Fading of nipple areolar reconstructions: the last hurdle in breast reconstruction?

Transcription

1 British Journal of Plastic Surgery (2002) 55, The British Association of Plastic Surgeons. Published by Elsevier Science Ltd. All rights reserved. doi: /bjps Fading of nipple areolar reconstructions: the last hurdle in breast reconstruction? N. R. Dean, T. Neild*, J. Haynes, C. Goddard and R. D. Cooter Department of Plastic and Reconstructive Surgery, Royal Adelaide Hospital; *Department of Human Physiology, Flinders University; Department of Anatomical Sciences, Adelaide University; and Co-operative Research Centre for Tissue Growth and Repair, GroPep Ltd, Adelaide, Australia SUMMARY. Fading of nipple areolar reconstructions is commonly reported, but there are few formal studies of this phenomenon. The purpose of this study was to determine whether deficiencies in nipple areolar reconstruction and pigmentation were perceived by patients, their partners and independent observers, and whether a technique could be developed to measure nipple areolar colour reliably. A total of 57 patients, 32 partners and four independent observers completed questionnaires about the appearance of the patients breast reconstructions in general and specifically about their nipple areolar reconstructions. Scores for the general attributes of the breast reconstruction were used as internal controls for the scores of the nipple areolar reconstruction. A computer software package was developed to analyse colour in photographs of the reconstructions. Independent observers thought that nipple areolar reconstruction improved the appearance of a breast reconstruction 81% of the time. Considerably fewer patients were happy with their nipple areolar colour than were happy with the more general attributes of the breast reconstruction (P 0.005). Colour analysis objectively demonstrated measurable mismatch between normal and reconstructed nipple areolar skin, which was positively correlated with time since surgery due to fading of the nipple areolar reconstruction. In our patients, the quality of nipple areolar reconstruction, in particular its pigmentation, is seen as inferior to that of the rest of the breast reconstruction in the eyes of patients, their partners and independent observers. The poor colour match and fading of reconstructed nipple areolar skin are phenomena that can be measured using colour analysis The British Association of Plastic Surgeons Keywords: nipple areolar complex, pigmentation, breast reconstruction, melanocytes. Nipple areolar reconstruction is a beneficial component of breast reconstruction. 1 Surgeons report that nipple areolar reconstructions, both with and without tattoos, fade over time. 2 5 We investigated perceptions of nipple areolar colour in patients undergoing breast reconstruction, and sought to devise a practical method of measuring colour and fading in nipple areolar reconstructions. The first part of our study was questionnaire based and, in order to compensate for the variation that normally occurs between subjects scores regardless of any real variation in opinions (e.g. dependent on personality or experience with questionnaires), we chose other attributes of the breast reconstruction as internal comparators for the subjects ratings of the nipple areolar reconstruction. The second part of the study employed specially developed computer software to standardise images of the patients and to measure the hue and saturation of the nipple areolar and breast skin. Techniques of nipple areolar reconstruction The skate flap 6 (Fig. 1) is one of many local flaps described for reconstructing the nipple areolar complex. 4,7 14 Skin Presented in part at the Royal Australasian College of Surgeons Annual Scientific Congress, Melbourne, Australia, May flaps are raised from a disc on the reconstructed breast mound and used to reform the prominence of the nipple. This leaves a denuded area around the new nipple, which requires a skin graft to provide neo-areolar skin. In the quest to produce a good nipple areolar reconstruction without skin grafting (hence avoiding donor wounds), several techniques have been developed involving intradermal tattooing as the only adjunct to the local flaps for nipple reconstruction. 3,15 17 However, shape distortion and loss of nipple areolar projection have been reported as drawbacks of these methods. 16,18 Intradermal tattooing has also been used in conjunction with nipplesharing techniques. 5 Patients and methods This study was designed as a clinical review of patients who had undergone breast-mound and nipple areolar reconstruction. The study was approved by the Ethics Committee of the Royal Adelaide Hospital, and all participants gave written informed consent for the study and for clinical photography. All living patients who had had nipple areolar reconstruction performed by one surgeon (RDC) were invited to participate. This surgeon did not use nipple areolar tattooing. Four of these patients could not be traced, and, of the remaining 62 patients, 57 (92%) agreed to participate. 574

2 Fading of nipple areolar reconstruction 575 In the questionnaire, patients were asked about specific attributes of their breast and nipple areolar reconstructions, followed by two more general satisfaction questions. Patients were first asked what they thought of symmetry in a bra, symmetry naked, appearance in a bra and appearance naked. These parameters were taken from a previous study examining the outcome of breast reconstructions. 20 The responses to these four questions were combined to give a rating for general attributes. The other questions asked the patient to rate nipple areolar colour, fading of the nipple areolar reconstruction, nipple areolar projection, loss of projection over time, satisfaction with the breast-mound reconstruction and satisfaction with the final result of the breast reconstruction. Figure 1 Skate-flap reconstruction. Of these 57 patients, 10 agreed to complete questionnaires only, and 47 participated in the full study protocol. Demographics and background information The mean age of the patients was 51 years; 43 were married, 10 were divorced or separated and four were single. Thirty-eight patients reported having a regular sexual partner, 15 reported having no sexual partner and four did not answer the question. There was a higher percentage of professional, associate professional and managerial women in the study than would be expected in the general population of Australian women. 19 The TRAM flap had been used for breast-mound reconstruction in 39 patients (37 free and two pedicled flaps), 14 patients had had implant reconstructions, two had undergone latissimus dorsi reconstructions, and two patients had required no breast-mound reconstruction because they had undergone small central resections of the breast. Eight patients had undergone bilateral breast reconstructions. In total, 55 patients had had nipple areolar reconstruction using a skate flap; the remaining two patients had undergone other local flap techniques. Patients questionnaire The patients were reviewed individually in the Department of Plastic and Reconstructive Surgery at the Royal Adelaide Hospital, and asked to complete a questionnaire on aspects of their breast reconstruction. There were 10 questions in total, and patients were asked to respond by making a mark on a 10 cm visual analogue scale with opposing responses at either end. Photography After the patients had completed their questionnaires, clinical photographs were taken using standard views with and without a bra (Fig. 2). These photographs were used by a panel of four independent observers to assess attributes of the breast reconstructions. In addition, one photograph with a colour standard included was taken for colour analysis (see below) (Fig. 3). All photographs were taken under standard lighting conditions with the same photographic background. All photography was performed by professional clinical photographers in a studio, using a Nikon (Japan) F5 camera with a 60 mm/f2.8 micronikkor lens. Independent observers and partners assessments Four people working closely with our unit, but with no knowledge of the study, were asked to act as independent observers. Two were male and two were female, and they included a nurse, a physiotherapist, a postgraduate student of anthropology and a ward clerk. The observers reviewed 10 slides of each patient, and rated various attributes of the breast reconstructions on questionnaires similar to those used by the patients. Slides of each patient before and after nipple areolar reconstruction were also shown where available (n 37) to allow the impact of the nipple areolar reconstruction on the completed breast reconstruction to be rated (Fig. 4). Slides were shown in a standardised manner with a set time allocated to each set of slides. Questionnaires were used by the patients partners to record their assessments of the breast reconstructions. These questionnaires were taken home by the patients and returned by post in prepaid envelopes. Partners were not asked to rate fading of the nipple areolar reconstruction over time because many of them had not initially been allowed to view the reconstructions. Scoring of questionnaires The position of the mark on the analogue scale was measured in mm from the end representing an undesirable outcome. Thus, for all measures a low rating indicated an unfavourable outcome: for example, for the attribute of fading of the nipple areolar complex, a mark towards the no, not faded at all end of the scale translated as a high

3 576 British Journal of Plastic Surgery Figure 2 (A I) Standardised photography for the assessment of breast reconstruction by independent observers.

4 Fading of nipple areolar reconstruction 577 the usual clinical goal (rather than producing a high mean score, which is clinically meaningless). Colour analysis The photographs that included the colour standard were scanned into a computer to enable standardisation and comparison of colour between images. Using a customdesigned computer programme, 21 the image colours were adjusted according to the colours within the colour standards. Colours in selected regions of the image were then measured, and expressed in terms of hue, saturation and value. Hue distinguishes between colours in terms of their dominant wavelength, e.g. between red (360) and yellow (420) (Fig. 3B). Normal Caucasian skin has a hue of around 380 (between red and yellow). The hue represents a position, in degrees, around a colour wheel, with red at 0 /360, yellow at 60, green at 120 and blue at 240. The 0 /360 discontinuity at red falls within the range of skin colours, so we have modified the scale slightly by allowing numbers greater than 360, up to 420 ( : yellow). Saturation indicates the difference between colours such as red and pink. Bright red is highly saturated, with a saturation of 1; pink has the same hue but a lower saturation. Fading of a colour is typically seen as a decrease in saturation. Value, the third parameter, describes dullness. Well-lit areas of an image have a value close to 1, but in the shadows value falls and colours become dull. Value is needed to describe the colour completely, but in practice it depends more on the lighting than on the properties of the skin. Statistical analysis A 2 test was used to compare the number of patients happy with nipple areolar colour and the number of patients happy with the general attributes of breast reconstruction. For other results, where measurements were paired (e.g. comparing normal with reconstructed breasts), paired Student s t-tests were used, with the twotailed P value given; where data were unpaired, either two-sample t-tests or analysis of variance was used, depending on the number of variables. Regression analysis was used to analyse the results over time. Figure 3 (A) Image of patient with colour standard in colour-analysis software program and (B) colour wheel. rating and a mark towards the yes, faded a lot end of the scale translated as a low rating. Ratings were grouped as shown in Table 1. Those patients who gave ratings of greater than 75 in at least three out of the four general attributes were regarded as being happy with the general attributes of the breast reconstruction. For statistical analysis, the frequency of patients happy with the result was used because producing results that patients are happy with is Results Time since surgery and method of reconstruction The mean time since nipple areolar shape reconstruction was 25 months (standard deviation: 16 months; range: 1 63 months). Of the patients who had undergone nipple areolar reconstruction using a skate flap, in 21 a skin graft from the groin had been used to pigment the nipple areolar reconstruction, in 21 a skin graft from the opposite areola had been used and in 13 there had been no deliberate attempt to recreate areolar pigmentation, with nonspecialised redundant skin being used to cover the denuded area of the skate flap.

5 578 British Journal of Plastic Surgery Figure 4 Images of patient (A) before and (B) after nipple areolar reconstruction, for assessing the impact of adding the nipple areolar complex. Figure 5 Patients ratings of the different aspects of breast reconstruction. Table 1 Rating Grouping of questionnaire ratings Group 0 25 unhappy with results moderately unhappy with results moderately happy with results happy with results Patients assessments Analysis of variance of the raw data revealed a highly statistically significant difference between the patients ratings of the different attributes of their breast reconstructions (P 0.001), with ratings for aspects of the nipple areolar reconstruction being lower than those for the other aspects of breast reconstruction. The distribution of the ratings for each attribute is shown in Figure 5. Most patients gave a high rating for satisfaction with their breast reconstruction, and there was no difference between satisfaction with the breast mound and satisfaction with the final result of the reconstruction. Table 2 2 test comparing the number of patients happy with the general attributes of their breast reconstruction and the number of patients happy with their nipple areolar colour General Nipple-areolar Total attributes colour actual number happy with result number not happy with result total expected number happy with result number not happy with result total test: P There was a statistically significant difference between the number of patients happy with the general attributes of the breast reconstruction and the number of patients happy with nipple areolar colour ( 2 test: P 0.005; Table 2). There was a smaller statistically significant

6 Fading of nipple areolar reconstruction 579 difference between the number of patients happy with nipple areolar projection and the number of patients happy with general attributes ( 2 test: P 0.05). For the whole sample, there was no significant correlation between time since the procedure and the patient s rating of nipple areolar colour. In patients who had a deliberately pigmented nipple areolar reconstruction (i.e. excluding those with non-specialised skin grafts), fading was found to be worse with increasing time since the procedure (slope of regression line differing from zero: P 0.005). Patients gave ratings at one time point only, and no conclusions can be drawn about the ratings that individual patients would give over time. Independent observers and partners assessments We received 28 completed questionnaires from patients partners. Three patients did not wish their partners to complete questionnaires, and the remainder either had no partner or failed to return the questionnaire. The patients partners gave similar ratings to the patients themselves, with a similar disparity between ratings for general attributes and ratings for the colour of the nipple areolar reconstruction (Fig. 6). The responses of the panel of independent observers showed a statistically significant difference between the ratings given for different attributes of the breast reconstruction on analysis of variance (P 0.001). Generally, the independent observers gave lower ratings for the breast reconstructions than did the patients (Fig. 6). Fading was frequently given a low rating, as was the colour of the nipple areola complex. The nipple areolar reconstruction was thought by external observers to improve the appearance of the breast reconstruction in 81% of cases, a finding that is unlikely to occur by chance (P using binomial distribution). Colour analysis Slides with colour standards were taken for 45 patients. In one case, there was a fault in the photographic processing that made the slides unusable, and, of the 44 slides scanned, four were found to have too high an error value on the colour-analysis programme to give accurate results. Colour analysis of scanned images of the patients was completed in 40 cases. Of these, three had had bilateral reconstructions; therefore, the reconstructed and normal sides could not be compared. The following data are from the remaining 37 patients. In the sample, there was found to be a statistically significant colour mismatch between the normal and the reconstructed nipple areola complex on paired t-tests. The mean hue was 373 on the normal side and 378 on the reconstructed side (P 0.001). This means that the reconstructed side was, on average, slightly redder than the normal side. The mean saturation was on the normal side and on the reconstructed side (P 0.001), indicating that the reconstructed side was measurably paler than the normal side. The colour saturation mismatch depended significantly on the source of pigmentation for the nipple areolar reconstruction (areolar skin, groin skin or non-specific skin) (P 0.05 on analysis of variance). This can be partly explained by the difference between the nonspecific skin-graft group and the rest of the sample (P 0.01 on two-sample t-test). There were no significant differences in hue mismatch between the different sources of nipple areolar pigmentation. In those patients with pigmented skin grafts (groin or opposite areolar grafts, n 27), the pallor of the reconstructed nipple areolar complex increased with increasing time since the procedure, so that the mismatch in colour saturation between the normal and the reconstructed sides increased over time (Fig. 7). This was measurable and statistically significant on regression analysis (P 0.05). In informal interviews with patients, three patients reported wound-healing problems with the skin graft to the areola, and four reported having had infections or other wound problems at the groin donor site. This equates to a total of seven wound complications in 57 patients (12%). Discussion Several techniques have been described to pigment nipple areolar reconstructions. The opposite areola, 4 the Figure 6 Ratings by patients, their partners and a panel of independent observers. Figure 7 Colour saturation mismatch over time in patients with pigmented skin grafts (n 27).

7 580 British Journal of Plastic Surgery vulva 22 and the buttock 9 have all been used as donor areas for neo-areolar skin. In this study, the patients who had areolar grafts using the skin of the opposite areola were those who had undergone delayed contralateral breast reduction. The use of upper inner thigh skin to replace the areola was proposed by Broadbent et al in 1977, based on the higher than average amount of melanin naturally found there. 23 Becker introduced the concept of using non-specialised skin (i.e. skin that does not have an intrinsically high melanin content) for grafting the areola, in conjunction with intradermal tattooing. 24 This technique was then developed by Spear et al, 25 who have advocated the importance of high-quality nipple areolar reconstructions for over a decade. 6 Henseler et al have recently highlighted the difficulties of obtaining a good colour match when tattooing the nipple areola complex, and examined the normal nipple areolar colour in a population of Scottish women. 26 Initial colour match of the nipple areolar reconstruction with later fading has been reported as a problem when a pigmented skin graft is used. 2,3,15 Techniques using pigmented skin necessitate a donor wound, usually in the groin where the scar may be painful and visible and where infection is a risk. 10 Little s group initially used pigmented skin to reconstruct the areola in their local flaps, before changing to using an intradermal tattoo. 6 Reconstructions that involve tattooing are also not immune to fading. In Becker s initial paper on intradermal tattooing, it was stated that approximately 25% of patients needed to have touch-up tattoos or re-tattoos in order to improve the colour match. 24 Bhatty and Berry reported having to re-tattoo 10% of their patients. 5 Hugo et al stated that 40% of their patients receiving intradermal tattooing required re-tattooing. 17 None of these publications detail the criteria used for deciding whether to re-tattoo (i.e. whether it was performed at the patient s request or only if the surgeon deemed the result unsatisfactory). In the only study of the outcome of nipple areolar tattooing where patients were formally asked about the colour of their nipple areolar reconstruction, 60% reported that the tattoo was too light, although the majority were still satisfied with their outcome. 2 It may be argued that these problems with nipple areolar reconstruction mean that it should not be performed. However, our finding that the addition of the nipple areola complex has a positive impact on the appearance of the breast reconstruction as a whole indicates that nipple areolar reconstruction is worth carrying out, in spite of its shortcomings. This is supported by the evidence of Wellisch et al, who found that satisfaction with breast reconstruction was increased by the addition of a nipple areola complex. 1 When we asked about satisfaction with the breast reconstruction, we found patients ratings were high, both for the breast mound and for the final reconstruction. This concurs with the high satisfaction rates found in a study of tattooed patients by Spear and Arias. 2 However, satisfaction has been shown to be a nebulous concept, the measurement of which may be difficult to validate. 27 Certainly, positive responses to satisfaction questions are encouraging for surgeons, but positive results in such measures should not preclude the pursuit of improvement in a procedure. It was interesting to find that independent observers ratings were generally lower than those of the patients and their partners. Perhaps the experience of prolonged medical treatment and a mastectomy scar means that patients are more accepting of the results of breast reconstruction than are the general population. The subjective impression of fading of the nipple areola complex by patients, their partners and independent observers is mirrored by the objective results obtained using colour analysis. In clinical outcome studies, patient perception is all-important, but an objective measurement technique is useful for comparing surgical methods because it provides quantitative data, which are difficult to obtain from patient measures. The two techniques of assessment should, therefore, be seen as complementary rather than mutually exclusive. The patients responses showed that they are less happy with the nipple areolar colour than they are with the more general attributes of the breast reconstruction. It is this finding that nipple areolar reconstruction is the poor relation in breast reconstruction that leads us to conclude that it is important to strive to improve nipple areolar reconstructions. The use of in vitro cultured autologous melanocytes to pigment a disc of reconstructed breast skin to mimic the areola has been described in a single case. 28 Widespread use of the original procedure using naeval melanocytes in a mutagenic growth medium cannot be recommended because of the risk of malignant transformation. However, safe serum-free media are now available for melanocyte culture, and the redundant abdominal skin of a TRAM flap could theoretically be used as the donor for the melanocytes required. Co-culture of melanocytes and keratinocytes, and their recombination as a pigmented skin substitute with a melanocyte-to-keratinocyte ratio mimicking that of the nipple areolar skin, is feasible, although how such a construct would behave in the long term is unknown. Applications of this type of tissue engineering have, until now, been mainly limited to replacing lost skin in burns patients, 29 and the use of cultured melanocytes has been restricted to patients with hypopigmentary disorders, 30 but there is no reason why this technology could not be extended to the reconstitution of the nipple areola complex, if it can be established that the technique is safe. Tattooing, in the hands of some surgeons, can produce excellent results, 26 and it may be that the problems of colour mismatch and fading in tattooed patients will be obviated with improved nipple areolar tattoo pigments and innovations such as the use of Munsell colour charts. 26 The technique of colour analysis used in this study could be useful in the objective evaluation of tattoos and their performance over time. Regardless of whether tattoos become trouble-free in the near future, it may be worth exploring new methods of pigmentation to widen the repertoire of techniques available to breastreconstruction patients and their surgeons. In conclusion, we have shown that poor colour match and fading of nipple areolar reconstructions are measurable phenomena that are perceived by patients, their partners and independent observers. Despite these problems, nipple areolar reconstruction still contributes positively

8 Fading of nipple areolar reconstruction 581 to the appearance of breast reconstructions, and the investigation of new methods of reconstituting this specially pigmented skin is warranted. Acknowledgements The authors thank all participating patients, the Department of Clinical Photography at the Royal Adelaide Hospital and the four members of our panel of independent observers: Carol Brooker, Colleen Daly, Brendan Zeman and Carl Stephan. We also thank Heather McElroy and Professor Maciej Henneberg for assistance with the statistical analysis. Nicola Dean was funded by a Women s and Children s Hospital Medical Postgraduate Research Scholarship and received a stipend from the Cooperative Research Centre for Tissue Growth and Repair for this work. References 1. Wellisch DK, Schain WS, Noone RB, Little JW III. The psychological contribution of nipple addition in breast reconstruction. Plast Reconstr Surg 1987; 80: Spear SL, Arias J. Long-term experience with nipple areola tattooing. Ann Plast Surg 1995; 35: Bostwick J. Discussion of nipple reconstruction with the doubleopposing tab flap. Plast Reconstr Surg 1999; 104: Little JW III. Nipple areola reconstruction. Clin Plast Surg 1984; 11: Bhatty MA, Berry RB. Nipple areola reconstruction by tattooing and nipple sharing. Br J Plast Surg 1997; 50: Little JW III, Spear SL. The finishing touches in nipple areolar reconstruction. Perspect Plast Surg 1988; 2: Little JW III, Munasifi T, McCulloch DT. One-stage reconstruction of a projecting nipple: the quadrapod flap. Plast Reconstr Surg 1983; 71: Bosch G, Ramirez M. Reconstruction of the nipple: a new technique. Plast Reconstr Surg 1984; 73: Hartrampf CR Jr, Culbertson JH. A dermal-fat flap for nipple reconstruction. Plast Reconstr Surg 1984; 73: Cohen IK, Ward JA, Chandrasekhar B. The pinwheel flap nipple and barrier areola graft reconstruction. Plast Reconstr Surg 1986; 77: Smith JW, Nelson R. Construction of the nipple with a mushroomshaped pedicle. Plast Reconstr Surg 1986; 78: Weiss J, Herman O, Rosenberg L, Shafir R. The S nipple areola reconstruction. Plast Reconstr Surg 1989; 83: Vecchione TR. Reconstruction and/or salvage of nipple projection. Plast Reconstr Surg 1986; 78: Tanabe HY, Tai Y, Kiyokawa K, Yamauchi T. Nipple areola reconstruction with a dermal-fat flap and rolled auricular cartilage. Plast Reconstr Surg 1997; 100: Eskenazi L. A one-stage nipple reconstruction with the modified star flap and immediate tattoo: a review of 100 cases. Plast Reconstr Surg 1993; 92: Kroll SS. Nipple reconstruction with the double-opposing tab flap. Plast Reconstr Surg 1999; 104: Hugo NE, Sultan MR, Hardy SP. Nipple areola reconstruction with intradermal tattoo and double-opposing pennant flaps. Ann Plast Surg 1993; 30: Banducci DR, Le TK, Hughes KC. Long-term follow-up of a modified Anton-Hartrampf nipple reconstruction. Ann Plast Surg 1999; 43: Castles I. Women in Australia. 1993: Dean C, Chetty U, Forrest APM. Effects of immediate breast reconstruction on psychosocial morbidity after mastectomy. Lancet 1983; i: Davey RB, Sprod RT, Neild TO. Computerised colour: a technique for the assessment of burn scar hypertrophy. A preliminary report. Burns 1999; 25: Adams WM. Labial transplant for correction of loss of the nipple. Plast Reconstr Surg 1949; 4: Broadbent TR, Woolf RM, Metz PS. Restoring the mammary areola by a skin graft from the upper inner thigh. Br J Plast Surg 1977; 30: Becker H. The use of intradermal tattoo to enhance the final result of nipple areola reconstruction. Plast Reconstr Surg 1986; 77: Spear SL, Convit R, Little JW III. Intradermal tattoo as an adjunct to nipple areola reconstruction. Plast Reconstr Surg 1989; 83: Henseler H, Cheong V, Weiler-Mithoff EM, MacKay IR, Webster MHC. The use of Munsell colour charts in nipple areola tattooing. Br J Plast Surg 2001; 54: Meredith P. But was the operation worth it? The limitations of quality of life and patient satisfaction research in health-care outcome assessment. J Qual Clin Pract 1996; 16: Baltaci V, Kilic A. A new application for reconstruction of areola with transplantation of cultured autologous melanocytes. Plast Reconstr Surg 1998; 101: Boyce ST, Kagan RJ, Meyer NA, Yakuboff KP, Warden GD. The 1999 clinical research award. Cultured skin substitutes combined with Integra artificial skin to replace native skin autograft and allograft for the closure of excised full-thickness burns. J Burn Care Rehabil 1999; 20: Olsson MJ, Juhlin L. Repigmentation of vitiligo by transplantation of cultured autologous melanocytes. Acta Derm Venereol 1993; 73: The Authors Nicola R. Dean MB, ChB, FRCS, Research Registrar and PhD Student Rodney D. Cooter MD, FRACS, Associate Professor and Director of Plastic Surgery Department of Plastic and Reconstructive Surgery, Royal Adelaide Hospital, North Terrace, Adelaide, SA 5000, Australia. Tim Neild PhD, Lecturer in Human Physiology Department of Human Physiology, School of Medicine, Flinders University, GPO Box 2100, Adelaide, SA 5001, Australia. Julie Haynes PhD, Lecturer in Anatomical Sciences Department of Anatomical Sciences, University of Adelaide, North Terrace, Adelaide, SA 5000, Australia. Christopher Goddard PhD, Cell Biologist Co-operative Research Centre for Tissue Growth and Repair, GroPep Limited, PO Box BC, Adelaide, SA 5000, Australia. Correspondence to Dr Nicola Dean. Paper received 29 October Accepted 24 June 2002, after revision.