 |
|
Why Does Carbon Dioxide Resurfacing Work?
A Review
E. Victor Ross, MD;
Joseph R. McKinlay, MD;
R. Rox Anderson, MD
Arch Dermatol. 1999;135:444-454.
ABSTRACT
Despite the unquestionable efficacy of carbon dioxide laser skin resurfacing, mechanisms for cosmetic enhancement remain poorly characterized. Histological studies have provided some insight into the cascade of events from initial laser impact to final skin rejuvenation. However, there are few comprehensive studies of gross and microscopic wound healing. Additionally, the literature is fragmented; excellent individual articles appear in journals from widely disparate disciplines. For example, some reports relevant to laser skin resurfacing are "sequestered" in the engineering literature. This article is intended to update the physician on laser skin resurfacing based on the broadest review of the current literature. It proceeds from a discussion of initial laser-tissue interactions, such as collagen denaturation, to examination of long-term biological sequlae. At some cost to scientific rigor, mathematical models describing laser-tissue interactions are not presented.
INTRODUCTION
Short-pulsed, high-energy, rapidly scanned carbon dioxide laser skin resurfacing (LSR) has become an accepted treatment for cosmetic rejuvenation. By restricting residual thermal damage (RTD), predictable results can be achieved.1-5 Previously, continuous-wave lasers combining lowpower densities (approximately 50 W/cm2) and gated exposures were reported as effective in wrinkle reduction. However, efficacy and safety required strict adherence to the following conditions: (1) low total fluences (2-5 J/cm2), (2) only 1 pass being made, and (3) high operator skill (using magnification loupes and precise clinical end points, ie, slight graying of the epidermis).6
Despite the popularity of short-pulsed and rapidly scanned carbon dioxide lasers, the mechanisms for cosmetic improvement are poorly understood. Initially it was proposed that tissue removal (ablation) was the most important mechanism in LSR.7-11 This concept was embraced because of its intuitive attractiveness and the observation that wrinkle crests could be gradually sculpted down by multiple passes of the laser.11 It was later determined that ablation depths into the dermis were often much less than wrinkle depths, and yet marked cosmetic improvement was still observed.12-13 Also, microscopic studies failed to show significant dermal ablation (>50 µm), even after multiple passes.14-18 This led to the proposal that other mechanisms, namely, collagen shrinkage and wound healing–induced fibroplasia, were mainly responsible for cosmetic improvement.
At Massachusetts General Hospital, in Boston, we have examined the physics and biological sequelae of LSR, primarily through the use of live farm pigs, but also through clinical studies. The array of lasers in these studies has included microsecond-domain (TruPulse; Tissue Technologies, Albuquerque, NM) and millisecond-domain (Ultrapulse; Coherent, Palo Alto, Calif) high-energy pulsed carbon dioxide systems; a continuous-wave carbon dioxide laser scanning a microbeam (Sharplan, Allendale, NJ), and an erbium:YAG system (Con Bio, Dublin, Calif).13, 19-24 The following review is based on our data as well as published material from the disciplines of physics, collagen chemistry, dermatology, and wound healing. The first section examines physics of resurfacing, including an overview of collagen thermochemistry. The second section reviews wound healing in LSR.
PHYSICS: TISSUE HEATING AND ABLATION
Despite several excellent works,25-29 the dynamics of laser-tissue heating and ablation are complex and are not adequately characterized, either in theoretical models or through experimentation. This is particularly true in carbon dioxide LSR, where complicated physical relationships between tissue heating and ablation occur over the narrow range of fluences typically used in clinical practice. This is further complicated by the bell-shaped beam profile most often used in carbon dioxide LSR.
The carbon dioxide laser possesses intrinsic properties that enhance its usefulness in LSR, the most important of which is the high absorption coefficient for tissue water (approximately 800 cm-1 at room temperature).29 This permits minimal RTD if 1 of the following conditions is satisfied: (1) the power density (fluence divided by exposure time) is sufficient that vaporization significantly outpaces the speed of thermal diffusion; (2) the energy is delivered in less than the thermal relaxation time ( ); or (3) the fluence is small enough that only a thin layer of collagen is denatured.26, 29
Energy density (fluence or radiant exposure) is often cited when characterizing tissue effects in laser dermatology. Indeed, for pulsed laser applications where pulse duration is approximately , this parameter serves as an adequate determinant of the laser-tissue interaction.26, 28-31 However, in LSR, pulse duration (and its scanning counterpart, dwell time) may vary as much as 10-fold between different devices. It follows that power density is a more useful parameter in determining RTD and vaporization depth in LSR.32-34 This underscores the sensitivity of laser-tissue interactions to the rate of energy deposition as well as total radiant exposure. For example, even though the ablation "threshold" has been measured as 5 J/cm2 for 1-millisecond carbon dioxide laser pulses, it has been shown to increase to 15 J/cm2 for longer pulses.20, 29 On the other hand, it may be as small as 1 to 4 J/cm2 for the microsecond-domain laser. The instantaneous power densities of the newer-generation pulsed and scanned carbon dioxide lasers range from 10,000 to 50,000 W/cm2. This allows ablation to proceed efficiently once tissue water has been heated. As the power density drops to around 2000 W/cm2, there is a slowing of vaporization and a relative increase in heating so that tissue is denatured and finally charred.35 This is the case in most continuous-wave carbon dioxide applications in defocused mode (eg, wart treatment), where power densities range from 50 to 800 W/cm2. See Figure 1 for a graphical depiction of the role power density, pulse duration, and fluence play in ablation and heating.
|
|
|
|
Figure 1. A, This graph shows the effects of pulse duration and fluence for a low power density, 60 W/cm2. This scenario occurs, for example, when treating a wart with a setting of 5 W and a 3-mm spot (defocused). Note that by the end of 1 second, residual thermal damage (RTD) has increased to nearly 500 µm. Ablation is negligible at this power density. B, Note the depths of ablation and RTD for a high power density, 10,000 W/cm2 (like that used in laser skin resurfacing (LSR) over the same time scale. By the end of 1 second, ablation has increased to deeper than 6 mm, but RTD has remained relatively constant at 150 to 200 µm. In LSR, we restrict the exposure time to 0.5 to 5 milliseconds (part of graph denoted by arrow). With longer exposures, ablation depth increases rapidly. This might be desirable when cutting tissue, for example, and can be accomplished by focusing 5 W into a 0.2-mm beam. C, This graph depicts the influence of pulse duration and fluence (for power density x 10,000 W/cm2) over the narrow range of parameters used in LSR for photodamage. The graph is an enlarged version of the small region denoted by the arrow in B. Note the narrow ranges of both RTD and ablation.
|
|
|
Despite literature extolling the virtues of ablation in carbon dioxide LSR,2, 10-11,15-16,24, 36-37 dermal ablation is minimal (<20 µm per pass) in typical applications where relatively low average fluences are used (approximately 7 J/cm2). Moreover, it has never been established that dermal ablation is necessary for clinical efficacy. Accordingly, an understanding of carbon dioxide LSR can be based primarily on nonablative tissue heating. In short, for the carbon dioxide laser with typical fluences, the depth of RTD, rather than the depth of dermal ablation, largely determines the depth of injury. More precisely, the total depth of dermal injury is the sum of a thinly ablated layer (if ablation occurs at all) plus the thermally damaged base.
The most important feature of novel carbon dioxide lasers is that the depth of dermal RTD is regularly controlled to within 60 to 120 µm for a range of pass numbers and repetition rates (up to 1 Hz). Pulse duration plays a major role in restricting RTD. Optimally, the exposure time should not exceed for a heated layer of water (the primary chromophore for infrared radiation) at 10.6 µm. Thermal relaxation time (for a planar geometry) is defined by = 2/4 , where is the optical penetration depth (OPD) for homogeneously absorbing layers (such as tissue water for infrared applications). In contrast, for discrete absorbers, ie, the melanosome, is defined in terms of the particle size; "" is the thermal diffusivity, a quantity based on the thermal conductivity and specific heat of the medium.38-39 Conceptually, represents the time for a heated layer to cool to 37% of its peak temperature. The OPD of laser irradiation is defined as the depth where fluence is attenuated to 37% of its incident value. The often-used term "thermal relaxation time of the skin" is meaningful only when used for specific wavelengths (or specific skin structures, eg, the epidermis). When used with reference to the skin in general, the frequently reported of 1 millisecond is meaningful only for 10.6-µm radiation.
If energy is delivered in less than , commonly applied models predict that heating will be confined to the OPD during the laser pulse. However, the depth of RTD is usually 3 to 4 times the OPD. This is because there is some heat diffusion during the pulse even with very short exposures, and deeper tissue heating occurs after the pulse. In practice, is only an approximation and does not predict the absolute time it takes tissue to cool.39-41
The failure of the above equation to accurately predict tissue cooling has been confirmed by Welch et al,42 who showed that even 0.25 second after irradiation with a microsecond-domain carbon dioxide laser, the surface temperature exceeded baseline by 20°C. Moreover, they found experimentally that was 20 to 40 milliseconds for the carbon dioxide laser, not the 1 millisecond usually cited in the literature.
Choi et al43 examined surface temperature as a function of pass number and fluence with the same 60-microsecond pulsed carbon dioxide laser. In their study, they applied 3 passes, wiping after the first but not the second pass. Even though minutes elapsed between passes, the peak surface temperature on first, second, and third passes increased from 335°C to 341°C to 388°C, respectively, for fluences of approximately 4 J/cm2. The similar peak temperatures for the first and second passes were explained by rehydration from wiping; presumably the constant water content normalized the surface laser-tissue interaction. On the third pass, the markedly higher RTD and peak temperatures were explained by the presence of light char, which acted as a heat sink (very high absorption), and superficial desiccation (ie, decreased thermal conductivity, so that heat tended to accumulate at the surface).
There are notable differences between carbon dioxide and erbium:YAG ablation for the typical 5- to 8-J/cm2 fluence range.44-47 The more explosive vaporization and louder sounds observed with erbium:YAG ablation are due in part to the larger absorption coefficient of the erbium:YAG laser (approximately 13,000 cm-1 for water), resulting in a greater absorbed volumetric power density, defined as the surface power density times the OPD.27 Another contributor to the loud sounds is the pulse profile of the erbium:YAG laser, since the typical 250-microsecond erbium:YAG pulse is actually composed of a series of 1- to 2-microsecond micropulses, each ablating the skin in a "jackhammer" fashion. Laser parameters can be modified so that the carbon dioxide laser behaves more like an erbium:YAG laser (in essence, by increasing the absorbed volumetric power density levels to that of erbium:YAG). For example, if the millisecond carbon dioxide laser is applied with a 1-mm spot size and high pulse energy (350 mJ, fluence=40 J/cm2), a loud pop occurs similar to that produced by an erbium:YAG laser at 4 J/cm2. Alternately, the pulse duration of the carbon dioxide laser can be shortened to achieve the same "erbium:YAG-like" effect.
The thresholds for ablation for carbon dioxide and erbium:YAG lasers vary inversely with their optical penetration depths in tissue (20 and 1 µm, respectively). This assumes thermal confinement. It follows that less surface fluence is required for ablation with the erbium:YAG laser. With the carbon dioxide laser, we are operating at ablation threshold in typical resurfacing applications, so a large fraction of energy is invested in tissue heating. This results in low ablation efficiency, and only a small mass of dermal tissue is ablated. In contrast, the erbium:YAG laser operates well above threshold (approximately 8-10 times threshold for a fluence of 5 J/cm2), resulting in greater ablation and less thermal denaturation.48-49 In brief, the carbon dioxide laser at typical operating parameters performs a self-limited, controlled heating of the surface, whereas the erbium:YAG laser operates in an almost purely ablative manner. However, the ablation efficiencies (micrograms of tissue removed per Joule per square centimeter) for erbium:YAG and carbon dioxide lasers are nearly equal once both lasers are used at several times their respective ablation thresholds.50
The thermal relaxation time for a layer of heated water 1 µm thick (equal to the OPD of the erbium:YAG laser) is only 1 microsecond.50 The standard 250-microsecond erbium:YAG pulse considerably exceeds this defined by the OPD (see above equation). Although somewhat counterintuitive, it follows that relative thermal confinement for a 1-microsecond pulse from a carbon dioxide laser is superior to a 250-microsecond exposure from an erbium:YAG laser. This means that although a 1-microsecond carbon dioxide laser pulse produces more RTD than erbium:YAG, it is closer to its wavelength-specific theoretical minimum (20 µm, vs 1 µm for the erbium:YAG laser, defined by their respective OPDs). Thus, substantial heat "leakage" with the erbium:YAG laser (during the pulse) is responsible for greater RTD than predicted simply by the OPD. The use of a 250-microsecond pulse evolved largely from engineering considerations, as the erbium:YAG laser operates efficiently over this time. Also, fortuitously, it has contributed to larger zones of RTD (20-50 µm) that undoubtedly have improved hemostasis. The small OPD of the erbium:YAG laser allows it to be used in different regimes (ablative or coagulative) depending on fluence, pulse duration, and repetition rate.50 For example, with the erbium:YAG laser, if the pulse duration or repetition rate is increased (or fluence decreased below threshold), RTD can be increased incrementally with improved hemostasis.51-52 Companies are presently exploring "hot" ablation and subablative strategies with erbium:YAG lasers.
In contrast, with the carbon dioxide laser, RTD "starts off" at approximately 80 µm RTD (for 1-millisecond–domain laser pulses; less with a 60-microsecond pulse; approximately 30-50 µm after multiple passes). It follows that increases in pulse duration for carbon dioxide beyond CO2 will tend to increase RTD to dangerously high levels.
EXAMINATION OF THE EFFECTS OF A GAUSSIAN BEAM AFTER SINGLE AND MULTIPLE PULSES AS AN AID TO UNDERSTANDING CARBON DIOXIDE LSR
It is instructive to examine the tissue effects of a Gaussian beam (Figure 2) after single and multiple pulses in a single spot. In this manner the effects of a broad range of fluences can be analyzed in a single specimen in carbon dioxide LSR. On the other hand, the use of this profile has contributed to confusion in LSR regarding tissue ablation. A Gaussian beam (seen in the millisecond and continuous-wave systems, but not the microsecond system) distributes energy density like a bell-shaped curve. The spot diameter is defined in terms of the distance from the center of the beam at which the fluence falls to 14% of the maximum.29, 53 The average fluence is defined by the energy divided by the spot diameter. The fluence at the center is twice this average. Because of the energy distribution pattern and the unlikelihood of directly overlapping beam profiles in second and third passes when operating in a typical manner (complete 1 pass, then wipe and return to the site), a particular site on the skin surface is likely to "see" a subablative fluence in 1 pass, followed by a supra-ablative fluence on the next pass. The impact of the Gaussian beam becomes especially apparent when working with higher fluences (ablation regime >10 J/cm2 or when applying >20 passes with 7 J/cm2 and holding the handpiece in a constant position, eg, in attempting to remove a seborrheic keratosis). In this case, an ablation defect mirrors the beam profile, and the crater looks like a raindrop in the sand rather than a cylindrical defect (Figure 3, B).
|
|
|
|
Figure 2. The relative fluence of a representative Gaussian beam.
|
|
|
|
|
|
|
Figure 3. A, Pigskin dermis after 1 pass with millisecond-domain carbon dioxide laser, average fluence 10 J/cm2, 1-mm spot, Gaussian beam profile (arrow indicates where local surface fluence is approximately 3 J/cm2; original magnification x100). B, After 10 passes, average fluence 10 J/cm2 (original magnification x40) (bar=400 µm, hematoxylin-eosin stain, both A and B).
|
|
|
Both nonablative and ablative heating effects can be examined in 1 specimen. For example, in Figure 3, A, where the average fluence is 10 J/cm2 (the center fluence will be twice that, or 20 J/cm2), a continuum of changes is evident from the center to the edge. In the center there is slight dermal ablation. The remainder of the specimen shows little ablation but relatively constant RTD up to the very edge of the beam. This is consistent with recently reported data by Weisberg et al,54 where there was a rapid rise in RTD at a threshold. Also, a graph showing RTD vs fluence (Figure 4) is similar to the profile of basophilic-staining collagen in Figure 3, A; in fact, if the two are overlaid with either inverted, their profiles match. This supports the principle that so long as energy is thermally confined during the pulse, absorption coefficient, and not fluence or presence of ablation, is the primary determinant of RTD. The aforementioned relation between RTD and fluence also explains the uniformity of injury in carbon dioxide LSR despite the use of various fluences. By performing the standard 2 or 3 passes with a scanner, the result at 2 weeks typically reveals a very uniform erythema that is difficult to obtain through purely ablative modalities (dermabrasion or erbium:YAG laser).
|
|
|
|
Figure 4. Graph of residual thermal damage vs fluence for a millisecond-domain carbon dioxide laser. Arrow designates ablation threshold. Dotted line indicates predicted residual thermal damage with ablation. Note resemblance of the graph to the inverted profile of denatured collagen at the edge of Figure 3, A (arrow). This graph is based on a simple model for tissue heating using the Beer law.
|
|
|
In the 10-pass specimen (Figure 3, B), where the handpiece was maintained in the same exact location over tissue, the area can be ablated (or only heated, ie, at the wings of the crater) without an increase in RTD, so long as the pulses are delivered independently in time and the epidermis has been removed. There is considerable ablation at the center, but very little at the edges, where the fluence is much lower. The ablation crater mirrors the Gaussian beam profile (Figure 2). This is consistent with theory, since ablation, once threshold is established, can be shown to proceed as a linear function of fluence. On the other hand, RTD follows the Beer law, which relates an exponential attenuation of subsurface fluence as a function of depth in tissue.50 It follows that despite the wide range of fluences across the specimen (even after 10 passes), RTD varies slightly. This will hold true so long as the pulse duration is held to approximately 1 to 5 milliseconds. With longer pulse durations, an increase in RTD will occur at the wings of the crater.20
COLLAGEN DENATURATION
Collagen denaturation is a complex process not completely characterized despite decades of experimentation.50, 55-63 Studies have been performed under various conditions with different animal species. For example, most shrinkage experiments are performed under either isotonic, isometric, or isothermal conditions. These laboratory conditions may not accurately represent skin that is heated rapidly in high-energy pulsed infrared applications.
Increasing tissue tension secondary to collagen heating has been attributed to the phenomenon of denaturation, but it is actually owing to the rubber-elastic properties of collagen that denaturation confers on the collagen polymeric network.62 It is known that heated collagen transforms from the crystalline form (right-handed superhelix) to a more amorphous gel with concomitant loss of birefringence and a peculiar staining behavior with dyes (eg, hematoxylin-eosin).64 Denaturation results in thickening and shortening of fibrils. The transition from the crystalline to the amorphous state involves multiple steps and varies between mammalian species as well as between ages within a species.65 A detailed analysis of collagen rate kinetics can be found in the references.62, 66-67 The first step in collagen denaturation is the rupture of hydrogen bonds between strands of the triple helix so that the helix transforms into 3 random-coil molecules. The loss of these intrachain hydrogen bonds (which temporally coincides with the bulk of collagen denaturation by sensitive calorimetry methods) results in a rapid rise in tension if the fiber is held at constant length. In some animal studies, further heating induces partial relaxation through rupture of intermolecular heat-labile covalent cross-links (the tension actually begins to decrease as temperature rises beyond 70°C over several minutes).56, 62-63 Intermolecular cross-links in human skin, however, are heat stable so that further heating results in continued increases in tension (although at a slower rate than with hydrogen-bond rupture). Remarkably, even after denaturation is complete, in human skin there is additional fiber shortening with increasing temperature. This is most likely due to hydrolysis of peptide bonds.62
Denaturation begins well before visible collagen shrinkage and probably before globally apparent microscopic architectural changes on routine microscopy (including birefringence loss). Even with early electron microscopy changes (granular disintegration of cross-links), contraction is not visible.68 It follows that a critical mass of fibrils must be denatured before bulk shrinkage is observed. This has been confirmed by Allain et al,56 who observed that it is the number of participating molecules within the bulk-denatured extracellular matrix that results in the amplitude of tension and shrinkage. This is supported by recent observations by Weisberg et al54 that (1) additional shrinkage may be observed in skin despite no identifiable increase in RTD and (2) the fluence threshold for shrinkage exceeded that of RTD in their in vivo experiment.
One issue confounding the clinician is the often-quoted temperature of 65°C as the shrinkage point of collagen.69 This number is only accurate for relatively long exposures, on the order of several seconds, somewhat longer than for pulsed lasers. In fact, there is no true shrinkage temperature for collagen, but only a range of temperature-time combinations. Protein denaturation is a rate process characterized mathematically by the Arrhenius equation. For collagens in general, for every 5°C decrease in temperature, a 10-fold increase in time is needed to achieve the same degree of denaturation.70 However, it is unclear if this time-temperature reciprocity holds for very short time exposures, for which denaturation kinetics are unknown.71 Zweig et al72 have suggested that the shrinkage temperature for millisecond-domain exposures must exceed 85°C.
There are distinct staining zones observed after multiple-pass carbon dioxide LSR. These are related to temperature gradient as a function of depth. Typically, there is a very fine (1- to 2-µm) zone of intense basophilia, and sometimes char, overlying a much larger basophilic zone about 60 to 100 µm thick (zone I). Deep to this, there is a zone composed of a mixture of basophilia and hypereosinophilia about 10 to 20 µm thick (zone II). Finally, there is a third zone that stains hypereosinophilic and extends an additional 30 to 250 µm, depending on the laser parameters (zone III). Different staining properties are due to variable states of denaturation, where staining reaction is dependent on exposed sites for dye binding. These sites and the exact characterization of this binding have not been identified or formally compared with depth of denaturation by polarization microscopy (birefringence).
From our observations, the basophilic-staining zone (zone I) corresponds to the depth where little to no birefringence is observed. This is typically referred to as the coagulative zone (and is the entity commonly referred to as RTD in the literature). In the mixture zone (zone II), there is a slight decrease in overall birefringence, as this zone demonstrates focal areas of intact fibers admixed with areas with complete loss of birefringence. In the transition zone (zone III), birefringence is similar to normal-appearing untreated collagen.
Ultrastructurally, Kirsch et al73 differentiated 3 zones based on fibril alteration by electron microscopy. Their study concluded that a zone comprising a mixture of denatured and native fibrils occurred sandwiched between normal and completely denatured zones. The width of the zone featuring the mixed fibrils was only 10 to 20 µm; most likely this zone coincides with our zone II.
Zweig et al72 identified a transition zone by routine staining that coincides with our zone of hypereosinophila (zone III). They performed electron microscopy on tissue in this zone (which showed retained birefringence) and noted that the fibrils were only slightly thickened and retained their periodicity. This suggests that subtle tinctorial change on hematoxylin-eosin staining may be more sensitive than even electron microscopy in detecting early denaturation.
WHAT IS RESPONSIBLE FOR INITIAL WOUND SHRINKAGE?
Despite sometimes heated discussions regarding collagen shrinkage and its putative role in carbon dioxide LSR, no study has unequivocally shown that collagen denaturation is responsible for the immediate visible shrinkage observed in LSR. Moreover, if denaturation is responsible for immediate shrinkage, a critical question is from which layer (of the 3 described zones of microscopic staining) does the contraction derive? We suggest that initial shrinkage does result primarily from collagen denaturation, and more specifically that it derives from the coagulative zone (zone I). We support our position as follows (in addition to the arguments generated from the description of collagen denaturation in human skin in the previous section): (1) Experiments have shown that in wounds where RTD is confined to the epidermis, little (<3% by area) or no visible contraction is observed.21 (2) We have observed in our own nonablative resurfacing experiments (unpublished observations, 1997) when we combined a topical cooling device and deep-penetrating wavelength, there was gross shrinkage at the surface only with basophilic changes in the dermis. Slight hypereosinophilic staining was associated only with erythema and edema. (3) Shrinkage follows denaturation in vivo. Therefore it is unlikely that the deeper transition zone, where denaturation is minimal, is responsible for bulk wound shrinkage. (4) Wound contraction profiles show that wound areas expand just as the basophilic zone is sloughed (about 1-3 days after surgery).21, 74
We suggest that initial shrinkage serves primarily as a monitor for underlying denaturation (dermal heating), and we speculate that there is only a small direct role for the basophilic zone shrinkage in final wound contraction. Even in cases where there is apparent integration of basophilic-staining collagen within granulation tissue (Figure 5), it is not known that this denatured collagen persists long enough to serve as a shrunken template for newly deposited collagen.
|
|
|
|
Figure 5. Pigskin 7 days after carbon dioxide laser skin resurfacing, 3 passes, with wiping, 7 J/cm2, millisecond laser, with occlusive dressings (OpSite; Smith and Nephew, Largo, Fla). Note areas of retained basophilic-staining collagen intermixed and subjacent to neoepidermis (arrow). This may serve as a shrunken template for new collagen deposition (hematoxylin-eosin, original magnification x200).
|
|
|
After careful consideration of the literature and our own observations, we suggest that dehydration intrinsically plays only a small role in immediate wound contraction. This is supported by (1) the carbon dioxide irradiation of agar, which causes no shrinkage (personal observation, E. Victor Ross, MD, 1998); (2) the fact that thermally altered collagen is hydrophilic and undergoes shrinkage in a hot water bath with histological changes identical to those produced from LSR75; and (3) the fact that with rehydration, there is only partial restoration of immediate tissue shrinkage.76
On the other hand, dehydration might be responsible for some shrinkage in LSR, as there is water loss through vaporization in the upper 10 to 20 µm of tissue (histologically noted by "popcorning"). Moreover, a recent ultrastructural study77 suggests that the loss of glycosaminoglycans (and water) between adjacent collagen fibers is the primary cause of immediate shrinkage in carbon dioxide LSR. However, the authors failed to show the absence of fibril shortening. This study reveals one of the difficulties in assigning "responsibility" for contraction, namely, that water is intimately associated with collagen in tissue, even stabilizing some hydrogen bonds. During heating, water is set free from collagen, after which it is slowly re-bound to the amorphous protein with cooling.78 Thus water, collagen, and the denaturation process are inextricably linked.
As might be surmised from the above discussion, assigning "explanations" for various observations in carbon dioxide LSR is made difficult by the complex and simultaneous nature of heat-mediated events. For example, tissue heating simultaneously results in water vaporization, freeing of bound water from collagen (as part of denaturation), and the ejection of intact particles due to the heterogeneous nature of tissue. This has confounded efforts to rigorously describe tissue ablation and heating.26-27,46-47 Moreover, in the case of the erbium:YAG laser, there is intrinsic absorption for some collagen bonds, so that water is not the sole chromophore.79
BIOLOGICAL CHARACTERISTICS OF WOUND HEALING
The distinct zones of RTD in carbon dioxide LSR distinguish it from other resurfacing modalities. What follows is a narrative of wound healing with carbon dioxide LSR based on our own pigskin and human experiments. During carbon dioxide LSR (for  2 passes), the deepest portions of the basophilic-staining zone as well as the zone of mild hypereosinophilic collagen staining (transition zone, Figure 6) are the critical zones for modulation of wound healing. Within these areas immediately after treatment (within 30 minutes), vascular stasis and slight nuclear hyperchromasia of fibroblasts occur. Tissue viability stains show these fibroblasts to be metabolically active, whereas, as expected, fibroblasts in the upper and mid–basophilic-staining zone are necrotic.19, 80
|
|
|
|
Figure 6. Pigskin 2 days after carbon dioxide laser skin resurfacing, 3 passes, with wiping, 7 J/cm2, millisecond laser, with only petrolatum applied to the wound. Most of the basophilic zone has already sloughed. Note necrotic fibroblasts in the transition zone (which extends to the level of the short arrows, where there is a subtle change in collagen staining). At the base of the photograph are a few plump, viable fibroblast nuclei (long arrow) (hematoxylin-eosin, original magnification x200).
|
|
|
One day after treatment, the basophilic zone is still intact microscopically. Within and just subjacent to this zone, neutrophils extend to the deepest aspect of the zone of mild hypereosinophilia (transition zone). Within this transition zone, tissue vitality stains show fibroblast death but retained birefringence (thus confirming the greater sensitivity of cells to higher temperatures than collagen—typically only 10% of cellular protein must be denatured for death, and these are often not as thermally stable as the crystalline collagen matrix).
By 2 days after treatment, there is no further extension of fibroblast death. The basophilic-staining layer begins to slough in some regions, whereas in others, where there is less inflammation focally, a portion of this completely denatured collagen is retained (zone I). Overlying this remnant of the thicker basophilic zone seen on day 1, a compact layer of fibrinous debris is occasionally observed (consistent with the crust, grossly). In other cases, it appears that both this debris and the entire basophilic-staining zone have sloughed (Figure 6). Deep to this layer, the hypereosinophilic zone continues to show fibroblast death but retention of birefringence. Remarkably, necrotic fibroblast nuclei and even the collagen fibers show a more horizontal orientation than adjacent normal fibers of equivalent depth. This reorientation might be due to stresses by overlying attached shrunken collagen, and facilitated by slight changes in the deeper collagen fibers that increase their malleability. Interestingly, studies with and without occlusive dressings show different wound-healing responses. With dressings, the basophilic zone tends to remain intact and later become integrated within newly forming granulation tissue, and reepithelialization tends to occur superficial to this zone (Figure 5). These specimens suggest that the basophilic-staining zone might play a direct role in resetting the lattice size for new collagen deposition and thus preserving a portion of the initial wound shrinkage. In contrast, in wounds treated with less occlusive methods (ie, application of only petrolatum), there is more pronounced inflammation and sloughing of most of the basophilic zone after 2 days (Figure 6).
During and after carbon dioxide irradiation, there is immediate contraction roughly proportional to the amount of dermal RTD over a range of 20 to 120 µm (for similar laser devices, ie, pulsed vs scanning). The associated area percentage of contraction ranges from 5% to 18%. Beyond 150-µm RTD, immediate contraction does not exceed 30%. After a portion of the denatured collagen sloughs (days 1-2), some initial shrinkage is lost, and the area is restored to an intermediate value between its original and minimum size. This may be due to rehydration (the basophilic zone prevents further wound desiccation after injury) and release of the tethering effects of the basophilic zone. A second contraction then occurs after carbon dioxide LSR (days 5-10) that is presumably due to myofibroblasts.
BY 5 DAYS after treatment in multiple-pass carbon dioxide wounds, partial reepithelialization occurs and early granulation tissue can be observed. Focal remnants of basophilic-staining collagen persist underlying this neoepidermis. Within areas of persistent denatured collagen, viable fibroblasts have presumably replaced necrotic ones observed 1 and 2 days after the laser procedure. By 7 days after treatment, there is a robust zone of granulation tissue approximately 200 to 350 µm thick underlying the epidermis. Areas of denatured collagen are occasionally still identifiable by polarization or routine microscopy deep to the neoepidermis. Seventeen days after treatment, the granulation tissue has become more organized but retains a thickness of approximately 300 µm. There is greater fibroblast density, nuclear size, and vascularity vs wounds of similar depth achieved with dermabrasion or erbium:YAG laser. Between 17 and 60 days after treatment with multiple-pass carbon dioxide laser, there is a slow relaxation in wound contraction (increase in area) resulting in most wounds reaching 90% to 95% of their original sizes. Histologically, the thickened hypercellular dermis becomes more compact (usually 150-200 µm thick) and collagen rich. Also, there is increasing horizontal alignment of collagen and elastic fibers.
This sequence of events is different after dermabrasion and erbium:YAG resurfacing (carried to the same depth of injury, where depth of injury equals the depth of tissue removal plus the depth of zone I RTD). There is only slight immediate wound contraction (3%-7% of original area), less than 50 µm RTD, and imperfect hemostasis. By 1 day after surgery, there is an irregular band of polymorphonuclear leukocytes extending from the wound surface to 200 µm deep in the dermis. Also, fibrin can be seen as a layer up to 100 µm thick at the wound surface. In the absence of occlusive dressings, desiccation is pronounced compared with multiple-pass carbon dioxide wounds,81 and there is more rapid re-epithelialization (3-5 days). Wound contraction begins only 2 to 3 days after surgery; there is no delay or biphasic contraction pattern as noted for carbon dioxide wounds. Microscopy 60 days after injury shows less tightly packed collagen fibers with a less horizontal orientation than in multiple-pass carbon dioxide wounds. Also, the elastic fibers are finer and more vertically oriented than in untreated skin. By 60 days after treatment, the initial wound contraction almost completely regresses in all cases of conservative erbium:YAG and dermabrasion wounds (wounds with <200 µm of dermal ablation).
What differentiates carbon dioxide laser from other resurfacing strategies of similar depth? It has been suggested that the method of injury is irrelevant but, rather, the total depth determines the subsequent healing cascade. Certainly this is intuitively attractive, since such dissimilar injuries as dermabrasion, chemical peeling, erbium:YAG, and carbon dioxide laser all can achieve adequate improved cosmesis. However, we and others have performed studies, as outlined below, that suggest that RTD independently modulates wound healing.
The wound contraction profiles of carbon dioxide wounds and purely ablative injuries (ie, conventional erbium:YAG and dermabrasion) suggest an initial healing delay in carbon dioxide–treated sites; explanations include the time interval for basophilic collagen sloughing, small vessel occlusion, and reduction in early myofibroblast activity.82 After this initial delay, most studies show greater and more sustained increases in wound-healing factors in carbon dioxide wounds vs their scalpel counterparts. For example, one study showed more and earlier epidermal growth factor expression after carbon dioxide laser incisions vs scalpel wounds.83 Pogrel et al84 found that carbon dioxide incisions showed earlier and greater hyaluronidase activity than scalpel incisions in rats. Hyaluronidase activity was also increased over a longer period of time after surgery. They suggest that the prolonged activity might be related to the later neovascularization in laser wounds. Recently, Harmon et al85 showed higher levels of platelet cell adhesion molecule after 6 weeks in carbon dioxide vs erbium:YAG laser wounds where the depth of injuries (as assessed by depth of ablation plus depth of RTD) were similar. Smith et al86 have shown increased factor 13, vimentin, and actin expression after pulsed carbon dioxide laser injury in a pig but did not compare these findings with non–carbon dioxide wounds.
It is less clear how thermal injury modulates the pattern and quantity of new collagen deposition. Li et al87 have shown that there is up to 3 times the collagen content in burn wounds as freezing wounds in rat skin. Margolis et al88 showed a greater and more sustained elevation in collagen content after carbon dioxide injury vs 50% trichloroacetic acid and dermabrasion wounds in mouse skin. Luomanen et al89 showed more collagen formation at 4 weeks in carbon dioxide laser wounds than in comparable scalpel wounds in rat oral mucosa. Recently it has been shown that collagen fragments can stimulate new collagen deposition and the directed migration of fibroblasts and the associated alignment of fibrils.90 This P15 residue, which presumably is released during the thermal denaturation of collagen, may be responsible for induction of mechanical stresses that in turn increase the proliferation of fibroblasts over injuries where collagen degradation proceeds solely through enzymatic pathways. Furthermore, the tractional forces generated by the initial collagen shrinkage may exist long enough to reorient the newly synthesized extracellular matrix. Despite the above argument for a role for denatured collagen in directing fibroblast migration, it should be noted that a recent study91 showed no difference ultrastructurally between collagen fibers 180 days after treatment in dermabrasion and carbon dioxide laser sites in a pig.
In short, LSR wounds can be categorized into 3 types: (1) wounds combining little dermal ablation (<50 µm) and dermal RTD greater than 60 µm (typical carbon dioxide LSR); (2) wounds combining little dermal ablation and less than 60 µm RTD (conservative erbium:YAG LSR); and (3) wounds with moderate dermal ablation (50-150 µm) and less than 50 µm RTD (typical erbium:YAG LSR). In a human side-by-side erbium:YAG–carbon dioxide study (with all 3 wound types),92 we showed that re-epithelialization and resolution of erythema were slightly faster with the erbium:YAG laser, even where apparent depths of injury were equivalent. The final fibroplasia zones at 6 months were thinner for erbium:YAG vs carbon dioxide laser. This is consistent with the pig study discussed above and suggests that RTD in carbon dioxide laser wounds intrinsically affects healing, and that more long-term wound contraction and fibroplasia per micrometer depth of injury is achieved with the carbon dioxide laser. The final zone of fibroplasia for a specific carbon dioxide device roughly varies as the layer of RTD in carbon dioxide laser wounds, whereas for erbium:YAG laser wounds, the thickness of the zone reflects the depth of ablation. Generally, to create equivalent zones of fibroplasia and erythema, we have found erbium:YAG ablation depths must be at least 1.5 to 2 times that of RTD depth with carbon dioxide.
COMMENT
|
