In recent years, the literature on Infrared thermography (IRT) on the human body has increased because of its applications in several fields. In clinical medicine, IRT is engaged in postural (1), joint inflammation, degenerative pathologies, diabetes, breast cancer, dentistry, and cardiovascular system evaluation (2–5), evaluation. In sports science, through the detection of temperature asymmetries, IRT is useful in injury prevention by detecting inflammation of underlying muscle groups (6, 7).

Furthermore, in this field, IRT can provide information on muscle metabolism and vascular conditions (8). The interest behind this tool is that it is contactless, radiation-free, non-invasive, and inexpensive (9, 10). This tool detects body heat radiation (11), allowing the evaluation of the skin temperature in a pre-determined region of the body, the region of interest (ROI), through a summation of heat expression from the deeper tissues to the upper levels and blood flow (12). However, different factors could influence the IRT procedure, as a recent scoping review highlighted (2).

Standardization interests the data acquisition and processing and the creation of the region of interest (6) thus, standard operating procedure can help in this regard (13). The environmental conditions and settings, equipment accuracy, and physiological factors are some examples of counting factors (14). To minimize most of the confounding factors, the checklist entitled “Thermographic Imaging in Sports and Exercise Medicine (TISEM)” (15) was proposed. On the other side, factors such as the thickness or composition of the tissue (14) or other characteristics of the participants could also influence temperature detection. Body composition is being studied in the recent literature as one of them. However, it is unclear to date which body composition characteristics, in particular, should be attended to before performing IRT to have accurate and reliable results. According to different studies, there is a negative correlation between the total body fat percentage and the skin temperature evaluated through IRT (16).

However, just the total fat mass was evaluated as a possible body composition variable that affected the IRT. In addition, the tools used in these studies were error-prone and not discriminating between body structures, or composition, such as body mass index (BMI) (17), or expensive and not easy to find, especially in the sports context where IRT is more widely used, such as dual-energy x-ray absorptiometry or computed tomography scan (18, 19).

Another variable to date not investigated that could be an important variable to consider during IRT is the percentage of total body water (TBW). The TBW compartment consists of the extracellular water (ECW) and the intracellular water (ICW). The ratio ECW/TBW gives us information on cellular hydration.

ECW is considered to be related to dehydration; excessive fluid retention is known to cause increased morbidity (20). Normal values of the ECW/TBW ratio are about 0.38–0.40 (40%), values of up to 45% of TBW are tolerable, and ECW levels above about 45% of TB indicate water retention (21–24).

The bioelectrical impedance analysis (BIA) is an indirect, quick, and non-invasive method used for body composition analysis by measuring the resistance and reactance of biological tissue at multifrequency electrical currents. It is an accurate and cheaper tool that could be used in different health conditions (25). Individual tissues have peculiar electrical conduction properties. Resistance indicates the conductive characteristics of bodies and fluids; it decreases with increasing proportions of water in tissues. Reactance is related to the volume of ICW, indicating the capacity of the cell membrane, it is the opposition of the capacitor to alternating current. These two electrical values are used in the equations to calculate the electrical properties of tissues in the various BIA models (26).

The BIA is highly correlated with DXA (17, 27) and has been validated as an alternative to the subjective and biochemical indicators of hydration status (28, 29).

Water, along with the electrolytes dissolved in it, is particularly important in conducting electricity through the body. BIA could esteem TBW with an error of 1.5–2.5 kg by passing a small electrical current (50 kHz) through the body and measuring the impedance to that current (30). After estimating the TBW through established equations and models, the BIA estimates the volume of extracellular and intracellular fluids (31). All BIA models are based on the assumption that the resistance at 50 kHz is proportional to the TBW and from the TBW is distinguished, by subtraction, the resistances of ECW are richer in chloride, and the resistances of ICW are richer in more potassium. Thus, the estimation of TBW, ECW, and ICW is based on regression models taking into account the current passage of 50 kHz in which the formula is adjusted by population-derived indices (age, sex, height.) to reduce interindividual errors (26). Thus, the reference method to assess TBW, the deuterium oxide (D2O), is used to calibrate the various BIA models and make it possible to calculate impedance (30).

Adipose tissue and extracellular water have active electrical resistance. In contrast, reactance arises at the cell membrane of tissue with high water content, which acts like a capacitor composed of two covers and a dielectric layer.

Just one study (19) adopted BIA to associate body variables with thermal camera parameters. However, this study (19) considered the correlation just of the total body fat percentage (with no segmental limbs analysis) and the skin temperature variation after exercise in the upper limb, and only in male participants. Therefore, considering the current shortcomings in the literature, the study aimed to investigate the potential influence of body composition variables on thermal imaging detection after vigorous exercise by focusing on water retention and gender differences. This could allow us to understand whether IRT can be used as an effective tool to monitor the performance of participants with different physical characteristics.

Eventually, the study also wants to provide a value above which the thermal image loses effectiveness in this condition. For the first time, this study analyzed the influence of ECW and segmental body fat, assessed by BIA, on skin temperature variation during exercise in one of the anatomical segments most prone to fat accumulation, the thigh.

One hundred two participants were included in the data analysis [22.75 ± 5.05 years old, BMI 23.11 ± 3.65 (kg·m⁻²), Body Fat 19.48 ± 7.61 (%)]. Sex distribution was 58,8% male (n°60) and 41,2% female (n°42). A statistically significant difference was found between the gender groups in all body composition variables: body fat (%) (male: 15.63 ± 4.4, female: 24.98 ± 7.9; p < 0.001), total body water (%) (male: 59.91 ± 3, female: 51.66 ± 5.4; p < 0.001), extracellular water (%) (male: 40.94 ± 0.98, female: 43.44 ± 2.6; p < 0.001), dominant leg fat % (male: 13.27 ± 3.3, female: 32.30 ± 6.3; p < 0.001), and non-dominant leg fat % (male: 13.66 ± 3.6, female: 32.50 ± 6.2; p < 0.001).

The study aimed to investigate the influence of water retention and segmental fat mass on the skin temperature variation during and after vigorous exercise detected by IRT. A negative association between basal thigh skin temperature and lower limb segmental fat mass was found. The heat generated into deeper body tissues reaches the epidermis to promote thermoregulation, this occurs due to the temperature gradient and depends on the thermal conductivity of various body tissues (42), which depends on water content and affects body thermal conductivity. Fat tissue has low thermal conductivity. Therefore, the subcutaneous fat mass affects the heat emitted from the skin surface. This might justify our results with lower temperatures in regions with a higher fat percentage. We obtained the same results when we divided the group by gender, where the women's lower thigh temperature than men in all three-time points (p > 0.05) is explained by their higher limb fat %. Several studies found a similar negative association between total fat mass and segmental skin temperature (16, 18), adipose tissue acts as an insulating barrier, which influences heat modulation through the skin this alters the body's ability to respond to changes in environmental temperature at sites with excess adipose tissue (43). However, Salamunes et al. used the DEXA to evaluate the fat mass and considered only female participants (18); Reis et al. performed body composition analysis with DEXA considering just men. Neither study considered the influence of segmental body fat on the change in mean temperature during and post-exercise by also comparing males and females. Weigert et al. (19), in agreement with our results, showed that body fat and skin temperature have a negative correlation. However, he considered just the total body fat percentage, and he analyzed the temperature in the biceps brachii segments, which is not an excessive fat accumulation and retention point (44). In our study, the thigh was considered both for temperature and fat assessment and as a segment involved in exercise so for the first time segmental temperature has been associated with segmental fat of the limb direct involved in the exercise.

The thigh is considered among the points of greatest fat accumulation and water retention (44).

We found a negative correlation between ECW and the basal average temperature in both legs (Table 1) this could be justified by the characteristics of the water retention condition. Water retention is an excessive build-up of liquids, which generally accumulates in the circulatory system, body tissues, or specific cavities in the body, i.e., thighs (45). The retention is characterized by edema. This alteration was considered a component of the pathophysiology of adipose tissue diseases (46). However, there is not necessarily an association between retention and the amount of body fat, retention can have several causes also independent of fat accumulation (45). So it is reasonable to analyses the ECW variable independently from the fat % to understand it's influence on IRT. Extracellular fluid has altered resistance, the conductance of extracellular fluid is determined by the concentration of ions, with sodium being the most concentrated (47). Higher tissue sodium content results in lower impedance in legs affected by lipedema and is a major component of the composition of tissues affected by water retention. It's demonstrated that altered extracellular fluids contribute to the pathological skin characteristics of areas affected by lipedema (48) by altering the conductivity in these tissues and consequently the processes of skin thermoregulation (49). Skin temperature depends on the vascular supply of skin and subcutaneous tissues and the thermal conductivity of the skin. Areas affected by lipedema have been shown to have 35% lower blood flow than unaffected regions. Thermographic images of the participants with edema are characterized by inhomogeneous temperature with ischemic areas where the temperature is lowered (hypothermia) were the cold spots may be related to blood stasis (50). In our study the HWR group had a lower temperature than the other groups at all evaluation time points (Table 3). However, while T_base and T_rest, this difference was significant, there was no difference between groups at the end of vigorous exercise (T_exe).

Muscle region's thermal adaptation to exercise in our study showed a decreased temperature immediately after the exercise (T_exe), both when we considered the whole group and when we considered separate groups for gender (Figure 2) or ECW content (Figure 3).

This temperature trend after exercise and rest time agrees with several articles that show that temperature tends to decrease during and at the end of exercise and increase during the recovery phase (51–54). Particularly, the temperature decrease seems to be about −0.4°C or more (51, 52). In our study, this range was complied in the LWR and MWR groups but not in the HWR (Table 2) that had an inhomogeneous trend between the dominant and non-dominant limb of 0.15°C and 0.80°C respectively underling that water retention could result in inhomogeneous temperature areas. The reduction in T_exe values in all groups could be explained by a redistribution of blood flow from the skin to the muscles activated during exercise to allow for the adequate supply of substances necessary for the increased muscle work. indeed, an indirect relationship between skin temperature during exercise and exercise intensity has been demonstrated (55). In our study, the vigorous intensity of the proposed exercise produced this effect, by reducing the temperature from T_base to T_exe. Another mechanism that may contribute to a sudden decrease in skin temperature at the exercise start could be the evaporation of sweat that occurs due to the increase in metabolic activity during exercise, which causes the internal temperature to rise, producing sweat to reduce it at the skin level. However, in our study, this factor should be considered with caution, given the short duration of exercise. In addition, a relationship between sympathetic and parasympathetic activation and changes in skin temperature during physical exercise has been demonstrated (52). Usually, during the warm-up phase, there is a sympathetic neurogenic vasodilation, and during the cool-down phase, there is noradrenergic vasoconstriction in the muscles areas involded in exercise (56). Exercise influenced these events because it is related to several hemodynamic changes. Muscle activity results in an increase in body temperature due to the activation of the energy mechanisms involved (57), and this activates the cutaneous thermo-regulatory processes. Therefore, we have an altered skin temperature as a result of muscle activation during exercise. Thus, the decrease of skin temperature immediately following (T_exe) the vigorous exercise we administered may be associated with skin vasoconstriction attributable to an increase in vasoconstrictor hormones released following the performance of all-out exercise (58).

In contrast, when we considered the temperature after the recovery phase (T_rest), there was an increase in temperature in the whole sample, in the LWR and MWR groups but the temperature still decreased after 5-min recovery (T_rest) in the HWR group. This condition was the same for the dominant and non-dominant limbs. Recent literature demonstrated that in normal conditions, the thermal images during recovery phases showed an increased cutaneous temperature because a new metabolic condition arises, requiring specific mechanisms for the release of the metabolic and mechanical heat produced during exercise.

During this period, the redistribution of blood flow follows an opposite direction from that immediately at the start of exercise, i.e., flow increases from the deeper tissues to the skin tissue inducing the activation of thermal dissipation processes (59):

Exercise produces an increase in blood nitric oxide (NO) concentrations, which is an important vasodilator of skin arterioles. NO promotes cutaneous vasodilation during the recovery phase in regions undergoing exercise by directly modulating nerve activity acting on smooth muscles (60). The result is an increased blood supply that causes an elevation of local temperature, which can be measured with thermal imaging cameras.

Other mechanisms could influence the skin temperature increase during the recovery phase such as glycogen resynthesis because the heat generated by these mechanisms occurs in the anatomical structures just under the skin (61), as well is important to consider the excess post-exercise oxygen consumption (EPOC), which varies directly according to the intensity and duration of the stimulus, which will influence skin temperature by the necessary dissipation of heat towards the state of body homeostasis. It has already been verified that the responses of IRTs in continuous (54) and progressive (53) exercise show different thermographsic responses.

However, If we look at the trend between T_exe and T_rest evaluation in the HWR group (Figures 3a,b), we have an opposite trend_. A possible explanation is that HWR participants had ECW levels over the tolerable value of 45%, indicating retention, the subcutaneous tissue of areas characterized by edema resulting from retention is often ischemic areas with a lower temperature than the other regions of the same segment (Figure 1) where the subcutaneous circulation is impaired or altered blood stasis ineffective for the heat dissipation processes occurring during the recovery phase. However, further studies that would allow the investigation of subcutaneous tissue would be needed to accept this explanation. It is reasonable to assume that they may have slower local thermoregulation that cannot be detected in 5 min of recovery. Future studies should explore this relationship after different phases of recovery.

The importance of investigating how skin temperature reacts to exercise in participants with different characteristics could be linked to the accuracy of information we can derive about athletes and their thermal regulation by IRT (62). Because a winged subcutaneous tissue and circulation could give incorrect information regarding temperature symmetry between limbs. This element is an established method for injury prevention by quantifying training load (63) or the biomechanical imbalances between limbs (64) by IRT. Therefore, identifying a specific factor that could affect the detection of skin temperature by IRT could be a key point in the correct use of thermal imaging. In addition, identifying an ECW value above which the use of IRT may not be reliable would be crucial to having standardized protocols.

This study has some limitations. Participants were not divided according to the physical activity level; Future studies could stratify the participants according to the physical activity practiced or different characteristics of body composition. Only the acute effect of vigorous exercise protocol was proposed, in the future, the effect of other types of exercise should be studied. We focus only in the 5 min of recovery after the exercise, analysis of other recovery periods can add value to these initial results. Although all types of exercise are characterized by an increase in sympathetic activation and a subsequent decrease in parasympathetic activation, previous studies indicate high variability in IRT values before, after exercise, and during the recovery phase.

Therefore, it is difficult to establish segmental temperature values in the various phases of exercise to refer to. In addition, although the BIA agreed well with the reference method to assess TBW (the deuterium dilution) (65) the accuracy of predictive equations of BIA models is affected by several factors: the inhomogeneous nature of various body compartments, electrolyte balance independent of fluid changes, and large interindividual variation in differences in circumferences between various body segments. In addition, greater variation in total body volume may result in relatively smaller variations in body resistance and reactance that are below the level of BIA accuracy. However, a BIA segmental approach, as the tool used in our study, reduce BIA tools bias related to interindividual variability in body shape and cross-sectional areas. Future studies would also explore the influence of water retention on skin temperature variation in other body sites, at different times during exercise and during other recovery times to set a specific thermal imaging protocol.

The present study identifies an association between segmental fat mass percentage and ECW with skin temperature, confirms the decrease in skin temperature at the beginning of exercise and its increase in the recovery phases; and, for the first time establishes that participants with ECW >45% exhibit an opposite temperature trend in 5-min post-exercise recovery in comparison with a normal trend. In employing thermal imaging as a tool for assessing post-exercise temperature, we should take into account the participant's water retention.