Infrared thermography in children: a reliable tool for differential diagnosis of peripheral microvascular dysfunction and Raynaud’s phenomenon?

Patients with age less than 16 years visiting our Paediatric Rheumatology outpatient clinic with personal history of episodes of finger discoloration upon cold exposure and undergoing thermography of their hands entered a cross-sectional study. PRP was defined when episodic reversible bi or-triphasic colour changes in the extremities were not associated with established or suspected connective tissue disease (CTD). SRP was termed when a defined diagnosis of CTD was present at moment of IRT examination. Acrocyanosis was defined when discoloration of extremities was referred as symmetric and persistent. Demographic data, autoantibodies profile and capillaroscopy results were collected. Exclusion criteria were presence of any skin or joint alteration potentially interfering with thermal analysis and any ongoing treatment for RP. The control group consisted of healthy age-matched subjects with absent or mild vascular disturbance on extremities, such as cold fingers after cold exposure but without colour changes. IRB approval was not needed as IRT examination is used in our standard assessment of patients with RP and the present study included only analysis of temperature data. Indeed, written informed consent was obtained from parents of all subjects taking part in the study.

All subjects were asked not to smoke or consume hot or caffeine-containing beverages as well as take drugs or make physical exercise for at least 4 h prior to the test. Thermographic images were acquired with the same infrared camera (ThermaCAM PM695, FLIR systems AB, Stockholm, Sweden) by same examiner (GM), as previously described [6]. After initial acclimatization of each patient in a temperature-controlled room (at 23 ± 2 °C) for 20 min, thermographic images were taken of the dorsal aspect of both hands (pre-test). The subject then put on latex gloves and placed his/her hands to the metacarpophalangeal (MCP) joints into 15 °C water for 1 min. Gloves were worn for the cold challenge to avoid evaporative cooling and removed immediately for re-warming, then thermographic images of the dorsal aspect of both hands were recorded at one-minute intervals for 10 min during re-warming (time T₀ to T₁₀ of the test).

Thermographic images were independently and blindly evaluated by two physicians with low or null experience in IRT (MC and RC) and no specific training was performed. For each image acquired before and during the test (pre-test, T₀ to T₁₀) the following data were measured: mean temperature at dorsum of MCP and distal interphalangeal joints (DIP) of the II, III, IV and V finger of both hands as showed in Fig. 1a.

From these measurements the temporal variations of mean temperature at MCP and DIP joints were calculated for each finger in order to evaluate and compare the re-warming patterns. Temporal variation 1 (ΔT₁) was calculated by deducting temperature at time n from temperature at time n-1 (T_n-T_n-1), with n being the minute after cold challenge. Temporal variation 2 (ΔT₂) was calculated by subtracting the temperature at time n from temperature pre-test (T_n- T_pre).

The distal-dorsal difference (DDD) was calculated by subtracting the mean fingertip temperature (DIP) from the mean temperature at dorsum (MCP), as previously described [17, 18]. Therefore, if the fingertip was colder than the dorsum, the DDD was positive (Fig. 1b).

Inter-rater reliability was determined by Intraclass Correlation Coefficient (ICC) and results were interpreted as follows: ICC values range 0.75–1 excellent reliability, 0.4–0.74 good reliability, < 0.4 poor reliability.

Forty-four patients and 15 age-matched controls were included in the study. Sixteen patients were affected by SRP, 14 by PRP and 14 by acrocyanosis (AC). Mean age was 11.9 years (3.6–16) in SRP, 12.2 years (6.1–15.6) in PRP and 14.2 years (6–16.6) in AC patients. Patients with SRP were affected by dSSc (10 patients), lSSc (3), MCTD (2), Overlap syndrome SLE/SSc (1). Mean disease duration was 3.3 years in SRP, 2.2 years in PRP and 1.5 years in AC patients. All SRP and PRP patients were currently reporting RP attacks, while digital ulcers had occurred in 6/16 SRP and in 2/14 PRP patients but none was active at the time of cold challenge. Thirteen patients with SRP were taking treatment (12 calcium channels blockers, 1 ACE inhibitors), 6/14 PRP patients were taking calcium channels blockers.

In SRP patients nail fold capillaroscopy showed scleroderma pattern active in 8 (50%) patients, late in 3 (18.8%), early in 1 (6.3%) and non-specific abnormalities in 4 (25%); in PRP patients showed no abnormalities in 8 (50%) and non-specific findings in 8 (50%), while in AC patients showed acrocyanosis pattern in 8 (mildly reduced capillary density and presence of dilated capillaries). The mean age of healthy controls was 12.4 years of age (8.5–15.8) and 5 of them reported to have “cold hands” but never colour changes. Two patients in the PRP group had low titre (< 1/160) positive antinuclear antibodies but none of them developed any clinical feature of CTD so far. Characteristics of subjects are summarized in Table 1.

Each examiner independently and blindly rated a set of 192 measures for each patient and control as II, III, IV and V fingers of both hands were evaluated at MCP and DIP joints at pre-test time and at T₀ to T₁₀ after cold challenge. All IRT examinations were performed in morning hours and without significant differences in seasonal distribution of execution of the procedure between the groups.

The inter-rater agreement for temperature measurement at DIP joints was excellent with mean ICC value 0.952 (0.942–0.962) for patients and 0.943 (0.936–0.950) for controls. Similarly, an almost complete agreement between examiners was observed for temperature measurements at MCPs as the mean ICC was 0.955 (0.947–0.964) in the group of patients and 0.945 (0–939-0.951) for controls.

The mean basal temperature at both MCP and DIP joints was significantly lower in patients with PRP, SRP and even more with acrocyanosis compared to controls (p < 0.001), as reported in Table 2. Moreover, analysis of DDD showed that, at baseline, patients with PRP presented higher temperature at DIPs and lower at MCPs compared with those with SRP and AC and therefore in PRP the DDD values were significantly higher (p < 0,001).

The analysis of temperature temporal variations showed that IRT was able to clearly differentiate patients (PRP and SRP and acrocyanosis considered together) from controls. In fact, the re-warming pattern was significantly slower in patients’ group as showed by analysis of ΔT₁ in which controls presented gain of basal temperature significantly earlier at MCPs, but even more at DIPs (p < 0.05) (Fig. 2a and b). This different trend was more evident in the comparison of ΔT₂, with healthy controls reaching higher temperatures and more rapidly than patients both in MCPs and DIPs (p < 0.001) as showed in Fig. 2c and d, respectively.

The analysis of re-warming pattern showed that patients with PRP and SRP significantly differed from AC particularly looking at ΔT₂ temporal variation. Indeed, subjects with both PRP and SRP presented some gain of temperature over time particularly at DIPs and this allowed PRP, but not SRP patients, to achieve the basal temperature by the end of the re-warming time. Inversely, in patients with AC the fingertips temperature after cold challenge showed only null or minimal changes over time. (Fig. 3a-d).

Furthermore, the analyses of DDD confirmed that patients with PRP, SRP and acrocyanosis significantly differed from healthy controls as for fingers temperature recovery pattern, as showed in Fig. 4a and b. After cold challenge, in PRP patients the fingertips were initially colder than dorsum but difference progressively reduced during the re-warming phase, achieving higher values than basal, as in fact DDD values became negative after 10 min. In SRP patients, the DIPs temperature, after a transitory reduction, gradually increased and DDD got back to the baseline values at the end of the test. Conversely, in acrocyanosis the DIPs temperature increased very slowly over the whole re-warming period (p < 0.05).