Infrared dermal thermometry is highly reliable in the assessment of patients with Charcot neuroarthropathy

Charcot neuroarthropathy (CN), commonly referred to as ‘Charcot foot’, is a serious limb-threatening complication seen in individuals with peripheral neuropathy. Although diabetic neuropathy is the most common cause [1, 2], it can also result from other conditions with neuropathic manifestations, such as alcoholism and renal failure [1, 3]. The estimated prevalence of CN ranges from 0.08% in the general diabetes population to 13% in the diabetes high-risk foot population [1, 4].

CN is a progressive and destructive bone and joint disease that can affect single or multiple joints of the foot and ankle [1, 2, 5]. It is characterised by acute fractures, joint subluxation, dislocation and instability, and bony destruction [2, 6, 7]. Acute CN typically presents as a warm, erythematous, and oedematous foot [5]. Misdiagnosis in its early stages can lead to gross foot deformity, ulceration and amputation [7, 8]. Therefore, early detection and management are paramount.

Hand-held infrared dermal thermometry is a non-invasive skin temperature assessment used to assist clinical diagnosis, monitor disease progression and resolution, and guide management principles [7, 9‐11]. A temperature difference of more than two degrees Celsius between the affected and non-affected Charcot foot is used as a clinical marker to detect acute (i.e. active) CN, while temperature differences of less than two degrees Celsius support safe withdrawal of immobilisation (e.g. total contact cast) [7, 10, 12, 13].

Commercially available, low-cost, hand-held infrared thermometers have been shown to have good accuracy, reliability and performance in assessing temperatures within a laboratory setting [14]. Other studies have shown a comparative level of agreement and repeatability between infrared thermometers and thermistor-type thermometers [15, 16]. Despite the wide clinical use of infrared dermal thermometry in the diagnosis, monitoring, and management of CN, there is surprisingly a lack of high-quality evidence to support its reliability. In addition, there is no consensus as to which anatomical sites should be included when performing this assessment. Common testing sites reported in the literature include: the hallux, medial 1st metatarsal head, lateral 5th metatarsal head, plantar metatarsal heads 1, 3 and 5, dorsal midfoot, 1st metatarsocuneiform joint, talonavicular joint, cuboid, plantar heel, and ankle [7, 11, 17, 18]. However, there have been no studies to date that have objectively compared the reliability of these testing sites.

Given the limitations of current evidence, this study aimed to investigate: (i) intra-rater and inter-rater reliability of infrared dermal thermometry using a ‘touch’ and ‘non-touch’ technique, and (ii) provide recommendations on the most reliable anatomical sites to test in patients with CN.

We collected data on 32 adults with diabetes mellitus and CN. The mean age was 59.9 (SD, 10.5) years, 68.8% were male, and the average body mass index was 33.1 (SD, 7.9) kg/m². The majority had type 2 diabetes (71.9%). The mean duration of diabetes was 20.6 (SD, 15.1) years and the average glycated haemoglobin was 7.7% (SD, 1.3). Peripheral neuropathy, peripheral arterial disease and previous foot ulceration were highly prevalent (93.8, 43.8 and 50%, respectively). Overall, there were a total of 36 Charcot feet among the 32 participants (four participants had bilateral CN). Charcot foot most commonly affected the tarsometatarsal joints (38.9%), followed by the naviculocuneiform, talonavicular, and calcaneocuboid joints (27.8%). The median duration of Charcot foot at the time of the baseline assessment was 2.8 (IQR, 1.3 to 5.9) months. A large proportion of Charcot feet were receiving total contact casting treatment (69.4%). A quarter of the sample (n = 8) had a previous (i.e. resolved) Charcot foot. Participant characteristics, foot history, and foot examination information are shown in Tables 1 and 2.

The average room and outside temperatures recorded at the first dermal temperature assessment were 22.4 °C (SD, 1.0) and 17.3 °C (SD, 5.4), respectively. Following the minimum 15-min acclimatisation period, the median time to performing the dermal temperature assessments was 1.1 (IQR, 0.3 to 5.0) minutes. The results of the relative and absolute reliability of the dermal temperature assessments in the units of measurement degrees Celsius (°C) are shown in Tables 4 and 5.

This study found that infrared dermal thermometry is a highly reliable tool in the clinical assessment of patients with CN. Overall, there was ‘good to excellent’ intra-rater and inter-rater reliability for the touch technique, and ‘excellent’ intra-rater and inter-rater reliability for the non-touch technique. In addition, measurement error was relatively low across the 10 anatomical sites tested. These findings suggest that either a touch or non-touch technique can be used confidently in clinical practice or research settings.

Intra-rater reliability was found to be better than inter-rater reliability for both techniques. Given that infrared thermometry is used to compare dermal temperatures between an affected and non-affected Charcot foot at a particular point in time (rather than across different days), and that the full assessment is most commonly performed by one clinician at any one point in time, intra-rater reliability is most relevant to clinical practice when performing this assessment. That being said, this study also showed high inter-rater reliability for both techniques, therefore, clinicians can remain confident with the temperature readings if the assessment is shared between clinicians.

Interestingly, a non-touch technique was observed to have slightly higher reliability and lower measurement error than the touch technique. This finding was unexpected as the non-touch technique required examiners to estimate a 5 mm distance from the device to the skin surface. One possible explanation for this finding is that the amount of pressure being applied by the raters for the touch technique may have varied, and therefore caused some inconsistencies in the temperature readings. This finding is in contrast to recommendations outlined in the Exergen Corporation® DT-1001 user manual, which states for “maximum accuracy the probe must contact the surface at the point of interest” [29]. While a touch technique may improve accuracy (i.e. temperature measured is reflective of the true temperature), our findings suggest that a non-touch technique is slightly more reliable than the touch technique. The current study focused on investigating the reliability of dermal temperature assessments (i.e. test-retest performance), which from a clinical perspective, is considered most important in establishing temperature differences between an affected and non-affected Charcot foot.

Measurement error was found to be relatively low across the 10 testing sites for the two raters and techniques, as indicated by the MDC values obtained (Tables 3 and 4). The MDC value represents an estimate of the amount of change required (in this case to dermal temperatures) for the change to be considered ‘real’, which is over and above measurement error [30]. As an example, if the MDC value is equal to 0.5, a change in Charcot foot temperature that is ≤0.5 °C in a test-retest scenario (e.g. trial 1 = 28 °C then trial 2 = 27.5 °C) would be accountable to measurement error (e.g. operator error). Inversely, any change in temperature that is greater than 0.5 °C would be due to a true change in temperature. Of the 10 anatomical sites assessed in this study, sites 1 and 2 (plantar 1st and 3rd metatarsal heads) were observed to have consistently high reliability across the raters and techniques. These sites were less likely to be affected by CN in this study, as a large proportion of participants presented with midfoot Charcot. Therefore, a potential explanation for this finding is that sites correlating to joints affected by CN may be less reliable and requires further investigation. Considering previous studies have found that sites of elevated dermal temperature correlate with the joints affected by CN [7, 11, 17], the presentation or pattern of CN is often variable with multiple joints affected, and the number of sites that should be tested in patients with CN is currently unclear, dermal temperature assessments should include an appropriate number and range of relevant sites to account for this. Given that there was no discernible difference in reliability across the 10 anatomical sites used in this study, and the assessment is non-invasive and quick to perform, the 10-site protocol presented in this study can be confidently used in clinical practice.

There are several potential limitations of this study. First, we did not assess different acclimatisation periods and whether they have an impact on the reliability of dermal temperature assessments. As per a previous study’s [26] recommendations, we used a minimum 15-min acclimatisation period to ensure stabilisation of foot temperatures prior to testing. Despite our best efforts to be consistent in the timing of dermal temperature assessments among participants, there was some variability from the end of the 15-min acclimatisation period to the start of the dermal temperature assessment. To adjust for this, room and outside temperatures were recorded prior to data collection to ensure no significant changes to foot temperatures may have occurred. Second, both raters in this study had high levels of clinical experience using infrared dermal thermometry to assess patients with CN. Therefore, it remains unclear whether less experienced clinicians or those with no experience would have reduced reliability when conducting this assessment. Third, we did not assess for the accuracy of the infrared dermal thermometer, therefore, our results can only be used for interpreting reliability of this assessment. Fourth, it remains unclear whether room or outside temperature changes may have had an impact on the dermal temperature measurements. However, as room and outside temperatures were recorded for each assessment by the two raters, and all participants were assessed in the same clinical environment, this was unlikely. In addition, temperature differences between an affected and non-affected Charcot foot is the key clinical factor when assessing CN, therefore, climate control of the environment is not as essential. Fifth, this study used either a touch or non-touch ‘pin-point’ testing method, therefore, our results are not generalisable to clinicians that use the ‘scanning’ method (i.e. moving probe around anatomical site to get highest reading). Sixth, our findings are only generalisable to patients with diabetes-related CN (i.e. not from renal disease, alcoholism etc.). Finally, recall bias may have been present (e.g. participants self-reported previous foot ulcers and CN duration), however medical records were referred to if clarification was needed, so this was unlikely.

There are several strengths of this study. To our knowledge, it is the first study to investigate intra-rater and inter-rater reliability of infrared dermal thermometry in CN. This study had a rigorous inclusion and exclusion criteria, sufficient sample size [28], robust study protocol, and our findings are able to be generalised to clinical practice as our study included patients with CN from a HRFS.

As this is the first study to evaluate the reliability of infrared dermal thermometry in the assessment of patients with CN, there is limited capacity to compare the current study findings to previous literature. Our finding that infrared dermal thermometry is a highly reliable assessment tool, is consistent with a previous laboratory-based study [14]. This study found high reliability (r > 0.989) between temperature change measurements of two raters when comparing nine commercially available infrared thermometers (including the Exergen Corporation® DT-1000) [14]. Infrared thermometry has also been shown to have high correlations between: (i) different thermometers (r > 0.80) in assessing other high-risk foot conditions such as peri-wound temperatures and (ii) the assessment of skin temperature via palpation (r_s = 0.81, p < 0.000) [27, 31].

Further research is needed to establish an evidence-based protocol for dermal temperature assessment in patients with CN. Future research may be directed towards investigating the reliability of: (i) a touch and/or non-touch technique of the ‘pin-point’ versus ‘scanning’ method, (ii) the comparison of an affected and non-affected Charcot foot, (iii) different acclimatisation periods, and (iv) the level of clinician experience and/or comparisons between different health professionals in performing dermal temperature assessments.

This is the first study to assess the reliability of dermal temperature assessments in patients with CN. Infrared dermal thermometry can now be used with confidence in clinical and research settings to provide a reliable assessment of skin temperature in patients with CN, using either a touch or non-touch technique at 10 commonly used testing sites. A non-touch technique, however, was observed to have slightly higher reliability indicating it may be associated with less measurement error than the touch technique.