Metabolic Syndrome (MS) is a clinical syndrome characterized by the presence of a group of metabolic disorders, including hyperglycemia (Diabetes Mellitus, DM, or Impaired Glucose Regulation, IGR), dyslipidemia (hypertriglyceridemia and/or low high density lipoprotein cholesterol), hypertension and central obesity (1). Evidence indicates that MS significantly promotes the onset and progression of type 2 diabetes mellitus (T2DM), cardiovascular and cerebrovascular diseases. Compared with patients without MS, MS patients had a higher risk and mortality of cardiovascular disease (CVD) (with a relative risk of 2.35 and 2.40, respectively) (2). In 2013-2014, the prevalence of MS adults was 31.5% in the US (3). In Mainland China, the prevalence of MS among the population aged 15 years and older has reached 24.5% in 2016 (4). According to the Yearbook issued by China’s National Health and Family Planning Commission, the total cost of hospitalization for diabetes, acute myocardial infarction and cerebral infarction in China has reached RMB 49.1 billion in 2016 (5), more than three times higher than that in 2012 (6). The increasing prevalence of MS implicates a heavy global health burden and socioeconomic cost, making it of great significance for early diagnosis and intervention of MS.

Nevertheless, in China, 2015, the awareness rate, treatment rate, and control rate of hypertension among adults were 51.6%, 45.8%, and 16.8%, respectively (7). In the 2013 national survey, undiagnosed diabetics accounted for 63% of the total (8). In 2010, the awareness rate, treatment rate, and control rate of dyslipidemia in Chinese adults were 10.93%, 6.84%, and 3.53%, respectively (9). The unsatisfying control station of MS might be the results due to asymptomatic MS patients of early-stage and complicated diagnostic procedure, which involves multiple assessments of body indexes such as blood glucose, blood lipid, blood pressure, waist circumference, etc., some of which are invasive examination and inconvenient to patients. Hence, it’s meaningful to explore a method that can effectively and quickly pick out MS patients as well as the high-risk groups of MS from so-called “healthy people”, assisting early detections and evaluation of the MS population.

Infrared thermography (IRT) is a technique capable of capturing infrared radiation emanating from the human body and converting it to temperature with the output of thermal maps (10). This accurate (with a thermal resolution of 0.03-0.1°C), non-contact and non-invasive technique has been used to early screening and efficacy evaluation of numerous diseases such as breast cancer, diabetic neuropathy, liver metastases, cardiovascular disease and so on (11, 12). The body surface temperature is mainly affected by local blood perfusion, but the influence of muscle activity and metabolic activity cannot be excluded (13). It might serve as an indicator of the function of corresponding parts of the body. Therefore, it’s possible to detect diseases before the noticeable occurrence of structural changes.

According to recent studies, the metabolic indexes which are used in the diagnosis of MS do have correlations with skin temperature: temperature of palms of DM patients tend to be lower than that of normal people (14). While overweight females have lower mean abdominal temperature and higher hand temperature than the lean ones (15), and average skin surface temperature (°C) waveform of hypertension participants varies from normal participants (16). Besides, skin temperature of the anterior supraclavicular shows the ability to detect brown adipose tissue (BAT) activation, which is involved in the body weight, glucose and lipid regulation, reflecting the metabolic changes (17, 18). Therefore, IRT exhibits the potential of indicating metabolic disorders, and may have some advantages and application prospects in MS screening.

Based on the current findings, we conducted this study which sought to explore the temperature distribution characteristics of the face, trunk, palms and feet of MS, hope to establish the correlation between specific thermograph patterns and metabolic disorder, and provide reference to further studies and the application of IRT in early assistant diagnosis and evaluation of MS population.

Our study demonstrated that: With the increasing number of abnormal metabolic indexes, the mean temperature gradually increased in the face, bilateral palms and dorsum of feet, and decreased in the anterior trunk. And for the MS patients, the temperature of the anterior trunk and bilateral palms became approximated. This specific tendency may indicate the severity of metabolic disorders.

The gender difference in temperature distribution was the basis of our subsequent analysis. In this study, there were obvious gender differences that the mean temperatures of face, palms and dorsum of feet were significantly higher in males than in females, while the mean temperature of anterior trunk showed the contrary characteristic. Several studies have also reported gender differences in temperature distribution. A study on healthy participants showed that the mean temperature of the chest was significantly higher in women than in men, while in the other analyzed body surface areas, including back, abdomen, and four limbs (anterior and posterior), the mean temperatures were significantly lower in women, which may be related to less skeletal muscle and thicker subcutaneous fat in females (20). Since we did not separate the anterior trunk region into chest and abdomen, the temperature characteristics of the anterior trunk of these two studies are not comparable, while the results that males have higher temperature in palms and dorsum of feet are consistent with this study. Another study found lower mean temperatures in the trunk and four limbs in the female group, which may have a bearing on higher body fat rates in females (21). The result of the lower mean temperature of the anterior trunk in females did not conform to our study. Consequently, further researches refer to the gender difference in body surface temperature distribution with a more detailed delimitation of ROI remain to be done.

The features of the thermal maps of the MS population were the key points of our study. And we did observe specific characteristic changes in the thermal maps of the MS population. There were few articles in allusion to the characteristic of infrared thermography of the MS population. Still, several studies on thermal characteristics of obesity, hyperlipidemia, hyperglycemia and hypertension have been done. Eduardo Borba Neves et al. (21) have discovered that, in males, the temperature in hands (anterior and posterior) had a tendency of increasing along with the body fat rate (BF%). Besides, BF% had a negative correlation with temperature in the anterior trunk, which was basically consistent with the results of our study. Eun-Mo Song et al. (22) found that the visceral fat area (VFA) was negatively correlated with the temperature of the acupoint CV4, which is located in the lower abdomen. This also corroborated the correlation between obesity and the decrease of temperature in the abdomen. Thiruvengadam J et al. (12) demonstrated that HDL was negatively correlated with the surface temperature of the left hand and anterior feet, which coincide with the conclusion of this study that severer metabolic disorder leads to higher mean temperature in palms and dorsum of feet. Aleck Ovechkin et al. (23) found the body surface temperature of the “Yin-Tang” acupuncture point which is located in the forehead negatively correlates with the severity of intracranial hypertension syndrome. This observation suggests varied temperature distribution characteristics of the face in hypertension patients compared with normal ones. While in this study, no change in the mean temperature of the face was found in MS participants, suggesting changes in body surface temperature of certain diseases might take place in some particular small areas, which could be covered when chosen ROIs are quite bigger. In summary, the characteristics of higher mean temperatures in palms and dorsum of feet, and lower mean temperature in the anterior trunk might be results of multiple pathological mechanisms as above. Studies on one particular metabolic disorder like hyperlipidemia, hypertension, and hyperglycemia may help in understanding the underlying mechanism of characteristic changes. However, clinical trials on large scales are still in lack, which conduces to a more accurate interpretation of the thermography characteristic of MS.

Through correlation analysis and linear regression, in both male and female groups, we found a positive correlation between the mean temperature of the palms and BMI, and negative correlations between mean temperature of the trunk and DBP, UA, TG, waist circumference, hip circumference, weight and BMI. These patterns can be consistent with the above-mentioned law of increasing palm temperature and decreasing trunk temperature with the aggravation of metabolic abnormalities. The unique results in male groups are: a positive correlation between T_f and DBP, and a negative correlation between T_t and SBP. In females, the unique results are: T_f was positively correlated with height and weight, the mean temperature of palms and dorsum of feet was positively correlated with hip circumference, UA, weight, BMI, waist circumference, FPG, TG and age, negatively correlated with HDL-C, and the average temperature of the trunk was negatively correlated with FPG, LDL-C, age and height. In general, the correlation between T_t and multiple metabolic indexes is obvious regardless of gender. While the mean temperature of palms and dorsum of feet showed correlations with a sizable amount of indexes only in the female group, and correlated with age at the same time. Given the uneven age baseline of female subjects, the relationship between palms and feet temperature and metabolic abnormalities still needs to be verified based on more homogeneous sample studies.

It is also of note that heterogeneity was found among the age of four female groups. The average age of the female MS group was 55.58 ± 10.431 years old, mostly in perimenopausal and postmenopausal stages (start at 47.5 years old on average (24)). The particular age of this group may be due to the prevalence increase of MS in perimenopausal and postmenopausal women. A meta-analysis carried out by Hallajzadeh et al. (25) found the prevalence of MS among postmenopausal women was 37.17% and the pooled OR for MS in postmenopausal women in comparison with the premenopausal was 3.54. In perimenopausal and postmenopausal women, decreased estradiol level attenuates the protective effect of estradiol, which may account for the inclination of carbohydrate and lipid metabolism disorder and contribute to the development of MS (26, 27). However, due to the small sample size of this group, there remains the possibility that the sampling error led to the elder age of women in this group. Accordingly, further study with an enlarged sample size is necessary.

Owing to the retrospective nature of this study, the results could be influenced by time, temperature, humidity and state of equipment, etc. Due to the lack of research on MS using IRT, the sample size was not calculated in advance. While we calculated the sample size afterward based on the results of this study. Except for the comparison of T_lp between M0 and M3, T_t between F0 and F3, T_rf and T_lf between F2 and F3, the statistical powers of the other comparisons were higher than 0.80. Moreover, the sample size and age of each group were not evenly distributed, which may also have some bearing on the outcome. A methodological problem is that we did not capture the thermograph of the dorsum of feet in a vertical angle, for the retrospective nature, also. This could affect the output temperatures with camera (28), while since all of the thermographs were captured in approximate angles, the tendency of temperature changes we observed is still advisable.

This work was a pilot study of temperature characteristics in the MS population, and it proved the feasibility of screening and evaluating metabolic disorders through IRT. The key findings of this work were the gender difference in temperature distribution, the sequences of the mean temperature of the face, anterior trunk, palms and dorsum of feet in subjects of different metabolic conditions, the trends of the temperature changes in the above body parts with increasing number of abnormal metabolic indexes, and the certain differences in mean temperature of these body parts between the normal group and MS group. And the secondary findings are the correlations between measurement or laboratory data related to metabolic disorders and the mean temperature of different body parts. With the popularization of IRT and the deepening of the research, this correlation may help screen patients with metabolic disorders, further postponing the development of diseases.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee of Dongzhimen Hospital. The patients/participants provided their written informed consent to participate in this study.

All authors developed the concept for the study. Y-hX and W-zZ offered the design of this manuscript. RC provided suggestions. M-jG and Z-jC conducted the experiment. H-zX drafted the manuscript. B-yJ did literature search. B-hM provided technical support on the software in this study and guided the revision process of this manuscript. Y-cS reviewed the manuscript and made corrections. All authors contributed to the article and approved the submitted version.

This work was supported by State Administration of Traditional Chinese Medicine JB063 programme and the Fundamental Research Funds for the Central Universities 2020-JYB-ZDGG-117 programme.

Author RC is employed by Beijing Wholelife Medical Science Co., Ltd.

The remaining 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.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.