Thermography, Temperature, Pressure Force Distribution and Physical activity as Predictors of Ulcers in Diabetic Foot: A Systematic Review

1. Introduction

One of the main complications of diabetes mellitus (DM) is the development of peripheral neuropathy (PN). Patients with DM and PN are more susceptible to developing plantar ulcers because they cannot perceive pain and be a trigger for lower extremities amputation [1,2,3]. According to Brand [4], three factors determine the possibility of actual breakdown probability and ulceration of the plantar soft tissue when there is a loss of painful sensation:

• Diabetic neuropathy and its complications in the sole of the foot.
• The magnitude of the forces reaching the foot.
• The distance walked that triggers a state of tissue inflammation.

Nowadays, foot pressure is still considered an important risk factor for the appearance of foot complications. However, this is a poor tool by itself to predict ulcer apparition because predictive values cannot be determined [5,6]. The poor specificity and sensitivity that these studies show for maximum pressures as pathological suggest that pressure cannot be the only factor associated with the development of plantar ulcers. In addition to plantar pressures, shear forces also play an important role in the development of diabetic foot lesions. They can predict higher shear pressure zones and increase plantar temperature accelerating tissue breakdown and leading to ulceration. Therefore, a significant correlation is established between shear forces and temperature increase [4,6,7]. Taking into account that inflammation is the first process for skin wound appearance and, above all, in the absence of pain, early identification of inflammation areas can help prevent diabetic foot ulcers. The use of thermography, as a tool for the analysis of load patterns on the soles, leads to a better understanding of the biomechanical factors of diabetic foot ulcers, for which thermal cameras are used to identify inflammation foot areas and critical points before the appearance of any injury [8,9,10]. Physical activity reduces diabetic foot ulcer risks. Therefore, people with diabetes mellitus should be encouraged to exercise. The optimal intensity of such activity remains unclear and controversial [11]. Hall et al. [12] show that there are significant temperature increases after walking for a short time and distance. They report that six minutes of walking increases 5ºC foot temperature. Burnfield et al. [13] show that walking speeds faster ones result in higher plantar pressures. No studies link both risk factors after analyzing physical activity in diabetic patients. So, it is necessary to assess physical activity (based on the temperature and pressure increase response in diabetic patients) [11]. The main objectives of this study are: 1. to analyze if plantar skin response after physical activity in diabetic patients is a prognostic tool of ulcer apparition 2. To asses if there is a correlation between exercise intensity, plantar pressure and plantar thermographic response.

2. Materials and Methods

Protocol and identification of the problem

This systematic review was conducted according to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement: An updated guideline for the publication of systematic reviews [14]. A systematic literature search has been carried out in order to analyze the scientific evidence on the relationship between plantar pressure and temperature as predictors and daily physical activity as a prognostic tool for plantar ulcers in diabetic subjects. Population, intervention, comparison and outcomes (PICO) strategies were defined for data research, as depicted in Table 1.

Research question

Is thermographic plantar cutaneous response a prognostic tool to avoid foot ulcers after physical activity in diabetic patients? Literature search strategy. The databases used were PubMed, Scopus and Web of Science (WOS). The Medical Subject Headings (MeSH) descriptors used were: physical activity, exercise, diabetic foot, plantar pressure, plantar temperature and plantar thermography, Table 2, being combined using the Boolean operators “AND” and “OR”. The search strategy used was performed in databases during the third quarter of 2023. Three search strategies were used for all databases: 1. (Physical activity) AND (Diabetic foot) AND (Plantar pressure) AND (plantar temperature) OR (Plantar thermography), 2. (Exercise) AND (Diabetic foot) AND (Plantar pressure) AND (Plantar temperature) OR (plantar thermography) and 3. ((Physical activity) OR (exercise) OR (activ*) OR (exerc*)) AND (Diabetic foot) AND ((Plantar pressure) OR (press*)) AND ((Plantar temperature) OR (plantar thermography) OR (temp*) OR (thermo*)). Eligibility and selection criteria: the population to be studied consisted of diabetic and healthy adult patients, without distinction of sex and regardless of the type of diabetes, with or without diabetic neuropathy. Studies in Spanish or English language were accepted. All randomized and non-randomized clinical trials (RCTs), prospective and retrospective studies that referred to the use of plantar pressures and plantar temperature and/or plantar thermography as a tool for the early foot ulcer detection or injury were included, as well as those that referred to physical activity, regardless of the year of publication. Articles that analyzed thermographic response of items like socks or foot insoles were excluded.

Data collection

A Word document table was created to extract data from studies. The main data extracted were: author name, year of publication, the purpose of study, number and type of included participants, methodology, intervention and outcomes.

Assessment of study methodology and quality

PEDro scale was used, it is a validated scale developed through a Delhi consensus by Verhagen et al. [15].

3. Results

Search strategies identified a total of 894 references. After screening, a total of 14studies were included in the review, as shown in Figure 1. Of the 14 selected articles, 12 (85.71%) are descriptive observational studies, while the remaining two (14.28%) are randomized clinical trials. Of the 12 descriptive observational studies, eight are cohort studies (66.66%), one is a cross-sectional study (8.33%), and three are case-control articles (25%).

The studies included were evaluated with the PEDro scale, and the quality of the analyzed studies is GOOD (Average 6.7 points). Two articles had a rating of 9 points (14.28%), another four articles had a rating of 7 points (35.71%), another six articles had a rating of 6 points (42.85%) and the rest 5 points (7.14%). Studies with a score between 9- 10 have an excellent methodology, between 6-8 have a good methodological quality, between 4-5 have a fair quality, and below 4 points have poor methodological quality [16,17]. Despite the relative heterogeneity of analyzed studies, there have been a certain number of consistent quality criteria in all of them: eleven studies meet the random-ness criteria in the selected sample and only in three articles was the sample selected for convenience, which could limit the external validity of results. However, all articles comply with the items on statistical information, which would overcome the bias of convenience sampling and support external validity because results of the study can be applied to the interest population. All studies met item 9 “results were presented for all subjects who received treatment or were allocated to control group, or where this could not be done, data for at least one key outcome were analyzed by 'intention to treat”. Only 2 studies fulfilled item 3 “Allocation was concealed” and no study fulfilled criterion 6 “All therapists who administered therapy were blinded” (Table 3).

Table 4 shows the characteristics of the eight studies included in this review. These were classified by authors and year of publication, the aim of studies, methodology, participants, intervention and results.

4. Discussion

This review aimed to use foot ulcers as a prognostic tool to analyze the response of plantar tissue to daily physical activity while considering different plantar pressure and temperature variables in healthy and diabetic patients. The results from different studies showed that the amount and type of activities analyzed to date were heterogeneous [18,25] and so accumulated load in plantar tissues differed greatly among individuals. All the research we considered [18,25] agreed that plantar pressure and temperature increase during walking and other weight-bearing actions. This activity can be studied by thermography, which may help to evaluate skin degradation and prevent injuries in plantar tissues. Maluf et al. found that the maximum pressures recorded during walking were weakly associated with the pressures recorded in the same participant during other daily activities [18]. This finding suggests that foot pressure during walking may inaccurately reflect loads on plantar tissues during other activities of daily life. According to these authors, the magnitude and distribution of pressures in healthy individuals vary with different types of activity and walking speeds, which may explain why patients at risk of skin ulceration could not be identified based on a predetermined pressure threshold during ambulation and other physical activities. Reddy et al. registered the foot temperature in two different participant age groups (30–40 years and 162 over 40 years) while walking on a treadmill for 45 minutes. They observed that walking cadence caused a progressive temperature increase (5°C over the first 23 minutes at a cadence of 80, 100, and 120 steps/minute) but after 45 minutes of walking, the final temperature did not vary when compared with the baseline [19]. This result was consistent with the findings of a subsequent study by Li et al., although most of the participants in the latter work were older (M = 70 years) and the increase in temperature observed after walking 30 minutes at 3 km/h was also lower (4°C) [20]. Nonetheless, both these studies suggest that there was a significant stabilization or plateau effect on foot temperature during walking. The concept of a foot temperature plateau seems intuitive given the physiological mechanisms that support heat generation, including muscle contractions, frictional forces, or viscoelastic heating, as well as factors that control heat loss such as radiation, conduction, and evaporation. However, further investigation would be required to confirm and explore this idea. Regardless, blood flow can either increase or dissipate temperature in different contexts. In this line, Nemati et al. developed a model to describe heat transfer in shod feet that could predict the temperature of the sole of the foot, especially at the level of the plantar arch and metatarsal area [21]. These authors also considered the heat loss caused by sweating during walking and running, measuring a maximum increase in the temperature of the sole of the foot after 30 minutes of 6°C, 8°C, and 11.5°C for speeds of 3, 6, and 9 km/h, respectively. The mean age of the participants in this latter work was 23.8 years, and like previous studies [19,20], the increase in temperature during walking and running was greater in younger participants. Nemati et al. [21] showed that cooling of the plantar arch and metatarsal area as a thermoregulatory mechanism—through sweating—was minimal at a speed of 3 km/h but appeared in the plantar arch area after 15 minutes at 9 km/h and at 20 minutes at 6 km/h, although sweat in the metatarsal are was negligible at all these speeds. Thus, based on all the above, it can be deduced that sweating plays a fundamental role in the thermal regulation of the plantar arch but is insignificant in the metatarsal area. Niemann et al. [22] concluded that plantar temperature decreased significantly during prolonged standing periods which could cause plantar injuries, especially in individuals that work in an upright position. These authors observed a pronounced temperature decrease in the maximum pressure areas (fifth metatarsal head and big toe) after alternating 6 periods of prolonged standing (5, 10, 20, 5, 10, 191 and 20 minutes) with sitting periods (5 minutes each) in healthy and diabetic participants. Pressure sensor recordings revealed that during longer standing periods there was intermittent pressure as the result of discomfort in healthy individuals but not in patients with diabetes and polyneuropathy. Given the relevance of the blood supply to tissues, it can be imagined that patients with diabetes may experience critical tissue ischemia when exposed to their own body weight for prolonged periods of time because they do not seem to experience foot discomfort in the same way as healthy volunteers [22]. Indeed, in the study by Carbonell et al. thermographic images were taken in both diabetic patients and a healthy control group before and after walking 100 m barefoot on a treadmill at a self-selected speed [23]. The temperature of the feet of the diabetic individuals decreased more than in healthy individuals at all the regions of interest (ROIs) analyzed (left and right big toe, forefoot, midfoot, and rearfoot), with the rearfoot and forefoot presenting the greatest temperature differences between the groups (−1.0°C), although not significantly so. In turn, in their study using infrared images to analyze feet in a military training environment, DiBenedetto et al. explained that any sudden increase in weight-bearing activities (such as walks or long runs) caused greater emission of heat in the areas that received the highest amount of pressure during the gait cycle, as reflected in thermographic imaging [24]. These authors determined normal foot parameters by dividing them into four regions: the toes, heel, forefoot, and rarefoot, observing that the temperature difference in the whole thermographic foot image was 1.1°C. In contrast, the normal differences between the hottest and coldest foot areas were 3°C to 5°C, with changes greater than 6°C suggesting the presence of a pathology. Interestingly, these data coincide with those from Reddy et al., where plantar temperature increased by 5°C during walking [19], with DiBenedetto et al. adding a temperature threshold of 6°C at which point a foot injury would be detected [24]. These authors showed that the warmest areas were shown in injured feet and at the greatest pressure points, despite the absence of a feeling of discomfort in these patients [24]. Of note, when stress fractures were suspected but the radiological findings were negative, thermography was still sensitive to this pathology. Indeed, correlations between stress fracture, thermography, and gammagraphy were positive in 66% of cases. Thus, this situation reflects the need to evaluate thermograms qualitatively and quantitatively to relate their results with clinical findings. Vasomotor changes in foot injuries are not only associated with hyperthermia but can also be related to hypothermia. This was also studied by DiBenedetto et al. in 1,000 military recruits during basic training, with this phenomenon being referred to as a reflex sympathetic dystrophy [25]. The authors pointed out that qualitative thermogram interpretation showed a high sensitivity for anomaly recognition but a low specificity in diagnosis terms. They also found that it is difficult to distinguish temperature changes caused by bone or soft tissue injuries and to draw conclusions about the severity of said lesions [24]. This is because hot areas simply signify a greater vascular response to injury but this is not always limited to the location of the lesion. Compared to bone injuries, soft tissue injuries often cause even hotter areas to appear on thermograms. This may be because of the bone depth compared to more superficially located and more readily injured soft tissues. Hence, thermography can be a good complement to understanding foot stress pathophysiology, although it must be accompanied by detailed clinical examinations and complementary tests because there are no consistent criteria to help establish a foot diagnosis when using thermograms. Importantly, as Reddy et al. showed, increased plantar temperature and vertical foot pressure are moderately correlated [19]. These authors suggested a relationship between foot load, foot pressure, and plantar temperature changes could exist in terms of shear forces. In turn, Yavuz et al. used a thermal infrared camera and a custom-made shear force plate to study a possible association between plantar tensions and increased foot sole temperature after exercise [26]. After patients walked for 10 minutes, pressure, stress, and shear data were recorded and moreover, thermograms were performed before and after exercise. This led to the identification of a moderate linear relationship between shear forces and maximum post-exercise plantar temperature (r = 0.78, p = 0.0002). It is thought that increased plantar temperature resulting from physical activity depends more on shear rather than triaxle plantar stresses, although thermal cameras are not reliable for predicting shear stress forces. However, in contrast to the findings by Yavuz et al. [26], the results from Priego Quesada et al. highlighted an inverse relationship between foot eversion and thermal symmetry immediately after running, with the medial ROIs showing higher temperatures than the lateral ROIs (p < 0.05) 246 [27]. In the same line, Catalá-Vilaplana et al. analysed how using unstable versus stable shoes affects the temperature of the sole of the foot after walking for 10 minutes at 1.44 m/s with each shoe type (with two hours of rest between each test) [28]. These authors only found significant differences in the results for the midfoot area (p = 0.004, ES = 1.2). One possible explanation for this outcome could be that the side of the foot experiencing less contact time during the stance phase may be exposed to greater friction with the footwear. This would result in a greater temperature increase [27] that would also remain more constant in unstable shoes because of the change in plantar distribution they create [28]. Similarly, Cuaderes et al. studied diabetic patients who either performed weight-bearing activities at a moderate intensity (athletes) or who did not perform these activities (non-athletes) [29]. This work found significant differences in pressure, temperature, and the plantar surface skin temperature gradient between these cohorts. This was especially true among athletes where the temperature gradient was higher in the first metatarsal head of these participants than in those who did not exercise (athletes: M = 1.66, SD = 1.31 vs. non-athletes M = 1.20, SD = 1.20, p = 0.02). These authors only found one significant linear relationship: between plantar pressure and weight-bearing activity in athletes in the second metatarsal head (r = 0.237, p = 0.02) and the third metatarsal head (r = 0.264, 262 p = 0.01) of the right foot in both cases. In the same vein, Perren et al. performed pressure tests in healthy and diabetic participants with and without complications (peripheral arterial disease or neuropathy) who walked for 15 minutes on a motorised treadmill [30]. One minute after stopping walking, thermograms of the hallux, medial, central, and lateral part of the forefoot and heel were recorded. There was a positive correlation between temperature and plantar pressure in the ROIs of the hallux, second to fifth metatarsal head and heel in diabetic patients with complications, while this relationship was only seen in the fifth metatarsal head in the healthy and diabetic participants without complications (p < 0.05). There was also a positive correlation for all the ROIs in the group with diabetes (both with and without complications) and only for the second to fifth metatarsal head in the group of healthy participants (p < 0.05). This work demonstrated that there was a positive correlation between pressure and temperature: as pressure increased, temperature also increased. There was a tendency for pressure areas to warm up in healthy individuals, including those with uncomplicated diabetes, and this correlation was stronger in diabetic patients with complications [30]. Similarly, Cuaderes et al. observed higher plantar pressures in sportswomen [29] while Jimenez et al. observed higher plantar temperatures in women versus men for all foot ROIs (forefoot: 12.3 vs. 9.8°C, p = 0.004; midfoot: 9.3 vs. 8.0°C, p = 0.047; and rearfoot: 10.7 vs. 8.7°C, p = 0.003) [21]. Therefore, these data indicate that sportswomen presented higher plantar pressures and temperatures than sportsmen. Jimenez et al. went on to study the relationship between pressure and plantar foot temperature and/or shoe sole type in athletes [31]. The participants were recreational runners simulating regular training on a treadmill for 30 minutes at a 1% incline (after a 6-minute warm-up). The highest plantar pressure values were recorded for the forefoot (p < 0.001), although these authors did not distinguish specific ROIs as in previous studies. Nonetheless, there were no differences between the forefoot and midfoot for plantar temperature, although lower values were found in the hindfoot (p < 0.001). Thus, as also concluded by Reddy et al. [19], plantar temperature did not seem to be related to plantar pressure. Finally, it is important to recognize that there were several limitations to this review that must be considered when interpreting the results. One of its main restraints was the low number of manuscripts included because we limited them to studies that included the key words “plantar pressure” and/or “plantar temperature” and “physical activity”. Nevertheless, the quality analysis we performed allowed us to establish relationships with implications that will be important in clinical practice. These findings can help predict skin deterioration in diabetic patients with peripheral neuropathy and allow the recommendation of personalized physical activity regimens to each patient according to the physiological responses of the soles of their feet. Another limitation was that the studies we analyzed were mostly cross-sectional and so future descriptive studies should take a surveillance perspective in order to reveal more information about the use of physical activity as a prognostic tool for assessing diabetic foot ulcers.

5. Conclusions

The results of this systematic review reveal that the magnitude and distribution of pressures in healthy and diabetic subjects vary between different types of activity, just as load patterns are modified by walking speed. Similarly, foot temperature increases by 5ºC when walking, and a temperature threshold of 6ºC may be useful to detect or foresee a foot injury. For all these reasons, thermographic evaluation of plantar temperature and pressure changes during exercise may help to prevent plantar lesions.