RANDOMIZED, PAIRED COMPARISON OF NO-STING BARRIER FILM VERSUS SORBOLENE CREAM (10% GLYCERINE) SKIN CARE DURING POSTMASTECTOMY IRRADIATION

Transcription

1 doi: /s (03) Int. J. Radiation Oncology Biol. Phys., Vol. 58, No. 1, pp , 2004 Copyright 2004 Elsevier Inc. Printed in the USA. All rights reserved /04/$ see front matter CLINICAL INVESTIGATION Breast RANDOMIZED, PAIRED COMPARISON OF NO-STING BARRIER FILM VERSUS SORBOLENE CREAM (10% GLYCERINE) SKIN CARE DURING POSTMASTECTOMY IRRADIATION PETER GRAHAM, F.R.A.N.Z.C.R.,* LOIS BROWNE, PH.D.,* ANNE CAPP, F.R.A.N.Z.C.R., CHRISTOPHER FOX, F.R.A.N.Z.C.R., JENNIFER GRAHAM, B.H.SC.,* JULIE HOLLIS, DIP.APP.SC., AND ELIAS NASSER, F.R.A.N.Z.C.R. *Cancer Care Centre, St. George Hospital, University of New South Wales, Kogarah, New South Wales, Australia; Cancer Care Centre, Wollongong Hospital, University of New South Wales, Wollongong, New South Wales, Australia Purpose: Postmastectomy irradiation provides an excellent model for irradiated skin care practices because of the relatively uniform surface and radiation compared with other situations in which radiation-induced moist desquamation is common. We designed a study to test the effect of prophylactic 3M Cavilon No-Sting Barrier Film (No-Sting) on the rates of moist desquamation compared with sorbolene cream (with 10% glycerin). Methods and Materials: The irradiated chest wall was divided into medial and lateral halves. Sixty-one women were randomized to have No-Sting applied to either the medial or lateral half, with the alternate half treated with sorbolene. Results: For all patients, the skin toxicity, calculated as the area under the curve, mean No-Sting and sorbolene score was 8.1 vs. 9.2, respectively (p 0.005, Wilcoxon signed rank test). The total number of weeks of moist desquamation for the 61 patients was 40 vs. 45, equating to a mean of 0.65 week vs week per patient in the No-Sting and sorbolene treated areas, respectively. The rates of moist desquamation were 33% vs. 46% (p 0.096, McNemar s Exact test). For 58 fully assessable patients (minimum of 7 weekly observations), the area under the curve and rates of moist desquamation were significantly different statistically (p and 0.049, respectively). No statistically significant differences were noted in the pain scores. The pruritus scores were significantly reduced in the No-Sting area (area under the curve, p 0.011). Conclusion: No-Sting reduces the duration and frequency of radiation-induced moist desquamation Elsevier Inc. Radiation, Toxicity, Skin, Barrier. INTRODUCTION Skin reactions are a common and undesirable result of radiotherapy (RT). The postmastectomy skin surface is relatively uniform and a good test bed for skin care protocols. We hypothesized that skin abrasion contributes to the overall reaction. Therefore, the prolongation of the life of the intact keratinized surface may increase the chance that the underlying stem cell pool would regenerate sufficiently to reduce the likelihood of moist desquamation. An additional possibility is that maintenance of the intact surface would help to maintain both moisture levels and a scaffold for migration/repair to maximize the efficiency of the repair processes. In other words, modification of the physical environment of the skin rather than the radiobiology of the radiation injury may modify the phenotypic expression of the radiation injury. Reprint requests to: Peter Graham, F.R.A.N.Z.C.R., Cancer Care Centre, St. George Hospital, Gray St., Kogarah, NSW 2217 Australia. Tel: ; Fax: ; grahamp@sesahs.nsw.gov.au Acknowledgments This was an investigator-initiated study; 3M is A barrier film may physically retard the normal desquamation rate by reducing the effects of abrasion. A number of skin protectors are used to reduce the trauma to skin from adhesive tape. No-Sting is a superior product by a factor of at least two compared with five other similar products for this purpose, as measured by transepidermal hydration after repeated adhesive tape applications and stripping (1). We selected this product to compare with our standard skin care using sorbolene cream. Sorbolene is a generic label given to skin barrier creams marketed in Australia composed of a mixture of water and oils that generally contain 10% glycerin. METHODS AND MATERIALS On the basis of previously recorded skin reactions in our prospectively recorded RT toxicity database, the baseline acknowledged for support provided in the form of 3M Cavilon No-Sting Barrier Film at no cost to the study. Received Feb 11, 2003, and in revised form Jun 3, Accepted for publication Jun 23,

2 242 I. J. Radiation Oncology Biology Physics Volume 58, Number 1, 2004 Fig. 1. Schematic of test layout. rate of moist desquamation for postmastectomy RT at the St. George Cancer Care Centre was taken as 60%. The sample size was chosen to provide a power of 80% to detect a reduction in moist desquamation rates to 30%. The South Eastern Sydney and Illawarra Area Health Service ethics committees approved the study, and all participants provided written informed consent. Sixty-one patients were recruited at the RT departments of St. George (n 50) and Wollongong (n 11) Hospitals. The postmastectomy chest wall was divided into medial and lateral compartments approximately about the mid-clavicular line (Fig. 1). These areas were randomized to have either No-Sting Film or sorbolene applied. The internal control method accounted for interpatient variation such as the use of chemotherapy, and the internal randomization was to protect against the potential for the skin reaction to be systematically worse in either the medial or lateral compartment. Randomization was by telephone to a separate data management unit. The film was applied twice weekly if allocated to the medial compartment and three times weekly if allocated to the lateral compartment. This was because of pilot experience demonstrating the accumulation of the product applied more frequently to the medial compartment. No-Sting Film was applied by the nursing staff to ensure both compliance and correct and even placement of the colorless film. Sorbolene was applied twice daily. On RT days, the nurses also applied the sorbolene once daily, with patients applying the sorbolene at other times. The products were applied from the start of RT until 2 weeks after completion. For those skin areas that developed moist desquamation, the skin care in that area was changed to a hydrocolloid dressing. RT to the chest wall was administered using 6-MV photon beams applied tangentially. The prescribed chest wall dose was 50 Gy in 25 fractions. All patients had a 1-cm bolus applied daily. This was either to the entire chest wall (n 17) or to a strip 4 cm wide on either side (8 cm total width) of the full length of the scar (n 44), at the discretion of the oncologist. Of the 61 patients, 49 received supraclavicular fields and 17 axillary fields. Six patients had an electron boost of 10 Gy in five fractions applied to a chest wall recurrence excision. These patients were only eligible if the electron area was split evenly by the No-Sting and sorbolene surfaces. Physician skin scoring assessments were made weekly using the Radiation Therapy Oncology Group acute skin scores. Patients completed 5-point Likert scales (minimal value 1, maximal value 5) weekly for pain and pruritus for each area of the skin. Polaroid photographs were taken weekly. Assessments were from Week 1 until complete resolution of the skin reaction or symptoms or to a maximum of 12 assessments. If the skin reaction had resolved completely before Week 12, subsequent assessments to Week 12 were allocated zero scores. Outcomes were analyzed by both dichotomous (threshold) and area under the curve comparisons. Threshold comparisons were tested using McNemar s exact test for paired categorical data. Area under the curve values were obtained by summing the values of the weekly scores. Each week was considered as one unit of time. The Radiation Therapy Oncology Group skin scores provide a value for the area (height in Radiation Therapy Oncology Group score 1 week of time) for that unit of time. If the reactions were scored as 0, 0, 0, 1, 1, 1, 2, 2, 1, 0, 0, 0 for Weeks 1 12, the sum of these provides a value of 8 (total area under the curve provided by a histogram plot of the weekly scores). These were compared using the Wilcoxon signed-rank test. For 14 patients, a full set of three thermoluminescent dosimeter and GAF film (GAF Chemicals Corp., Wayne, NJ) dose measurements were obtained at 12 positions across the chest wall along the central axis as part of a concurrent dosimeter comparison study. For those patients in whom a strip bolus was used, because this followed the scar superolaterally, the bolus moved away from the central axis. This report was not concerned with the dosimeter performance, but a summary of the thermoluminescent dosimeter results is presented in brief to demonstrate the relative dosimetry across the chest wall on the central axis. The equal and random allocation of the two skin care products to the medial and lateral compartments protected against systematic variations in dose among other factors such as friction areas to account for the variation in skin reactions. RESULTS The mean patient age was 58 years (range 30 88). The treatment area was left sided in 27 patients. Forty-one patients were postmenopausal. Eight were Asian, 1 was black, and the remainder were white. Thirty-seven patients received chemotherapy, which was administered, at least in part, concurrently for 13 and entirely sequentially for 24. Nineteen patients received anthracycline chemotherapy, but none received this concurrently with RT. The mean RT dose was 49.7 Gy (range 38 60). Only 2 patients did not complete RT as planned. Both of these stopped early because of skin reactions. Both were white and had received concurrent chemotherapy. Four patients had interruptions in RT. In

3 Radiotherapy skin care P. GRAHAM et al. 243 Table 1. Analysis of all patients recruited (n 61) Outcome Area under curve p RTOG skin score Sorbolene * No-Sting 8.1 Pain Sorbolene * No-Sting 10.6 Pruritis Sorbolene * No-Sting 11.9 % Skin score 2 Sorbolene No-Sting 33 Pain score 1 Sorbolene No-Sting 36 Pruritis score 2 Sorbolene No-Sting 21 * Wilcoxon signed rank test. Exact McNemar s test. two, this was primarily because of the skin reaction in the No-Sting area. Thirty-one patients were randomized to medial application of No-Sting and 30 patients to lateral application of No-Sting. The results for the analyses for all predefined end points using all patients (intention-to-treat analysis) are summarized in Table 1. The analyses of the fully assessable patients, excluding those who had no data available beyond 6 weeks, are summarized in Table 2. Irrespective of the Table 2. Analysis of fully assessable patients (n 58) Outcome Area under curve p RTOG skin score Sorbolene * No-Sting 8.4 Pain Sorbolene * No-Sting 10.9 Pruritis Sorbolene * No-Sting 12.3 % Skin score 2 Sorbolene No-Sting 33 Pain score 1 Sorbolene No-Sting 34 Pruritis score 2 Sorbolene No-Sting 21 * Wilcoxon signed rank test. Exact McNemar s test. analysis, for skin reaction and pruritus, the scores were less in the No-Sting areas. No evidence was found of a statistically significant trend in the relationship to pain scores or analgesia requirements. Nine patients required analgesics for skin reaction in the sorbolene area and 8 patients for the No-Sting area. Figure 2 shows the mean dosimeter values across the chest wall if all locations were under full bolus (standard deviation [SD] 0.196). Figure 3 shows the overall mean dosimeter values per fraction, including locations proximate to the beam edge or not under bolus (SD 0.34). It is clear that the dosimetry across the chest wall did not vary systematically, except by bolus status. Although the dose on the central axis seemed lower overall laterally, the bolus did extend from the medial to the lateral borders, but not always on the central axis at which the dosimeter strips were located. Sixty-four dose observations were made in areas without a bolus. The mean dose per fraction in the sorbolene-treated skin was 1.46 Gy (SD 0.219) and in the No- Sting treated skin was 1.50 Gy (SD 0.257), not clinically or significantly different statistically (p 0.55, t test). The values of the area under the curve and the rates of moist desquamation by side and skin care product are summarized in Table 3. None of the differences were statistically significant, although the trend appeared to be for a greater skin reaction in the lateral compartment to be somewhat reversed in the No-Sting group. DISCUSSION Preventive measures are often used with the aim of reducing radiation-induced skin reactions, although few controlled trials have been done. Prophylactic topical ascorbic acid vs. placebo was tested in 85 patients receiving cranial RT without a significant difference found (2). Aloe vera has not been shown to be superior to placebo in randomized double-blind comparisons to either placebo gel or no treatment (3). Topical hyaluronic acid in a mixed population was shown in a double-blind randomized trial to reduce peak radiation skin reactions and global efficacy (4). The possible limitations of the hyaluronic acid study included minor imbalances between the treatment allocation arms and the lack of a regression analysis to exclude other skin reaction predictors having influenced the result of the study. Nonetheless, hyaluronic acid has a number of properties that could have contributed to the positive clinical trial result. Hyaluronic acid combines with fibrin to provide a mechanical support favoring migration of the cells involved in repair. Hyaluronic acid stimulates migration of phagocytes, fibroblasts, and endothelial cell proliferation (5 8). An apparently important physicochemical property is its ability to avidly retain water (9). The primary hypothesis for the possible benefit of No- Sting in irradiated skin was by the reduction in the superficial skin layer loss owing to normal abrasion. The apparent pattern (not statistically significant) of reactions that were a little worse in the lateral compartment being reversed by

4 244 I. J. Radiation Oncology Biology Physics Volume 58, Number 1, 2004 Fig. 2. Mean thermoluminescent dosimeter dose per fraction measures on central axis medial (position 1) to lateral (position 12) where under full bolus. No-Sting would fit with the hypothesis that No-Sting is effective by the mechanism of reducing surface attrition owing to friction, because this is thought to be greater in the lateral compartment. Although we have not tested the mechanical support model for enhanced cellular migration of cells, if the skin surface remains intact, it is reasonable to predict more efficient repair underneath. In addition, although No-Sting is not aimed at binding water, it has been shown to reduce transepidermal water movement. Irrespective of the underlying mechanism, this study has provided strong evidence that No-Sting Film does reduce the phenotypic expression of radiation injury, even though no radiation protection is present at the molecular or cellular level. No evidence exists to indicate whether sorbolene should be applied more or less frequently to minimize skin reactions. This study simply continued what was standard practice in our institutions for the use of sorbolene. The film, in contrast, is designed to last several days. More frequent application of the film would result in accumulation and tend to cause the film to crack and peel, thus reducing its potential effectiveness. Irrespective of whether the application frequency was optimal for both products, in this study, using the specified applications, the skin reactions were less in the film-treated areas. This study continued the use of a standard product in a standard manner as the control and tested against this a new product in what was thought to be Fig. 3. Mean thermoluminescent dosimeter dose per fraction measures on central axis medial (position 1) to lateral (position 12) for all measures, regardless of bolus.

5 Radiotherapy skin care P. GRAHAM et al. 245 Table 3. Pattern of skin reaction by skin compartment Compartment Product n AUC Moist desquamation (%) Lateral Sorbolene No-Sting Medial Sorbolene No-Sting Abbreviation: AUC area under curve (score). the method most likely to obtain benefit from the product. It is possible that different application frequencies may have resulted in a different result; however, that does not alter the interpretation of the current study or invalidate the result. This investigation suggests that the film is better than sorbolene as applied in this study, but it is possible that neither is superior to no prophylactic skin care. We are unaware of any comparison of sorbolene to no treatment. However, because sorbolene was the existing standard of care in our departments, it would have been ethically difficult to have a no-skin-care arm. A no-treatment area would have also been more difficult to have patients agree to participate, particularly when patient information booklets and brochures provided to our patients, both internally developed and provided by various state cancer information bodies, describe the use of creams for skin care. Therefore, although scientifically it may be purer to have a no-treatment area, this was not a viable design within our departments. However, irrespective of the value of each compared to no treatment, this was still a valid comparison of the two products. Sorbolene was not applied in sufficient amounts to provide a bolus effect and does not accumulate to any significant bolus thickness. The skin dosimetry data in the nonbolus areas confirmed that no difference was present in the buildup effect (or lack thereof) between sorbolene and No- Sting in those areas outside the skin covered with the bolus material. No differences were noted in the dose in the areas without bolus to explain the results of this study. The study protocol did not prospectively define the minimal number of assessments below which a patient was not assessable for skin response. It is very clear from the data, however, that peak skin reactions occurred several weeks after RT completion (Fig. 4). Including patients for analysis who did not have evaluations beyond Week 6 would therefore underestimate the occurrence of skin reactions and also underestimate the potential effect of the No-Sting Barrier Film. We therefore presented analyses for all patients (intention-to-treat analysis), irrespective of the number of assessments, as well as for the 58 patients, with the exclusion of 3 patients who had only 2 6 weeks of assessments. The primary reason for the missing data on these patients was failure of the treating physician to record the skin assessments. Once this was detected by trial monitoring, no additional patients treated by that physician were recruited to the study to avoid the problem of missing data eroding the capacity of the study to test the hypothesis as designed. Fig. 4. Mean weekly physician skin scores. The reduction in pruritus scores was counterintuitive. Because the No-Sting Film has a dry feel on the skin, and emollients that No-Sting Film lacks are commonly used to treat skin pruritus, it was expected that the pruritus scores might increase in the No-Sting treated areas. The lack of impact on pain may relate to the altered sensation in the postmastectomy skin flaps. Moist desquamation was selected as the cutpoint for the dichotomous analysis of skin reactions because of its clinical relevance and significance. There were no a priori data on patient pruritus or pain in this clinical situation. The cutpoints for these outcomes were chosen to divide patients into similar size groups for analysis. The area under the curve analyses were not subject to potential bias in the selection of cutpoints for analysis and were highly consistent with the dichotomous analyses. The reason for a lower moist desquamation rate in our sorbolene-treated area compared with historical controls may relate to the reduced extent of bolus used, because the clinician who treated most of the patients who provided the historical data to design the study used full chest wall bolus, but his patients comprised 10% of those recruited. This study was not designed to assess the clinical value of the bolus and how much to use in this clinical context. The use of bolus in Australia is highly variable, although most departments and clinicians use bolus (10). Although this may influence the size of the benefit to be obtained in this clinical scenario, this study was intended to model the potential benefit that may in turn be applied to any skin area at significant risk of moist desquamation, such as in head-and-neck RT or radiation to inguinal areas. The presentation of both dichotomous and area under the curve analyses was chosen for the following reasons. Frequency of a threshold skin reaction score is readily understood by clinicians and is more easily understood in terms of the size of the effect. However, exceeding a threshold may mask variations in the size of the area involved and the time for resolution of that reaction. Area under the curve analyses reflect both the degree of reaction and the duration of the reaction, which would generally reflect the overall injury manifestation, al-

6 246 I. J. Radiation Oncology Biology Physics Volume 58, Number 1, 2004 Table 4. Cost effectiveness of sorbolene vs. No-Sting Film skin care Patients (n) Skin care Cost ($AUS) Weeks dressings (n) Dressings cost* ($AUS) Average cost per patient* ($AUS) 100 Sorbolene $ ( ,585) 147 (95 235) 100 No-Sting $ ( ,838) 145 (99 222) Abbreviation: $AUS Australian dollars. * All nursing care administered in outpatient department; range (in parentheses) from no nursing input to community nursing home visits. though clinicians may not have the same intuitive appreciation of the resulting scores. This study was designed using the threshold of moist desquamation, because no previous data were available for the area under the curve analyses. To illustrate the immediate clinical care impact, if the difference found is true and is extrapolated to 100 patients, 9 weeks of dressings required for moist desquamation could be avoided. Individual patient cost data were not prospectively recorded; however, estimations of relative costs can be made (Table 4). Using our current moist desquamation management practices, the No-Sting strategy costs 1% less than the sorbolene strategy, at an average cost of $145 AUS vs. $147 AUS per patient if all nursing care is administered to our outpatients. Depending on the amount of nursing resources allowed for (ranging from the unlikely value of zero, to 20 min in outpatients, to community nursing visits in the home), the use of No-Sting would range from a 4% increase to a decrease of 6% in overall skin care cost. Our usual policy is to arrange for community nurses to perform dressings once the RT course is completed, so No- Sting in our system would represent a cost savings. The costeffectiveness of No-Sting Film may vary according to local moist desquamation management practices and health care systems. Apart from monetary considerations, this would have significant nursing time (20 outpatients to 45 community hours per 100 patients) and patient convenience implications. Overall quality-of-life comparisons were not possible with this study design, because each patient was their own control; however, we have reported on those symptoms specific to the treated skin areas most commonly reported by patients. The substantial advantage of this particular study design compared with previous randomized skin care trials for radiation skin reactions was the use of the internal control to exclude the effect of other variables influencing the findings. In addition, a very homogenous test bed of postmastectomy patients was used, rather than skin in multiple different sites such as the head and neck or groin or patients with very variable radiation techniques, doses, and fractionations. Both of these aspects make the findings quite robust. Unfortunately, because of the very different appearance and application frequency between the two products, it was not possible to blind the study. Because of these limitations, although we would recommend No-Sting as an effective form of skin care during RT, the study will be repeated as a larger multicenter Trans-Tasman Radiation Oncology Trial with several enhancements. Cavilon No-Sting Durable Barrier Film is now available in a cream format with emollients, which is applied more frequently. This facilitates a double-blind design comparing CBDC with sorbolene. Three hundred patients will be recruited with digital imaging for additional blinded scoring. REFERENCES 1. Grove CL, Leyden JJ, Lutz JB, et al. Comparison of the skin protectant properties of various film forming products. Data on file 3M. 2. Halperin EC, Gaspar L, George S, et al. A double-blind, randomised prospective trial to evaluate topical vitamin C solution for the prevention of radiation dermatitis. Int J Radiat Oncol Biol Phys 1993;26: Williams MS, Burk M, Loprinzi CL, et al. Phase III doubleblind evaluation of an aloe vera gel as a prophylactic agent for radiation-induced skin toxicity. Int J Radiat Oncol Biol Phys 1996;36(Suppl. 1): Liguori V, Guillemin C, Pesce GF, et al. Double-blind, randomised clinical study comparing hyaluronic acid to placebo in patients treated with radiotherapy. Radiother Oncol 1997; 42: Trabucchi E, Preis B, Baratti C, et al. Topical treatment of experimental skin lesion in rats: Macroscopic, microscopic and scanning electron-microscopic evaluation of the healing process. Int J Tissue React 1986;8: Venturini D. Osservazioni cliniche sull impiego dell acido ialuronicao in terapia dermatological. G Ital Dermatol Venereal 1985;120: Weigel PH, Fuller GM, LeBoeuf RD. A model for the role of hyaluronic acid and fibrin in the early events during the inflammatory response and wound healing. J Theor Biol 1986; 119: West DC, Hampson IN, Arnold F, et al. Angiogenesis induced by degradation products of hyaluronic acid. Science 1985;228: Nakamura M, Nishida T. Recent developments in the use of hyaluran in wound healing. Exp Opin Invest Drugs 1995;4: Delaney G, Blakey D, Drummond R, et al., on behalf of the Faculty of Radiation Oncology, Radiation Oncology Advisory Group to the National Breast Cancer Centre. Breast radiotherapy: An Australasian survey of current treatment techniques. Aust Radiol 2001;45: