Abnormalities of both vascular structure and function are well recognized in systemic sclerosis (SSc). Abnormalities include endothelial impairment, lumen narrowing, and fibrosis of the microvasculature (1–3). The most characteristic clinical expression of this vascular dysfunction/damage is severe Raynaud's phenomenon (4). Recent investigations have indicated that this damage may be responsible for the abnormal production and regulation of vasodilators and vasoconstrictors such as cytokines, growth factors, and endothelin 1, and that these may in turn be the cause of increased collagen production and fibrosis (5, 6).
Laser Doppler imaging (LDI) and single-point laser Doppler flowmetry (LDF) are both sensitive methods of measuring cutaneous blood flow in patients with SSc and help to differentiate between primary (idiopathic) Raynaud's phenomenon and Raynaud's phenomenon secondary to SSc, as compared with healthy controls (7, 8). These techniques are particularly useful when monitoring a patient's response to dynamic stimuli that cause significant changes in blood flow, such as occlusion, heating, cooling, and iontophoresis (9–12). In addition, both methods can be used to monitor treatment response (13–16).
LDI and LDF utilize the Doppler effect, where changes in wavelength due to movement between the source (in this case, the red blood cells, from which light is back scattered) and observer are used to determine the speed of blood flow. The laser Doppler technique was first applied by Stern in 1975 to quantify changes in blood flow (17) and has since been used extensively. LDI involves perfusion mapping of areas, rather than examination of blood flow at a single point as with LDF; by taking into account blood flow in adjacent points within an image and performing some averaging techniques, LDI is more likely than LDF to give a clinically meaningful image of perfusion, since, due to the heterogeneity of cutaneous blood flow, LDF is subject to large fluctuations as a result of minute changes in the probe position (e.g., because of movement artefacts) (18). However, LDI scans take a finite time to acquire, and therefore cannot provide the continuous monitoring that is possible with LDF.
In both the LDI and the LDF method, blood flow is represented by flux (flow through an area) and is measured in arbitrary perfusion units, which are directly proportional to the product of the mean speed and concentration of red blood cells. In LDI, an image of relative blood flow is built up point by point as the laser scans over an area of skin, whereas in LDF, relative blood flow at a single site of interest is plotted over time. For our study, both techniques were noncontact, although some LDF systems used in previous studies require laser delivery via an optical fiber, and therefore require skin contact. Of the laser Doppler studies involving patients with SSc, very few have utilized LDI (most have used single-point monitoring with LDF).
In this study, we describe the use of a novel dual-wavelength LDI system in the investigation of the microvascular abnormalities of SSc. We provide only a brief background describing this new system, since it has been described in detail elsewhere (19). The majority of LDI studies have been carried out at a single wavelength, which images blood flow in certain types of vessels depending on the penetration and absorption of the laser wavelength. Green, red, and infrared wavelengths have penetration depths of ∼0.3, 0.5, and 1.0 mm, respectively, in excised skin (20), and it has been shown, in in vivo studies with dual-wavelength LDF, that longer wavelengths allow for examination of deeper perfusion (21).
Within the skin, in the absence of melanin, it is hemoglobin that is the main visible light absorber, particularly of shorter wavelengths. In the skin, which is relatively avascular and transparent to light, green light is approximately twice as likely to be absorbed as red; however, in hemoglobin, green light is 20 times more likely to be absorbed (21,22). Light that enters the dermis (i.e., that is not absorbed by melanin in the epidermis) can be assumed to penetrate until it reaches a blood vessel, at which point the light is either scattered or absorbed. Upon scattering, the light will then either undergo further scattering, perhaps back toward the surface of the skin and a detector, or be absorbed. Green light that reaches the detector can be thought of as having come from smaller microvessels, such as capillaries, rather than a larger, thermoregulatory-type microvessel, in which it is more likely to be absorbed due to the higher chance of interaction with hemoglobin, although it may be scattered from the peripheral flow of a larger microvessel. Red laser light is less likely to be absorbed by hemoglobin, and it is more likely to be back scattered to the detector from larger microvessels (both from peripheral and from axial flow), having passed through the capillaries without significant interaction, although some of the signal will originate from the nutritive capillaries and other small microvessels encountered in the upper layers of the dermis. Therefore, it can be assumed that images obtained by green laser are mostly of small microvessels, whereas images obtained by red laser are of larger microvessels (19,23).
The advantage of a dual-wavelength system is the ability to image different components of the microcirculation. Dual-wavelength LDF and LDI therefore hold promise for application in SSc and are now being used to study other diseases in which microcirculatory abnormalities occur (22,24,25). Our novel red (633 nm) and green (532 nm) dual-wavelength system enables measurement of perfusion in both larger, thermoregulatory vessels and smaller, nutritive capillaries (19). The present study was undertaken to investigate the hypothesis that cutaneous microvascular perfusion in the dorsum of the hand and in the digits is impaired in patients with SSc, as quantified by the response to the dynamic stimuli of local heating and occlusion, respectively, when compared with a healthy control group.
- Top of page
- PATIENTS AND METHODS
Our choice of dynamic tests was informed by the results of previous studies, which have shown that both local heating and occlusion of the digits are reproducible techniques that result in distinct, dynamic responses in blood flow while also offering a practical option in terms of time constraints.
As was observed in our previous study (in healthy control subjects), the maximum increase and AUC for the perfusion–time curve were both lower for data from scans with the green wavelength (19). This indicates that vessels imaged by the red laser experience a larger change in perfusion due to heating and take longer to return to baseline than the capillaries imaged by the green laser. The difference between the values of peak blood flow imaged by the 2 wavelengths is most likely due to the mechanism of perfusion control; arterioles dilate and therefore undergo an increase in both the speed and the volume of blood flow, whereas capillaries, which do not have smooth muscle to allow them to vasoconstrict and vasodilate, have precapillary sphincters to control flow. Consequently, the capillaries do not have the capacity to undergo the same increase in blood flow as that of larger vessels.
No difference was found between the healthy controls and SSc patients in the local heating data (AUC, PEAK1, and PEAK2 values) obtained with either wavelength. Both peaks were examined because they reflect different aspects of vasodilation. It has been observed in several studies that during sustained local heating, vasodilation is mediated by 2 time-dependent mechanisms (27, 29). During the first 5 minutes of heating, activation of local sensory nerves is responsible for the rapid increase in perfusion forming the main peak (PEAK1). The smaller, second peak (PEAK2) has been observed to occur when heating is continued beyond 10 minutes, since the heat is conducted to deeper levels of the skin. This second peak is dependent on nitric oxide (NO) production. Although the protocol for this study involved heating for only 6 minutes, it was apparent from the repeat scans that by 10 minutes after the start of heating, there was still a significant increase in blood flow within the heated area. It can be assumed that an increase in temperature, from baseline, accompanies this increase in perfusion. This is a consequence of the fact that although the skin's thermal relaxation time, for a pulsed heat source (such as a laser), is in the order of microseconds to seconds (30) (depending on the source and location of heating), the rate of temperature increase in the skin is too high for the blood to be able to convect heat away from the interaction site at the same rate, and therefore it accumulates until heating is ceased. Thus, the presence of PEAK2 is due to the fact that the skin is still experiencing an increase in temperature after the heating pad has been removed.
The second peak (PEAK2) was of interest due to the recent hypothesis that NO levels are elevated or reduced in patients with SSc when compared with healthy controls; the direction of change in NO levels is most likely dependent on the duration and severity of disease (31–34). However, no statistically significant differences in PEAK2 values were found between patients and controls. This could be due to the fact that subjects were not maximally vasodilated due to curtailed heating, and therefore the effect could not be differentiated.
Since there were no significant differences between the patient and control groups in the peaks of the perfusion curve or the AUC, it can be concluded that microvasculature function as assessed at the dorsum of the hand is not affected by the SSc disease process (in contrast to the fingers). This is consistent with the findings of several other studies in which the dorsum of the hand or other more proximal sites were examined with blood flow monitoring, and no difference in blood flow was observed between groups (35–37). Albrecht et al (35) used local heating on the dorsum (proximal to the digits) and found no difference in the flux values following heating in SSc patients compared with controls. Geirsson et al (36) investigated the effect of local heating in 5 locations of the body, comprising the forehead and sites on the wrists and ankles bilaterally. No difference was found between the control group and patients with SSc. They concluded that the function of the precapillary sphincters and the smallest vessels in the SSc patients was preserved. Walmsley and Goodfield (37) used LDF to study the increase in perfusion during dorsal foot warming in patients with SSc; although the baseline perfusion before heating was significantly higher in the control group, and patients were found to have a lower threshold for vasodilation, no significant difference in blood flow was found at the increased temperature (36°C) in patients with SSc compared with control subjects. In contrast to these studies, Wollersheim et al (38) found that blood flow in the distal digits (assessed with LDF) during whole-hand heating was significantly reduced in patients with SSc compared with controls. Therefore, the results from these previous studies are consistent with our findings.
By occluding the finger rather than performing brachial occlusion, it is possible, in addition to measuring hyperemic response, to easily image the area being occluded with well-defined boundaries (19). The decrease in blood flow was more pronounced, both in patients and in controls, in the vessels imaged by the red wavelength than in those imaged by the green wavelength. As discussed above, this finding is most likely explained by the fact that capillaries do not have the capacity to dilate and constrict in the same way as larger, thermoregulatory vessels, because their mechanism for increasing or decreasing blood flow is reliant on sphincters and one-way valves. The residual Doppler signal under occlusion is the biologic zero and is attributable to several factors, including collateral circulation through bone, movement of fluid within the interstitial space, Brownian motion of red blood cells, and residual vasomotion (39).
No significant difference in blood flow during occlusion was found between patients and controls on images with the red wavelength, whereas a difference was found on the green-wavelength images. SSc affects primarily the microvasculature (1), and it is therefore a fair assumption that the smaller vessels are more sensitive to challenges that cause disturbances to microvascular flow and compromise it further. This result indicates that the smaller vessels are more impaired than the larger ones, which corresponds with the observations in a study carried out by Kristensen (40), who examined vascular resistance in the digits of patients with SSc, using the Xe washout technique. In response to occlusion, he found that vessels in the subdermis, i.e., larger vessels, were more reactive than those in the dermis, i.e., small vessels, indicating that small vessels were more impaired than larger vessels. In addition, he found that vessels in the hand were more reactive than those in the fingers, further supporting the suggestion of the digit vessels being more impaired than the dorsal vessels. Stucker et al (41) found, with digit subtraction angiography, that 90% of patients with SSc had arterial narrowing in the fingers, 31% had narrowing of the ulnar artery, and none had narrowing of the radial artery. These findings, together with our own, suggest that in SSc there are abnormalities of the digit microvasculature and also of the larger vessels. It is therefore likely that both large- and small-vessel disease is responsible for digit ischemia in SSc.
Our observation of a significant decrease in PEAK/OCC values in the SSc patients compared with the controls, using both wavelengths, indicates that SSc affects both small vessels such as nutritive capillaries and larger vessels such as those in the thermoregulatory circulation. This is consistent with the findings of other studies that also identified a reduction in hyperemic response. Using LDF, both Albrecht et al (35) and Wigley et al (9) found a significantly decreased response to postocclusive hyperemia in patients with brachial and digit occlusion. Wigley et al also performed occlusion following administration of a vasodilator; no significant improvement was observed, and thus they concluded that structural defects limit microvascular flow, causing dysfunction. Wollersheim et al (42), also using LDF to measure the occlusion response in the fingers, found a lower baseline and lower peak response, but concluded that since the percentage changes in both patients and controls were not significantly different, the peak increase was directly related to the baseline value of flow. Similarly, Goodfield et al (43), using digit arterial occlusion, found that the decreased response in patients occurred when resting blood flow was low, and that the absence of a hyperemic response could be alleviated by prewarming the patients' hands, after which responses compared well with those in controls; they concluded that this was attributable to abnormal vessel wall reactivity. Kristensen (40) concluded that decreased response to occlusion and abnormal hemodynamics were due to a local fault in the vasculature.
We found that 1) with local heating of the dorsum (monitored using LDI), there was no difference between groups in either the small (e.g., capillaries) or large (e.g., thermoregulatory) microvessels, 2) before and during occlusion (monitored using LDI), there was a significant difference between groups in the small microvessels, and 3) with regard to the hyperemic response on release of occlusion (monitored with LDF at the fingertip), there was a difference between groups in both the smaller and the larger microvessels.
It should be recognized that with dual-wavelength LDI, certain assumptions are made regarding the types of vessels that are imaged with each wavelength. However, these assumptions are based on spectral and penetration depth data from many scientific studies in the field of laser photomedicine, as discussed above. Tulevski et al and Ubbink et al (44, 45), whose studies compared near-infrared and green-wavelength LDF to capillary microscopy, obtained significantly different results between the 2 wavelengths, suggesting that the green wavelength measures more superficial flow. Although they concluded that the green wavelength did not exclusively examine superficial capillary perfusion, it should be noted that our previous study (19) showed that the type of vessel imaged was dependent on the power of the laser used; therefore, the situation is complex.
We believe these results warrant further studies to examine the microvascular responsiveness in patients with SSc. These studies could include simultaneous heating (using standard stimuli) at several sites (including the finger, dorsum, and forearm) and should assess correlation of the results with the degree of skin thickening. Although in our study we found no association between the degree of skin thickening of the dorsum of the hand (as assessed by the modified Rodnan skin score ) and hyperemic heating response (results not shown), examination of a larger number of different sites would be required to investigate this more fully. Avascular skin is relatively transparent to light, since there are no highly absorbing molecules present, even if thickened. Therefore, the effect of absorption or scattering in blood is much more relevant than thickness or fibrosis of skin in monitoring blood flow response. Thus, if future studies reveal any association between the degree of scleroderma and perfusion, this could be attributable to the clinical implications of fibrosis on blood flow responses, as opposed to increased optical absorption by the skin.
In summary, no differences in the dorsal heating response were found between groups, whereas differences were found in the finger occlusion response, suggesting that microvascular impairment in SSc may be selective to the digits. The additional observation that the postocclusive hyperemic response was decreased using both the green and the red wavelengths in SSc patients (even though, under occlusion, differences were observed only on green laser wavelength images) indicates that larger vessels are also involved, but to a lesser degree.
As discussed above, several other heating and occlusion laser Doppler studies have demonstrated differences between patient and control groups in the digits but not in the dorsum of the hand. Iontophoresis studies examining endothelial-dependent and -independent responses to vasodilation similarly have provided evidence to support this theory. In one study, forearm iontophoresis produced no differences between the control and patient groups (47), whereas similar studies with digit iontophoresis showed impaired vasodilation in the patient group (12, 48, 49). This selective microvascular impairment is likely to contribute to the severe Raynaud's phenomenon that is characteristic of the SSc disease process. Therapies should therefore be targeted specifically to the digit vasculature.