FM, 31.6 ± 1.2 °C for Group FM, 32.3 ± 1.5 °C for Group HP and 31.9 ± 1.5 °C for Group Control (see Fig. 4). The obtained differences Δ1 were: 0.1 °C between groups FM + HP and FM, 0.6 °C between groups FM + HP and HP, 0.2 °C between groups FM + HP and Control, 0.7 °C between groups FM and HP, 0.3 °C between groups FM and Control, and 0.4 °C between groups HP and Control.
The variance analysis showed that the samples are not significantly different from each other, with an occurrence probability (p) of 0.38 and a frequency of occurrence (f) of 1.01.
The average temperature of the two points (SP3 and SP10) in the BAT region was 31.7 ± 1.5 °C for Group HP + FM, 31.5 ± 1.2 °C for Group FM, 31.4 ± 1.0 °C for Group HP and 31.5 ± 1.5 °C for Group Control (see Fig. 4). The Δ1 was 0.2 °C between groups FM + HP and FM, 0.3 °C between groups FM + HP and HP, 0.2 °C between groups FM + HP and Control, 0.1 °C between groups FM and HP, 0.0 °C between groups FM and Control, and 0.1 °C between groups HP and Control (see Fig. 5). The analysis of variance showed that the samples are not significantly different from each other, with p = 0.51 and f = 0.76.   Analyzing the four groups, the first principal component (PC1) explains 42.96% of the total variation and, according to the eigenvectors, the weights variations of Group HP + FM are negatively high for this component. That is, the higher the average temperatures of the thyroid gland region and BAT were, the lower the score of the first component was. For Group FM, Group HP and Group Control, according to the eigenvectors, the weights variations are positive. Therefore, with an increase in temperature, the score of the first component will also increase (Fig. 6).

Discussion
The temperature difference found in the thyroid region of patients with FM and HP (Group 1) by thermography was + 1.1 °C with r = 0, so the temperature variation does not impact in average temperatures of this group and the variables had no correlation. That is, for this study group, thyroid metabolism variation and BAT temperature variation are independent of each other, independent of the patient´s diagnosis for HP and FM. However, it was found by other authors that average temperature values higher than 0.26 °C from normal human temperature (36.5 °C to 37.0 °C) already show alterations in thyroid gland metabolism 41 . It should be noted that in the study, the authors evaluated only the temperature variation of the thyroid gland, without correlating another variable as considered in this paper such as HP and FM.
The average temperatures of the three regions of the thyroid gland were similar to each other, as the analysis of variance and Δ1 between groups ranged from 0.1 to 0.7 °C. According to some results reported in the literature for the trunk region, temperature variations of 0.17 ± 0.042 °C showed abnormalities 38 , which do not agree with the results of other research conducted in the same anatomical region that found temperature variations of 0.5 °C to 1 °C 41 . The patients had homogeneous characteristics, besides, TSH serum levels controlled.
HP is among the thyroid dysfunctions that have nonspecific symptoms similar to symptoms of FM 33,42 . In the literature it can be found that HP dysfunction presents alterations in the hypophysis-hypothalamic axis where FM patients have less thyrotropin (TSH) 33,[43][44][45][46][47][48][49][50] . Therefore, in FM patients, pain is being investigated related to changes in the hypothalamic-hypophysis-adrenal axis, and it is still unknown if both cases of HP and FM hyperactivity are present in these axes. It is not known whether this hyperactivity is genetically based or whether it is a result of stress throughout life, or in a specific situation.
BAT is also a target of thyroid hormones, where they present a large number of 3, 5, 3′-triiodothyronine (T3) receptors that are 70% occupied at room temperature and approximately 100% occupied during cold exposure 45,51 . In fact, BAT has its own T3-generating mechanism, due to the local activity of the enzyme deiodinase type II, D2. The activity of this enzyme and T3 concentration increases 3 to 50 times during the sympathetic activation of BAT, resulting in an increase in the local impact of T3, without affecting the plasma concentration . Thermal imagens have been established as a valid alternative for diagnosis for BAT activity and FM 5,22 . In this research, small differences in temperatures were found between the groups, as the thermography equipment measures the body's infrared energy, demonstrating the thermal distribution of the skin surface by a high-resolution image.
In this way, the temperature can be measured from the energy emitted by the skin surface in a totally safe way, that is, without any contraindications 11 . Nevertheless, according to the results, there were no significant differences in the temperatures of BAT between the three groups, and r = 1, reinforcing that the average temperatures of the thyroid region and BAT are interdependent. That is, when there is a variation in the metabolic rate of the thyroid, there will be a variation in the rate of BAT activity. Therefore, elucidating this result requires knowing the impacts of thyroid hormones in human biology, in which acceleration of energy metabolism and ATP turnover are related as a result of energy transformation to heat production 52 . In these terms, thyroid hormones as mediators of homeotherm are present in homeothermic animals, and are capable of stimulating heat production. In humans with HP there is hypothermia and cold intolerance, lose part of their homeothermy and the ability to adapt to the environment 45,52 .
In the literature there are indications that other diseases should be excluded before starting FM treatment, because there are similarities of symptoms 33 to other diseases such as HP, hyperparathyroidism (parathyroid adenoma) and autoimmune thyroiditis due to autoimmune disease 46,53,54 . Therefore, it was necessary to know the patient's profile from the selected records to classify groups and to discard other diseases that could impact www.nature.com/scientificreports/ the results. Thus, the following symptoms were observed in the questionnaire for preliminary FM criteria: chronic pain in the body, showing hyper-radiant regions [54][55][56][57] , depression, tiredness, non-repairing sleep, signs of periocular congestion, constipation, diarrhea, headache and the correlation of chronic generalized pain such as myalgias, arthritis, arthrosis and rheumatism 33,45,49 . All patients maintained controlled treatment for thyroid dysfunction, especially for HP. According to the parameters of the questionnaire for preliminary FM criteria, a patient with FM is diagnosed as such if he or she presents a composition of factors such as the combination between a generalized index of pain (WPI) ≥ 7/19 + a scale of gravity of symptoms (SS ≥ 5) and/or WPI between 3 and 6 + SS ≥ 9, besides clinical exams 38,53,54,58,59 .
However, patients selected for this study had, in Group HP + FM, a mean WPI = 11.6 and a mean SS = 9.4, that in combination confirmed the disease. Group FM had a mean WPI = 10.4 and a mean SS = 8.8, also confirming FM but with normal HP laboratory results. Group HP had a mean WPI = 3.5 and a mean SS = 3.5, the patients included in this group showed no FM, and Group Control presented a mean WPI = 3.8 and a mean SS = 5.3; confirming normality for FM according to ACR 38 .
In this research, it was observed that FM patients presented mantle signals that could be indicated of neurovegetative disorders, besides periocular congestion (nonrestorative sleep), vasospasms in the extremities and other signs 30,52 .
Results of the thermography diagnostic examination, although not a definitive diagnosis, suggest that mantle phenomena together with peripheral vasoconstriction can support clinical diagnosis and play important roles in the follow-up of FM patients as markers of neurovegetative dysfunction present in the disease 54 .
This study followed the outlined objective using a database containing thermographic images and, to carry out the research, there was no contact with the patients. The database with the medical records did not include some data, such as the BMI and serum TSH levels. Future works can be done by analyzing other variables, such as temperatures of other parts of the body, other pathologies, as well as analyzing the impact of BMI.

Conclusions
Analyzing the results in this work, it was possible to conclude that the average temperatures of the three thyroid points (p = 0.38, f = 1.01) and the two BAT points (p = 0.51, f = 0.76) for the surveyed groups were not different, with a Pearson's correlation coefficient (r) equal to zero, signifying that thyroid metabolism variation and BAT temperature variation are independent of each other, independent of the patient´s diagnosis for HP and FM. A similar result was confirmed when applying the Pearson correlation coefficient between the data obtained from the thyroid gland and the BAT regions. The group composed by patients with HP and FM, showed r = 0, meaning also that the thyroid and BAT temperatures are not correlated. Therefore, a variation in the metabolic activity of the thyroid gland does not imply BAT activity.
As for the group composed of individuals with FM and without HP, the group for individuals with HP, and the control group (healthy subjects), it was found r = 1, meaning that the average temperatures of the thyroid and BAT are directly related, where a variation in the metabolic activity of the thyroid gland exhibits an interaction with the variation in the metabolic rate of BAT.
Thus, it was possible to conclude that this thermometry technique can be used for measuring the temperatures of the thyroid gland and BAT regions in order to evaluate their changes. This study has shown that that mean thyroid and BAT temperatures are similar and there was no correlation between thyroid temperature and the presence of hypothyroidism or fibromyalgia using thermometry.

Data availability
Not applicable.

Code availability
Not applicable.