Laser Doppler flowmetry for assessing localized scleroderma in children

Authors


Abstract

Objective

Assessment of disease activity is a major challenge in the management of children with localized scleroderma. The aim of this study was to evaluate the role of laser Doppler flowmetry (LDF) in comparison with infrared thermography in the detection of scleroderma disease activity.

Methods

In 41 children with localized scleroderma, 111 lesions were assessed on 2 separate occasions, by clinical examination, LDF, and thermography. Measurements from contralateral areas of unaffected skin served as intrapatient controls, and differences in blood flow and temperature were calculated between the corresponding sites. The sensitivity and specificity to detect clinically active lesions were compared between LDF and thermography.

Results

Seventy-five active lesions (34%) and 147 inactive lesions (66%) were identified clinically. The median relative increase in blood flow measured by LDF was +89% (range −69% to +449%) for clinically active lesions and +11% (range −46% to +302%) for clinically inactive lesions (P < 0.001). Thermography showed a median difference in temperature of +0.5°C (range −0.1°C to +4.1°C) and +0.3°C (range −1.9°C to +2.7°C) for clinically active lesions and clinically inactive lesions, respectively (P = 0.024). Using a cutoff level of 39% to indicate increase in blood flow, a sensitivity of 80% and specificity of 77% to detect clinically active lesions were observed; for thermography, no useful cutoff level was identified. The correlation between differences in blood flow and differences in temperature was small, but significant (r2 = 0.120, P < 0.001).

Conclusion

LDF is a helpful, noninvasive diagnostic technique that can be used to discriminate disease activity in children with localized scleroderma, and is more accurate than thermography for this purpose.

Localized scleroderma, or morphea, is a rare connective tissue disorder characterized by hardening and thickening of the skin due to an increased density of collagen (1, 2). The course of localized scleroderma includes an early inflammation stage with hyperemia of the skin, followed by fibrosis, sclerosis, and finally, atrophy. The condition usually begins in childhood and is much more common in the pediatric population than is systemic sclerosis. The disease shows a great variety in its clinical presentation and has been classified into the clinical subtypes of plaque or circumscribed morphea, linear scleroderma, including scleroderma en coup de sabre, as well as generalized morphea profunda (deep), pansclerotic, and combined forms (1, 3).

In general, localized scleroderma is considered to be a condition confined to the skin and subcutaneous tissue and characterized by a benign, self-limiting nature. However, it often involves underlying muscle and bone. Of note, extracutaneous manifestations of the disease can be found in almost one-quarter of affected children (4). At the more severe end of the spectrum, the disease can progress over several years and may cause substantial atrophy, growth retardation, irreversible structural deformities, joint contractures, and severe functional, cosmetic, and psychological disabilities.

In children, the current treatment for severe localized scleroderma is a combination of systemic corticosteroids and low-dose methotrexate (5–8). The management of severe localized scleroderma is challenging, and the detection of disease activity remains a fundamental problem, both in the evaluation of the need for treatment and the assessment of therapeutic efficacy over time. Clinical examination is subjective and sometimes remains unsatisfactory, and laboratory tests are not helpful for this purpose. Reliable and reproducible methods are needed to detect and monitor disease activity.

Thermography has previously been reported to be a helpful tool to assess disease activity in children with localized scleroderma (5, 9, 10). However, a limitation of this technique became evident when thermography was found to yield false-positive results in the assessment of patients with older lesions of localized scleroderma, characterized by marked atrophy of the skin, subcutaneous fat, or muscles (5, 9).

Laser Doppler flowmetry (LDF) is a noninvasive method for the measurement of cutaneous microcirculation and has a broad range of applications in dermatology and microvascular surgery (11–13). It remains unknown whether blood flow levels measured by LDF correlate with clinical findings related to disease activity in juvenile localized scleroderma. In this study we assessed disease activity in a cohort of pediatric patients with localized scleroderma, by clinical examination and, simultaneously, by LDF and infrared thermography. Our aim was to evaluate the role of LDF in detecting disease activity in juvenile localized scleroderma, and to define the sensitivity and specificity of this technique in comparison with infrared thermography.

PATIENTS AND METHODS

Patients.

The study was performed between April 2005 and November 2006 at the Pediatric Dermatology and Rheumatology Units at Great Ormond Street Hospital for Children, in collaboration with the Rheumatology Department at the Royal Free Hospital (London, UK). Children with a minimum age of 2 years were recruited. The diagnosis of localized scleroderma was made clinically by an experienced pediatric dermatologist (JIH) and pediatric rheumatologist (PW). Children with generalized localized scleroderma in whom a right-to-left–side comparison between affected and normal skin was impossible were excluded. Written informed consent was obtained from the children's parents. The study protocol was approved by the local ethics committee.

Study design.

All patients were assessed by the same investigators (LW and KJH) according to a standardized protocol, which included clinical examination, LDF, and thermography. During the study period, every patient underwent evaluations at 2 different time points, performed a minimum of 2 months apart.

All patients were assessed clinically by the same investigator (LW). The subtype of localized scleroderma, the site and extent of the lesions, and current treatment were noted. In every child, 3 measurement points were selected to represent the areas of suspected disease activity, as follows: 1) the border of the affected skin; 2) the areas in which lesions were spreading or had most recently spread; and 3) in the case of a single lesion, 3 different sites across the affected area, or in the case of multiple lesions, 3 different lesions in separate areas.

These selected sites were marked with small arrow-shaped aluminum foil stickers. For single, small lesions in localized scleroderma, as was identified in some children with scleroderma en coup de sabre, the number of measurement points was reduced, as appropriate. Corresponding measurement points were marked at the contralateral, unaffected side of the body, to serve as intrapatient control sites (Figure 1). All marked sites were assessed clinically, followed by thermography and LDF evaluations. At the second study visit, the same measurement points were assessed, with reference to digital photographs of these same sites.

Figure 1.

Representative images from thermography (A) and laser Doppler flowmetry (B–D) of a localized scleroderma lesion on the right forearm of a 10-year-old girl presenting with clinically active linear localized scleroderma. Thermography (A) shows a slight temperature increase along the affected area. Arrows mark the selected and corresponding control sites for the measurements. A laser Doppler flowmeter (Moor Instruments, Axminster, Devon, UK) (C) was used to measure blood flow, with the optical-fiber probe attached to the skin of the right and left forearm (B). The output on the computer screen (D) represents the blood flow, or flux (in arbitrary units), and demonstrates a marked difference between normal skin (right panel of B) and affected skin (left panel of B).

Clinical assessment.

The marked areas of localized scleroderma were described clinically as either active or inactive. Lesions that were spreading and/or showing erythema were defined as active, whereas all others were described as inactive. We noted which of the marked sites represented the most active lesion in a patient. This was defined as the site with the maximum extent of inflammation and/or spreading. In patients with no clinically active lesions, the most recently active area was noted.

Thermography.

The technique of thermography used in this study was carried out as has been previously described (9). Using this technique, we recorded the temperature over an area of 1 × 1 cm at all measurement points and control sites. The thermograms were obtained using the same infrared camera (FLIR SC 500 Thermacam; Flir Systems, West Malling, UK) to assess the skin of each patient (Figure 1).

Laser Doppler flowmetry.

LDF is a noninvasive method for the measurement of cutaneous microcirculation (11–14). The tissue volume assessed has a surface area of ∼1 mm2 and a depth of 1–2 mm, which means that, predominantly, the vasculature of the papillary dermal bed is assessed. We used an MBF3D laser Doppler monitor with a laser wavelength of 810 nm (Moor Instruments, Axminster, Devon, UK). Measurements were performed using an optical-fiber probe attached to the skin with a self-adhesive disc. Blood flow was monitored for 10 seconds at each skin site sequentially. Measurements were performed 3 times, and the mean flow, in arbitrary flux units, was calculated at each site (Figure 1). A maximum of 1 reading per site was discounted to remove “noisy” data attributable to movement artifacts.

Statistical analysis.

The absolute difference in temperature (expressed in °C) between the affected site and the corresponding, contralateral control site was calculated for each measurement point. Blood flow levels tend to show a broad variation among different body areas (15). We therefore calculated the relative difference in blood flow between a lesion and the corresponding control site as follows: [(blood flow of affected site − blood flow of unaffected site)/blood flow of unaffected site] × 100. Results are expressed as the percentage increase (+) or decrease (−) in blood flow.

We analyzed the data from all patients, but also in separate analyses of the children with scleroderma en coup de sabre and those with lesions on the trunk and/or limbs, since patients with en coup de sabre lesions may represent a distinct clinical subgroup in whom clinical signs of inflammation are often lacking despite the presence of active disease. We considered all assessed lesions in each patient (usually 3 sites per patient) as being representative of the overall clinical picture in the patient. Since multiple measurements per patient are not independent from each other, and since the number of assessed sites per patient was sometimes smaller than 3, it could be argued that this introduced bias into the analysis. We therefore undertook a subanalysis of the clinically most active site or the most recently active site in each child, thus analyzing only 1 site per patient.

Results are expressed as the mean ± SD, median (range), or percentage, as appropriate. Comparisons were made with the use of the t-test, Fisher's exact test, or chi-square test, as appropriate. Linear correlations were described by the Pearson's product-moment correlation coefficient (r and r2). The quality of the diagnostic tests was described by the area under the receiver operating characteristic (ROC) curve and its 95% confidence interval (95% CI). In addition, cutoff values were identified in order to obtain the highest sensitivity and specificity values in combined analyses. The null hypothesis was rejected by setting the significance level as a 2-sided P value less than 0.05. All analyses were performed with the use of SPSS for Macintosh OS X (version 11.0; SPSS, Chicago, IL).

RESULTS

Characteristics of the patients.

Forty-five patients with localized scleroderma were enrolled in the study. Four of the children were unable to attend a second study visit, and were therefore excluded from the analysis. A total of 41 patients was included, among whom 111 lesions were identified for analysis. The epidemiologic and clinical features of the patients are summarized in Table 1. Thirty-nine patients had either linear localized scleroderma or a combination with the linear subtype, while 2 patients (5%) had the deep subtype. In 18 patients (44%), >5% of the total body surface area was affected. The mean ± SD disease duration at the time of study entry was 5.1 ± 3.5 years. The mean ± SD time interval between the 2 study visits was 7.9 ± 5 months. Thus, a total of 222 affected skin lesions was assessed over the study period, and each was compared with its respective control site.

Table 1. Main characteristics of the 41 patients with localized scleroderma
  • *

    Patients were started on treatment with systemic corticosteroids and methotrexate (MTX) after the first study assessment.

  • Having stopped treatment with systemic corticosteroids and MTX.

Female-to-male ratio2.2:1
Disease subtype, no. (%) 
 En coup de sabre18 (44)
 Linear of trunk/limbs19 (46)
 En coup de sabre combined with linear of trunk/limbs2 (5)
 Deep2 (5)
Age at disease onset, years 
 Mean ± SD5.1 ± 2.9
 Median (range)4.5 (1.5–11.6)
Age at study entry, years 
 Mean ± SD10.2 ± 3.0
 Median (range)10.0 (4.7–15.7)
Treatment at study entry, no. (%) 
 MTX16 (39)
 Systemic corticosteroids and MTX6 (15)
 No treatment19 (46)
  Pretreatment*16 (39)
  Off treatment3 (7)

Findings of clinical assessment, LDF, and thermography.

Clinical examination of the skin sites revealed 75 active lesions (34%) and 147 inactive lesions (66%). The median relative increase in blood flow measured by LDF was +89% (−69% to +449%) for clinically active lesions and +11% (−46% to +302%) for clinically inactive lesions (P < 0.001) (Figure 2A). Thermography showed a median difference in temperature of +0.5°C (−0.1°C to +4.1°C) for clinically active lesions and +0.3°C (−1.9°C to +2.7°C) for clinically inactive lesions (P = 0.024) (Figure 2B). Table 2 demonstrates the differences in blood flow and temperature for the clinical subgroups of patients with en coup de sabre lesions and those with other types of localized scleroderma lesions.

Figure 2.

A, Relative difference in blood flow for clinically active lesions and clinically inactive lesions (n = 222; P < 0.001). B, Absolute difference in temperature for clinically active lesions and clinically inactive lesions (n = 222; P = 0.024). Data are shown as box plots. Each box represents the 25th to 75th percentiles. Lines outside the boxes represent the range (without outliers and extremes). Lines inside the boxes represent the median.

Table 2. Differences in blood flow and temperature between clinically active and clinically inactive lesions in patients with and those without scleroderma en coup de sabre lesions
 Relative difference in blood flow, median (range) %Absolute difference in temperature, median (range) °C
En coup de sabre lesions (n = 94)  
 Clinically active (n = 28)+105 (+14 to +331)+0.7 (−0.2 to +2.8)
 Clinically inactive (n = 66)+11 (−46 to +302)+0.4 (−1.9 to +2.7)
 P<0.0010.082
Lesions other than en coup de sabre (n = 128)  
 Clinically active (n = 47)+86 (−69 to +449)+0.4 (−1.0 to +4.1)
 Clinically inactive (n = 81)+10 (−45 to +219)+0.3 (−1.9 to +2.6)
 P<0.0010.136

The area under the ROC curve for the relative difference in blood flow, as a measure of the ability of LDF to detect clinically active lesions, was 0.80 (95% CI 0.73–0.87; P < 0.001), as shown in Figure 3A. A cutoff level of 39% as an indicator of increase in blood flow had a sensitivity of 80% and specificity of 77% to detect clinically active lesions. In contrast, the area under the ROC curve for absolute differences in temperature, as a measure of the ability of thermography to distinguish between clinically active and clinically inactive lesions, was only 0.59 (95% CI 0.52–0.67; P = 0.025). A cutoff level of 0.5°C as an indicator of increase in temperature had a sensitivity of 52% and specificity of 58% to detect clinically active lesions. Table 3 shows the details of the ROC curves for the differences in blood flow and differences in temperature in patients with en coup de sabre lesions and those with other types of localized scleroderma lesions.

Figure 3.

A, Receiver operating characteristic (ROC) curve for the relative difference in blood flow assessed by laser Doppler flowmetry in all lesions (n = 222). The area under the ROC curve was 0.80 (95% confidence interval [95% CI] 0.73–0.87, P < 0.001). B, ROC curve for the subanalysis of the relative difference in blood flow for the most active or most recently active lesions (n = 82). The area under the ROC curve was 0.88 (95% CI 0.80–0.96, P < 0.001).

Table 3. Details of receiver operating characteristic curves for the differences in blood flow and temperature in patients with and those without scleroderma en coup de sabre lesions*
 Relative difference in blood flowAbsolute difference in temperature
AUC (95% CI)PAUC (95% CI)P
  • *

    Values are the area under the curve (AUC) (95% confidence interval [95% CI]). P values are in comparison with the AUC values of the other group for each measure.

En coup de sabre lesions (n = 94)0.88 (0.81–0.95)<0.0010.62 (0.49–0.74)0.063
Lesions other than en coup de sabre (n = 128)0.75 (0.65–0.85)<0.0010.58 (0.48–0.68)0.136

Correlation between differences in blood flow and differences in temperature.

When comparing the differences in blood flow and differences in temperature, we found a small, but significant correlation between the 2 measures, with a Pearson's correlation coefficient of 0.447 (n = 222; r2 = 0.120, P < 0.001). The correlation between the LDF findings and the thermography findings did not markedly improve when calculated for the following subgroups: clinically active lesions (n = 75; r = 0.354, r2 = 0.125, P = 0.002), clinically inactive lesions (n = 147; r = 0.268, r2 = 0.072, P = 0.001), scleroderma en coup de sabre lesions, including combined forms (n = 94; r = 0.254, r2 = 0.065, P = 0.018), and localized scleroderma lesions other than the en coup de sabre subtype (n = 128; r = 0.405, r2 = 0.164, P < 0.001).

Subanalysis of a single lesion per patient.

In the subanalysis involving the clinically most active site or most recently active site in each patient, we measured a total of 82 lesions in 41 patients, in comparison with the respective control site. Blood flow assessed by LDF remained significantly higher in active lesions (n = 43) than in inactive lesions (n = 39) (median relative difference in blood flow +100%, range +2% to +449% in active lesions versus +13%, range −34% to +288% in inactive lesions; P < 0.001). However, the subanalysis assessment by thermography did not show any significant difference between clinically active lesions and clinically inactive lesions.

In this subanalysis, the area under the ROC curve for the LDF findings was slightly larger (0.88, 95% CI 0.80–0.96) than that described for all 222 lesions (P < 0.001) (Figure 3B). In contrast, the area under the ROC curve for the thermography findings was even smaller (0.48, 95% CI 0.35–0.61) than that described for all 222 lesions (P = 0.066). A cutoff level of 45% for the relative difference in blood flow had a sensitivity and specificity of 86% and 85%, respectively. In this subanalysis, there was no correlation between differences in blood flow and differences in temperature (P = 0.076). Moreover, the results of this subanalysis did not differ when patients with and those without scleroderma en coup de sabre lesions were analyzed separately (results not shown).

DISCUSSION

This is the first study to evaluate LDF as a tool to detect disease activity in a large cohort of children with localized scleroderma. Blood flow measured by LDF was significantly increased in clinically active localized scleroderma lesions. We found that LDF had a sensitivity of 80% and specificity of 77% to detect active lesions, if the blood flow was found to be increased by at least 39%. The subanalysis of only 1 lesion per patient showed an even greater sensitivity and specificity of LDF (86% and 85%, respectively). Thermography, in contrast, was not helpful in the detection of disease activity in the children in this study.

The clinical presentation of localized scleroderma is varied. Some lesions, particularly the scleroderma en coup de sabre type, often do not show any obvious clinical signs of inflammation, although the disease remains active and continues to progress. In this situation, as well as during followup of treatment, it is often difficult to determine whether a lesion is active, and this cannot be determined by clinical examination only.

LDF is a noninvasive method for the measurement of cutaneous microcirculation. It has been used to investigate a range of medical conditions, such as inflammatory skin disorders, vasospastic vascular disorders, neuropathies, the response of microcirculation to neurotransmitters, tumors, ulcers, and burns, and has been used to monitor the microcirculation in skin flaps or grafts following intestinal, orthopedic, or plastic surgery procedures (11–13). The use of LDF for the assessment of localized scleroderma in adults has been described in 2 small studies and 1 case report (16–18). Serup and Kristensen (16) performed LDF in 15 adults with localized scleroderma, and Kalis et al (17) applied this method in 16 adults with localized scleroderma. As in our study, those authors found blood flow to be increased in localized scleroderma lesions. However, they did not describe levels of blood flow in relation to clinical disease activity.

Serup and Kristensen (16) performed LDF measurements in the sclerotic center of plaques, as well as in perilesional areas, and found blood flow to be elevated in both sites. In our study, we limited the assessment to the borders of lesions, in order to standardize the measurements. Kalis et al (17) assessed blood flow in the center of localized scleroderma plaques, and distinguished between clinically progressive, static, and resolving lesions. Those authors found increased blood flow in equal proportions in all groups (mean +40% [±SD 10%]), whereas in our study, blood flow was significantly higher in the clinically active lesions, by a median of +89%. However, compared with normal skin, we also recorded a slight increase in blood flow in clinically inactive lesions. It remains unclear whether this is clinically relevant.

LDF values do not show any relevant intraindividual differences between corresponding body sites in healthy subjects (15). It is possible that the slightly higher blood flow in clinically inactive lesions reflects persistent changes in the microcirculation caused by disease progression. However, it is more likely that some inflammation is ongoing in a proportion of these lesions, and not detectable by clinical examination alone. In fact, sudden reactivation of clinically stable localized scleroderma is a well-recognized characteristic of the disease in childhood (5). This observation highlights the fundamental difficulty in assessing disease activity in localized scleroderma, and thus reiterates the need for reliable and reproducible methods. Histopathologic investigations represent the gold standard, but repeated skin biopsies are not an appropriate option to monitor localized scleroderma in children.

The results of this study suggest that LDF is an important adjunct to clinical examination for the purpose of detection of disease activity. Prospective studies are needed to further investigate the potential prognostic value of LDF in localized scleroderma. LDF can be used to investigate only a small skin area, and measurements are susceptible to motion artifacts. The more recently developed technique of laser Doppler imaging (LDI) allows evaluation of blood flow over a specific area of skin while avoiding contact with the skin surface (11). Nevertheless, current LDI devices are of limited use in children, because of their slow scanning times. With further development, LDI may become a valuable technique.

In the present study thermographic imaging did not prove to be helpful in distinguishing between active and inactive localized scleroderma lesions. This is in contrast to the findings in previous reports from our institution and others (5, 9, 10). Martini et al (9) reported a sensitivity and specificity of thermography of 92% and 68%, respectively, to detect active lesions. In that retrospective study, thermograms were visually assessed, and a lesion was considered thermography positive if it appeared at least 0.5°C warmer than the matching opposite body site. This is different from the method used in the present study, in which we limited the area of thermography to a small site within a localized scleroderma lesion and measured the actual temperature in °C instead of making a global visual judgment.

We suspect that the main reason for the thermography results obtained in the present study is that considerable numbers of long-standing localized scleroderma lesions were included. Previous studies showed that thermography produced false-positive results in patients with older localized scleroderma lesions, in whom marked atrophy of skin, subcutaneous fat, and muscle was present (5, 9). In these situations, positive findings on thermography probably represent increased heat conduction from deeper tissues, rather than active inflammation in the affected skin. This mechanism is likely to have contributed to the poor correlation between the results of LDF and those of thermography as seen in the present study. It was previously shown, in healthy volunteers, that skin temperature does not correlate well with cutaneous blood flow, which supports the notion that thermography is not a good measure of skin perfusion (14).

Although the precise childhood incidence is unknown, juvenile localized scleroderma is a rare condition, and our single-center analysis involved a large cohort of 41 children. A limitation of this study is the lack of investigator blinding. However, this is unlikely to have influenced our results, since the techniques of LDF and thermography are essentially operator independent (15). In this study all patients were assessed consistently by the same investigator, and we did not evaluate any interobserver variability. More studies are needed to identify a gold standard for the detection of disease activity in children with localized scleroderma, particularly a method that would be suitable to utilize in the clinical routine.

Our results thus show that LDF is a valuable, noninvasive diagnostic technique that can be used to discriminate disease activity in localized scleroderma, and is more accurate for this purpose than thermography. A prospective study in which we are evaluating the application of LDF in monitoring disease progression, response to treatment, and the relationship between changes in blood flow and skin structure in children with localized scleroderma is ongoing. By integrating spatial and flow data, the newer method of LDI represents an interesting technique for future study.

AUTHOR CONTRIBUTIONS

Dr. Weibel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Weibel, Howell, Visentin, Denton, Zulian, Woo, Harper.

Acquisition of data. Weibel, Howell, Visentin.

Analysis and interpretation of data. Weibel, Howell, Visentin, Rudiger, Denton, Woo, Harper.

Manuscript preparation. Weibel, Howell, Visentin, Rudiger, Denton, Zulian, Woo, Harper.

Statistical analysis. Weibel, Visentin, Rudiger.

Ancillary