Percentage of shrinkage of in vitro human skin per pass of the laser plotted vs the fluence of the laser. n=6 for each data point. A, Data for the short-pulsed laser (energy in millijoules noted at the top of the graph). B, Data for the scanner laser (intensity in watts noted at the top of the graph). Lines are fits using the equation in the "Calculations" section of the text. The fit parameters—1/2 (I½) and T (IT) (as explained in the text)—are indicated by arrows.
Percentage of shrinkage of in vivo porcine skin per pass of the laser plotted vs the fluence of the laser. n=6 for each data point. A, Data for the short-pulsed laser (energy in millijoules noted at the top of the graph). B, Data for the scanner laser (intensity in watts noted at the top of the graph). n=6 for each data point. Lines are fits using the equation in the "Calculations" section of the text. The fit parameters—1/2 (I½) and T (IT) (as explained in the text)—are indicated by arrows.
Histological sections of in vitro human tissue treated with 3 passes of the respective lasers at various fluences (Gomori trichrome stain, all sections photographed at ×50) (bar=100 µm). A, Pulsed laser, subthreshold (100 mJ); B, pulsed laser at threshold (330 mJ); C, pulsed laser above threshold (500 mJ); D, scanner laser, subthreshold (1.0 W); E, scanner laser at threshold (3.5 W); and F, scanner laser above threshold (7.0 W). Arrows indicate the lower level of collagen denaturation as denoted by tincture change.
Histological sections of in vivo porcine tissue treated with 3 passes of the respective lasers at various fluences (Gomori trichrome stain, all sections photographed at ×50) (bar=100 µm). A, Pulsed laser subthreshold (100 mJ); B, pulsed laser above threshold (500 mJ); C, scanner laser subthreshold (1 W); D, scanner laser above threshold (8 W). Each point is the average of 10 measurements per slide, with 2 pigs used for the experiment. Arrows indicate the lower level of collagen denaturation as denoted by tincture change.
Depth of histological thermal damage of in vitro human skin as determined by tissue morphometry on Gomori trichrome–stained slides plotted vs fluence. A, Pulsed laser (energy in millijoules at top of graph); B, scanner laser (intensity in watts at top of graph). Each point is the average of 10 measurements per slide, with 2 different samples used for each experiment. The fit parameters—1/2 (I½) and T (IT) (as explained in the "Calculations" section of the text)—are indicated by arrows.
Weisberg NK, Kuo T, Torkian B, Reinisch L, Ellis DL. Optimizing Fluence and Debridement Effects on Cutaneous Resurfacing Carbon Dioxide Laser Surgery. Arch Dermatol. 1998;134(10):1223-1228. doi:10.1001/archderm.134.10.1223
To develop methods to compare carbon dioxide (CO2) resurfacing lasers, fluence, and debridement effects on tissue shrinkage and histological thermal denaturation.
In vitro human or in vivo porcine skin samples received up to 5 passes with scanner or short-pulsed CO2 resurfacing lasers. Fluences ranging from 2.19 to 17.58 J/cm2 (scanner) and 1.11 to 5.56 J/cm2 (short pulsed) were used to determine each laser's threshold energy for clinical effect. Variable amounts of débridement were also studied.
Main Outcome Measures
Tissue shrinkage was evaluated by using digital photography to measure linear distance change of the treated tissue. Tissue histological studies were evaluated using quantitative computer image analysis.
Fluence-independent in vitro tissue shrinkage was seen with the scanned and short-pulsed lasers above threshold fluence levels of 5.9 and 2.5 J/cm2, respectively. Histologically, fluence-independent thermal depths of damage of 77 µm (scanner) and 25 µm (pulsed) were observed. Aggressive debridement of the tissue increased the shrinkage per pass of the laser, and decreased the fluence required for the threshold effect. In vivo experiments confirmed the in vitro results, although the in vivo threshold fluence level was slightly higher and the shrinkage obtained was slightly lower per pass.
Our methods allow comparison of different resurfacing lasers' acute effects. We found equivalent laser tissue effects using lower fluences than those currently accepted clinically. This suggests that the morbidity associated with CO2 laser resurfacing may be minimized by lowering levels of tissue input energy and controlling for tissue debridement.
CUTANEOUS carbon dioxide (CO2) laser resurfacing has gained increasing acceptance as an effective therapeutic modality for rhytidosis and photoaged skin. Unfortunately, troublesome adverse effects such as persistent erythema, pigmentary alteration (hyperpigmentation and hypopigmentation), and/or hypertrophic scarring have been reported with the use of these new lasers.1,2 Optimizing the laser technique may lead to improved clinical results with fewer adverse effects.
The working model for the CO2 laser's mechanism of action revolves around the high degree of 10.6-µm radiation absorption by water.3 Water absorption of the laser energy causes rapid, localized heat production with subsequent water vaporization. This exothermic reaction causes tissue ablation. As the dermis is penetrated and the relative concentration of water decreases, the laser energy is increasingly dissipated to the surrounding tissue as heat. This thermal transfer is probably responsible for dermal collagen denaturation. Denaturation of collagen has been theorized to contribute to tissue shrinkage and the ultimate clinical goal of wrinkle reduction.4 Thermal energy from the laser may also be the primary source of the noted cutaneous resurfacing complications.
While multiple articles have been published about the efficacy of CO2 laser resurfacing,5- 12 studies delineating optimal CO2 laser resurfacing parameters are lacking. We recently compared 2 types of CO2 resurfacing lasers using a human skin in vitro system.4 This study used that model to determine threshold fluences for 2 types of CO2 resurfacing lasers and to compare the effects of tissue débridement on threshold fluences. Because in vitro studies may vary significantly from in vivo studies, we compared our in vitro results with in vivo results obtained with porcine skin to confirm the applicability of our results to live subjects.
Human skin obtained as excess tissue from reduction mammoplasties or abdominoplasties was used for all in vitro experiments. Tissue was placed on saline-moistened gauze, double wrapped in aluminum foil, sealed in an airtight bag, and stored at 4°C for 2 days or less, or frozen at −20°C. All tissue was used within 2 months of freezing. All experiments were performed at room temperature.
A plastic template was used to cut 2.0×0.5-cm sections of tissue by scalpel incision. More than 80 tissue samples from at least 5 sources were used in this investigation. Tissue samples from at least 2 sources were used for each data point. Three samples from each source were used for each data point. Therefore, each data point consisted of 6 samples, treated and measured after each laser pass for 5 passes, making each value a least-squares fit of 30 measurements.
A Sharplan SilkTouch CO2 laser (Sharplan, Allendale, NJ) was the continuous-wave scanner laser used in these experiments. The 125-mm handpiece was connected to a Sharplan 1060 CO2 laser and set between 1 and 8 W with a 200-millisecond repeated scan. The "scan size" reported by the manufacturer for this handpiece is 3.7 to 4.0 mm. The scan size (as measured by burning on a tongue depressor and measuring the image on ×25 magnification with a surgical microscope) was a 3.4-mm spiral-scanning pattern.4 Laser fluences were calculated from the intensity as measured with a Power One energy meter with an LM-10 detector head (Coherent, Palo Alto, Calif). The TruPulse laser (Tissue Technologies, Albuquerque, NM) was the short-pulsed CO2 laser used for comparison to the scanner laser in these experiments. The laser settings for the TruPulse laser were between 100 and 500 mJ, with a fixed pulse duration of 100 microseconds. This laser produced a square spot size of 3×3 mm, confirmed using ×25 microscopy on a tongue depressor–burned image as above. Energy fluences for the TruPulse laser were calculated from the intensity as measured with a power meter (Molectron PM600, Toronto, Ontario). We measured energy fluences ranging from 2.19 to 17.58 J/cm2 for the SilkTouch laser and from 1.11 to 5.56 J/cm2 for the TruPulse laser.
For most debridement studies, debridement was defined as using a saline-soaked cotton swab to debride the tissue until no further desiccated tissue was removable. For some samples, a saline-soaked 4×4 gauze was used for debridement, as this method is often used clinically in CO2 laser resurfacing. The 4×4 gauze technique was not used for the majority of the experiments to avoid potential damage to the small in vitro samples of skin. In control tissue where no debridement was done, sterile saline solution was gently applied to the treated area with a cotton swab to rehydrate the tissue. The tissue was then blotted with a dry cotton swab.
The epidermal surface of the 2.0×0.5-cm strip of tissue was completely irradiated with nonoverlapping laser scans. Each tissue specimen received 5 laser passes. For all studies, each pass was followed by debridement or rehydration with a saline-moistened cotton swab (as above) and blotted dry prior to the next pass.
Piglets weighing 4 kg were used for the experiments. India ink tattoo dots placed 1.0 cm apart served as markers for contraction of the skin. The axes of the dots were oriented longitudinally on the animals. The 1.0×0.5-cm area inside the marks was irradiated with the lasers. The laser energy fluences, debridement techniques, and tissue and data analysis used were the same as for the in vitro experiments.
Tissue shrinkage was evaluated by measuring the linear distance change between 2 centrally placed India ink tattoos 1.0 cm apart. To measure shrinkage, the tissue was digitally photographed before irradiation and subsequent to each laser pass and debridement. The digital photographs were taken with a video camera (model MKC-301A, Ikegami, Tokyo, Japan) mounted to the sideport of a surgical microscope with a 400-mm focal length lens (Carl Zeiss Inc, Thornwood, NY). The camera was interfaced with the built-in video port of a microcomputer (MAC 840 AV, Apple Computer Inc, Cupertino, Calif). Distances were measured with Photoshop 2.0 software (Adobe, Mountain View, Calif).
The relative distance measured between the tattooed dots on the tissue was plotted vs the number of laser passes. These data were usually linear and were fit to a straight line. The slope of the line (S) gave the tissue shrinkage per laser pass for each laser fluence. Measured tissue shrinkage or depth of thermal damage (see "Histological Change" section) was graphed as a function of irradiation using the Henderson-Hesselbalch 2-state equation to obtain a best-fit line for the observed data points.
where S indicates tissue shrinkage or depth of thermal damage as a function of the laser irradiation (I); Smax, maximum tissue shrinkage or maximum depth of thermal damage; I, the laser fluence or intensity or energy; and I½, the fluence or intensity or energy where one half of the maximum shrinkage or depth of thermal damage is measured. The threshold value (IT) was defined as the point where 90% of Smax is reached: IT=1.95 × I½.
We used this 2-state model to describe the data because they appeared to follow a sigmoid curve. This model implies a 2-state situation in which (in state 1) the intensity of the laser is not enough to cause shrinkage or thermal damage, or (in state 2) the laser intensity is sufficient to induce shrinkage or thermal damage. The data were fit to the mathematical model using MacCurve Fit software (V.1.0.7, Kevin Raner, Victoria, Australia). This program finds the best-fit parameters and the SEs of the parameters.
Tissue samples for histological analysis received 3 passes with the respective lasers because this correlates with most clinical CO2 laser resurfacing in humans. Debridement was done between each laser pass on all the histological samples with sterile saline solution and a cotton swab. Tissue specimens were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with either hematoxylin-eosin or Gomori trichrome for light microscopic evaluation. Analysis was performed on a light microscope (Vanox AH-2, Olympus, Lake Success, NY). Morphometric analysis was performed with planar morphometry software (Southern Micro Instruments, Atlanta, Ga).
Tissue contracted linearly with the number of passes for both lasers, as previously described.4 The short-pulsed laser within its typical operating range of 250 to 500 mJ gave a maximum of 3.6% shrinkage per pass with a threshold of 220 mJ when aggressive debridement was used (Table 1 and Figure 1, A). Without debridement, the maximum shrinkage per pass was 2.3% per pass with a threshold of 990 mJ.
Within the scanner laser's typical operating range of 5 to 8 W, the observed plateau shrinkage per pass was 5.1% with a threshold of 2.7 W when aggressive debridement was used (Figure 1, B). This 5.1% shrinkage plateau per pass is seen below the energy settings typically used clinically with the scanner laser. When simple rehydration of the tissue without debridement was done, the observed plateau of shrinkage per pass was 2.4% with a threshold of 6 W. Using a 4×4 gauze for debridement increased the plateau of shrinkage to nearly 9% per pass (data not shown). The threshold did not measurably change from 2.7 W.
Because results obtained in vitro may differ markedly from the clinically observed in vivo results on patients, we confirmed our in vitro human skin sample results by doing in vivo resurfacing experiments with live pigs. Figure 2, A, shows the data for the porcine skin shrinkage with the short-pulsed laser using debridement between passes. The maximum shrinkage in the pig model was 2.6%, and the threshold was 340 mJ. The maximum shrinkage observed was only slightly lower than the maximum shrinkage of 3.6% seen in the human skin model. Results seen with the scanner laser in the pig model, also using debridement between passes, are shown in Figure 2, B. Here, we observed a maximum shrinkage level of 3.7% per pass with a threshold of 3.3 W for the pig skin contraction. Again, this is slightly lower than the maximum shrinkage of 5.1% observed in the human skin experiments.
Morphometric evaluation of the histological specimens was performed on the skin sections stained for collagen using Gomori trichrome. Collagen denaturation was observed by dermal tincture change. The amounts of collagen change observed histologically using the short-pulsed laser are shown subthreshold for thermal denaturation (Figure 3, A), at threshold (Figure 3, B), and above threshold (Figure 3, C). Histological results for the scanner laser are shown in Figure 3, panels D (subthreshold), E (threshold), and F (above threshold). Photomicrographs of the results seen in porcine skin are shown below and above threshold for thermal denaturation for the short-pulsed laser (Figure 4, A and B), and the scanner laser (Figure 4, C and D) for comparison. We observed that the depth of collagen denaturation plateaus above a threshold fluence similar to what was observed with tissue shrinkage (Table 1 and Table 2). Interestingly, the threshold for collagen denaturation was lower than the threshold for cutaneous contraction in vivo with both lasers, but the opposite was observed in vitro.
The plateau depth of collagen denaturation observed in the human breast skin treated with the short-pulsed laser was 25 µm (Figure 5, A); the threshold was 320 mJ (Table 1). The scanner laser produced considerably more thermal change. The observed plateau depth of collagen denaturation was 77 µm, with a threshold of 4.2 W (Figure 5, B).
Collagen denaturation observed in vivo in the porcine model was similar to that observed for the in vitro human skin model. For the short-pulsed laser, the plateau depth of collagen denaturation was 35 µm (Table 2). The observed plateau depth of collagen denaturation for the scanner laser was 80 µm. These depths of collagen denaturation were within experimental error in comparing the in vitro and in vivo models.
Our study revealed several interesting observations about cutaneous laser resurfacing. Most intriguing was that these lasers have a threshold energy requirement for both cutaneous contraction and thermal denaturation. Graphs of these studies demonstrate a best fit with the Henderson-Hesselbalch 2-state equation. This implies a 2-state laser-tissue interaction where the laser energy is either insufficient (state 1) or sufficient (state 2) to cause tissue shrinkage or thermal damage. Applying too little energy would be predicted to have an inadequate clinical effect. Once the energy fluence threshold is reached, our study shows that additional energy does no t appear to enhance the laser's acute effect. In fact, additional energy may increase the potential for undesirable outcomes such as hypertrophic scarring, pigmentary change, and persistent postoperative erythema, since more energy is being deposited into the tissue with little or no apparent gain. This excess laser energy may result in increased thermal damage that is not measured acutely with histological methods such as those used in this study, or may result in increased tissue ablation (which we cannot accurately measure at present), leading to a deeper dermal injury.
Many studies have compared the clinical efficacy and adverse effects of various CO2 resurfacing lasers. However, previous studies have not compared the threshold effects of the different resurfacing lasers as done in this study. We found that the scanner and short-pulsed lasers showed cutaneous shrinkage thresholds at energy fluences of 5.9 and 2.5 J/cm2 in vitro and 7.3 and 3.7 J/cm2 in vivo, respectively. We also found the threshold irradiance (intensity) for consistent depth of thermal change in the dermis using the scanner laser to be 6.0 kW/cm2 (4.2 W) in vitro and the threshold for cutaneous contraction was 3.8 kW/cm2 (2.7 W). This corresponds reasonably well to the data of Chernoff et al,7 who noted the threshold for char-free ablation with the scanner laser to be 5.0 kW/cm2 (3.5 W), where ablation was measured as the removal of tissue. Kamat et al13 reported an energy fluence of 4 J/cm2 as the lower limit for histological changes observed secondary to irradiation energy using a continuous-wave CO2 laser. Our data with the scanner and short-pulsed CO2 lasers also correspond well to their data.
Our threshold values correspond with the accepted operating range of the TruPulse laser of 250 to 500 mJ with an energy fluence of 2.78 to 5.56 J/cm2. Our measured depth of collagen change was 25 to 35 µm with the TruPulse laser. This is slightly less than the 50- to 100-µm change observed when nitroblue-tetrazolium chloride staining was used by Smith et al.14 Staining sensitivity differences may account for this variance. The clinical operating range of the SilkTouch laser is from 5 W to 8 W,2 with a corresponding energy fluence of 10.99 J/cm2 to 17.58 J/cm2, well above our observed 7.3 and 5.6 J/cm2 in vivo thresholds for shrinkage and thermal denaturation, respectively. The use of higher energy fluences than necessary with resurfacing CO2 lasers may explain some of the observed undesirable clinical adverse effects.
Energy fluences required for threshold clinical effect with resurfacing lasers would be predicted to be higher than our observed in vitro study due to differences in live tissue such as the cooling effect of blood flow. Our in vivo data using porcine skin confirmed this expectation. We found 35% and 18% higher fluence thresholds for contraction with the pulsed and scanner lasers, respectively. We also observed a 28% and 27% reduction in the amount of shrinkage per pass in porcine skin (in vivo) compared with human skin (in vitro) for the pulsed and scanner lasers, respectively. This is also expected, as the in vivo skin is surrounded by normal skin that offers resistance to contraction, in contrast to the in vitro skin, which lacks this resistance. Our human skin in vitro model approximated the in vivo acute effects of resurfacing lasers sufficiently well to be helpful for the prediction of clinical cutaneous effects with various resurfacing lasers.
Methods of debridement may cause wide variations in the effectiveness of CO2 laser resurfacing. For instance, we found that debridement with a saline-soaked gauze pad was better than using cotton swabs, which in turn was better than simple rehydration. Our data establish the concept that vigorous debridement of the desiccated tissue between laser passes will increase the amount of shrinkage obtained per pass and decrease the threshold energy fluence required for adequate shrinkage. This concept requires clinical testing for correlation with the cosmetic results.
In summary, our data predict that the clinical effects of the CO2 resurfacing lasers will be similar when operated at their respective threshold levels. Clinical adverse effects may be decreased by operating at these lower energy fluences. Clinical studies substantiating the threshold levels for each resurfacing laser are therefore necessary. For purposes of comparison, we suggest that clinical laser resurfacing studies should use lasers at the calculated threshold levels, and control for the amount of débridement.
Accepted for publication April 22, 1998.
These studies were supported by grants from the Department of Defense Medical Free Electron Laser Program administered through the Office of Naval Research (N00014-94-1023) (Drs Reinisch and Ellis), Arlington, Va, and the National Institutes of Health, Bethesda, Md (P30 AR41943).
The TruPulse laser was loaned from Tissue Technologies, Albuquerque, NM, and the SilkTouch handpiece was loaned from Sharplan, Allendale, NJ.
Presented in abstract form at the 18th Annual Meeting of the American Society for Laser Medicine and Surgery Meeting, San Diego, Calif, April 6, 1998.
The expert technical assistance of Lillian B. Nanney, PhD, and Mary McKissack is gratefully acknowledged.
Reprints: Darrel L. Ellis, MD, 3900 The Vanderbilt Clinic, Nashville, TN 37232-5227.