- Email: [email protected]
Topical negative pressure for treating chronic wounds: a systematic review
British Journal of Plastic Surgery (2001), 54, 238-242
9 2001 The British Association of Plastic Surgeons doi:10.1054/bjps.2001.3547
PLASTIC
SURGERY
REVIEW
Topical negative pressure for treating chronic wounds: a systematic review D. Evans and L. Land
Health and Social Care Research Centre, University of Central England, Birmingham, UK SUMMARY. Some wounds take a long time to heal, fail to heal or recur, causing significant pain and discomfort to the patient and cost to the National Health Service. This review assesses the effectiveness of topical negative pressure (TNP) in treating chronic wounds. The Cochrane Wounds Group Specialised Trials Register was searched for randomised controlled trials (RCTs) that evaluated the effectiveness of TNP on chronic-wound healing. Eligibility for inclusion, data extraction and details of trial quality were conducted by two reviewers independently. A narrative synthesis of results was undertaken as only two small trials, with different outcome measures, fulfilled the selection criteria. Trial 1 considered any type of chronic wound, trial 2 considered diabetic foot ulcers. Both trials compared TNP with saline-gauze dressings. Trial 1 reported a statistically significant difference in the percentage change in wound volume after 6 weeks, in favour of TNP. Trial 2 reported a difference in the number of days to healing and a difference in the percentage change in wound surface area after 2 weeks, in favour of TNE These two small trials provide weak evidence to suggest that TNP may be superior to saline-gauze dressings in terms of wound healing. However, due to the small sample sizes and the methodological limitations of the studies, these findings must be interpreted with extreme caution. The effects of TNP on cost, quality of life, pain and comfort were not reported. It was not possible to determine the optimum TNP regimen. Further high-quality RCTs that address these issues are required. 9 2001 The British Association of Plastic Surgeons
Keywords: topical negative pressure, chronic wound, healing, randomised controlled trial.
People are living longer, with more complex conditions, and are undergoing more extensive surgical procedures; therefore, the numbers with chronic wounds and wound complications continues to increase. 1 Despite the use of modem dressings, some of these wounds do not heal easily and cause much discomfort and pain. 2 Wound care in the UK costs in excess of s billion per year and therefore treatments that are both clinically effective and cost effective are required. 3 Since the biology of wound healing has become better understood many strategies have been developed to try to manipulate the wound-healing process. 3 One way of manipulating the wound environment is to apply a topical negative pressure (TNP) (measured in mmHg) across the wound surface. 1,4 To manipulate the wound in this way requires an open-cell dressing (e.g. foam, felt, surgical towel, gauze) to pack the wound, tubing connecting the dressing to a suction pump via a canister that collects any exudate, and an air-tight seal around the dressing. 5'6 All non-viable tissue needs to be removed beforehand. 1 TNP is generally contraindicated if the wound or surrounding tissues are cancerous, if there are fistulas to organs or
body cavities, if there is necrotic tissue or if there is untreated osteomyelitis.7 The concept of using negative pressure to create a suction force, enabling the drainage of surgical wounds in order to promote wound healing, is well documented.8,9 It has long been suggested that if excess fluid is not adequately removed from a wound following surgery, its components may serve as both physical and chemical deterrents to wound healing. 9 The basic concept of mechanical forces influencing the shape and growth of tissues is also well documented. 1~ TNP is reported to do both, i.e. remove excess interstitial fluid and transmit mechanical forces to the surrounding tissues with resultant deformation of the extracellular matrix and cells. 11 Both factors are thought to result in improved wound healing through a variety of mechanisms. 3 The transparent adhesive dressing usually employed to secure the dressing may also help to maintain a moist wound environment.3'7 Documented evidence of the effect of TNP on the wound environment has been obtained from several animal studies. Using an acute-wound swine model, Morykwas et al reported an increase in local functional blood perfusion, a significantly accelerated rate of granulation-tissue formation, a significant reduction in tissue bacterial levels and a significant increase in nutrient blood flow in TNP-treated wounds compared to wounds treated with saline-moistened gauze. 12,13 Fabian et al, using an ischaemic-wound model of the rabbit ear, reported a
A more detailed version of this review is available in electronic format on the Cochrane Library and will be continually updated with relevant RCTs establishing the effectiveness of TNP in treating chronic wounds. (The Cochrane Library (January 2001/1). Oxford: Update Software.)
238
Topical negative pressure for treating chronic wounds: a systematic review significant increase in the peak granulation tissue, a significantly smaller granulaton-fissue gap and a smaller epithelial gap in TNP-treated wounds compared to wounds treated with identical foam (but without TNP) and adhesive drape. 14 The use of TNP in patients with chronic wounds has been described by a number of clinicians. 1'15'16 The use of TNP for the in-home treatment of chronic wounds has been reviewed by referring to trials using non-concurrent/ historical control groups. 17 This review examines the impact of TNP on chronic wounds by referring to trials where the patients have been randomised to concurrent treatment and control groups.
Objectives The objectives of this study were to undertake a systematic review of all randomised controlled trials (RCTs) using TNP in the treatment of any patients with chronic wounds to determine: firstly, whether TNP was more effective than wound dressings in terms of improving healing rates, reducing cost, improving quality of life, minimising pain and providing comfort (summarised only when the original trials used validated measures); and secondly, whether there was an optimum TNP regimen in terms of dressing type, mode of delivery of negative pressure, degree, application (continuous and/or intermittent) and duration of negative pressure.
Methods The criteria for considering studies for this review were as follows.
Types of studies. Prospective RCTs that evaluated the effectiveness of TNP in treating chronic wounds. There was no restriction on the basis of language or publication status. Types of participants. Any patient with a chronic wound. For the purposes of this review, chronic wounds were as defined by Wysocki 18 and healing by secondary intention. Types of interventions. The primary intervention was TNP. All types of dressing, modes of delivery of negative pressure and negative pressure cycles that varied in degree, application and duration were considered. The intervention had to contrast with the control group. All types of control interventions (no treatment, sham, standard, other experimental) were included in the review. The strongest evidence was likely to be provided by trials in which the control dressing was the same as the one used to transmit TNP but which was not subjected to negative pressure. Types of outcome measures. The primary outcome measure of interest was the healing rate of the wound. As there is no consensus as to the most valid and reliable means of measuring healing rates of wounds, trials were included if they measured healing by some objective method, such as the time to complete healing or the rate
239
of change in wound area and/or volume expressed as an absolute or relative rate, or the proportion of wounds that were completely healed within the trial period. The secondary outcome measures of interest were cost, quality of life, pain and comfort.
Search strategyfor identification of studies The Cochrane Wounds Group Specialised Trials Register was searched for RCTs that evaluated the effects of TNP in the treatment of patients with chronic wounds. The Cochrane Wounds Group search strategy aimed to maximise the retrieval of completed or ongoing, published or unpublished studies with the appropriate design, wound type and intervention. For the intervention, the synonyms 'topical negative pressure', 'sub-atmospheric pressure therapy', 'sub-atmospheric pressure dressing', 'vacuum sealing', 'vacuum-assisted wound closure', 'vacuum-assisted closure (VAC or V.A.C.)', 'negative pressure dressing', 'foam suction dressing', 'vacuum compression', 'vacuum pack' and 'sealed surface wound suction' were used. Experts in the field and relevant companies were contacted to enquire about ongoing and recently completed relevant trials. In addition, citations within obtained papers were scrutinised to identify additional studies.
Methods of review References identified from searches were entered into a bibliographic software package (ProCite v. 4.0.3, Research Information System, USA). Titles and abstracts of studies identified by the Cochrane Wounds Group search strategy were assessed by one reviewer (DE), in terms of their relevance and design, according to the selection criteria. Translation was sought where necessary. Full versions of articles were requested if, from this initial assessment, they satisfied the selection criteria. Those titles and abstracts rejected were checked by another reviewer (LL). Full papers were checked to identify those studies that fitted the selection criteria (DE), again with translation being sought where necessary. This was repeated independently by another reviewer for verification (LL). Where uncertainty regarding the inclusion of abstracts or full papers remained, additional information was sought from the original authors. Any disagreement between the reviewers was resolved by discussion. Details of studies eligible for inclusion were summarised using a data-extraction sheet. The validities of the studies were assessed to detect potential sources of bias from the study design. This process was guided by a checklist to assess the validity of RCTs, outlined in the NHS Centre for Reviews and Dissemination Guidelines for Undertaking Systematic Rev,_'ews of Researcn on Effectiveness. 19 If any data were missing from trial reports, attempts were made to obtain that data by contacting the authors. Studies that had been published in duplicate were included only once. Data extraction and validity assessment were undertaken by one reviewer (DE) and repeated independently by another reviewer for verification (LL). Any disagreement was resolved by discussion.
240 Results
Description of studies No published full reports of RCTs were identified initially using the search strategy outlined above. Two relevant abstracts were identified and the authors contacted. One was subsequently published as a full paper2~ and the other is still ongoing. 21 Another relevant study was identified when an expert in the field made the reviewers aware of this study, which is awaiting publication. 22 Therefore, the two small trials eligible for inclusion in this review of the effect of TNP on the healing of chronic wounds were Joseph et al 2~ and McCallon et al. 22 Joseph et al assessed 24 patients with 36 chronic wounds (associated with pressure, dehiscence, trauma, venous stasis or radiation), defined as present for at least a month, in a randomised parallel-group study. 2~ Half (18) of the wounds received TNP (open-cell foam dressing with continuous suction (125 mmHg)) changed every 48h and the other half (18) received normal saline wet-to-moist gauze dressings (with an occlusive dressing used to secure the gauze) changed three times daily. It appears that where patients had multiple wounds, such wounds were treated during randomisation and data analysis as if they were independent from each other. At 3 and 6 weeks the percentage change in wound volume was measured by volume displacement of alginate impression moulds. Additional outcome measures were given by Joseph et al20 but they did not fulfil the selection criteria for this review. McCallon et al assessed 10 patients (age range: 18-75 years) with diabetic foot ulceration, defined as present for at least a month, in a randomised parallel-group pilot study. 22 Five patients received TNP (open-cell foam dressing with continuous suction (125mmHg) for the first 48h followed by TNP with intermittent suction (125mmHg)). Five patients received saline-moistened gauze dressings changed twice daily. The number of days taken to obtain satisfactory healing (calculated from the date of initial debridement to the date of definitive closure either by secondary intention or by delayed primary intention (suture closure, skin graft or muscle flap)) was recorded. In addition, at 2 weeks the percentage change in wound surface area was measured using sterile acetate, a pen and computer biometrics. Additional outcome measures were given by McCaUon et a122but they did not fulfil the selection criteria for this review. The two eligible trials both employed a commercially available device (VAC pump, Kinetic Concepts Inc, San Antonio, TX, USA) to apply the suction and McCallon et al22 stated that they used it in accordance with the manufacturer's general protocol for chronic wounds. Twelve studies reporting the effects of TNP on chronicwound heating, including cost, were excluded from the review, principally because they were not RCTs. In these studies either patients were not randomly allocated to the concurrent treatment and control groups, 23 or the control group was non-concurrent/historical, 24 ,25 or, as in the majority of studies, there was no control or comparison group at all. 1' 15' 16'26-28 Some of the studies were prospective RCTs on animals rather than humans. 12- 14 The remaining studies described TNP, proposed mechanisms of action or were case studies.
British Journal of Plastic Surgery Table 1 Quality assessment of randomised controlled trials evaluating the effect of topical negative pressure on chronicwound healing Joseph et al2~ patients in trial/arms inclusion/exclusion criteria stated a priori sample size calculation true randomisation allocation concealment appropriate baseline characteristics reported blinded outcome assessment main interventions well described adequate description of associated care appropriate outcome measures used withdrawals intention-to-treat analysis
McCaUon et a122
24 (36 wounds)/2 yes
10/2 yes
no yes a no b yes d
no no no c yes
yes yes
no yes
yes
yes
yes
yes
no no
N/A no
N/A: not appropriate (no withdrawals). aRandomised wounds rather than patients to treatments: inappropriate to randomise multiple wounds as if they were independent from each other when using a between-subjects design. bAllocation concealment unclear. CAllocation concealment absent. dReported for some variables, unclear how many patients in each arm of the study had multiple wounds.
The methodological quality of the included studies is summarised in Table 1.
Findings The findings of the review are presented with reference to the original objectives of the review. Synthesis of results using statistical pooling methods was not appropriate as only two small trials fulfilled the selection criteria and they used different outcome measures, so a narrative synthesis of the results is provided.
Was TNP more effective than wound dressings in terms of improving healing rates, reducing cost, improving quality of life, minimising pain and providing comfort? Joseph et al reported a significant difference in the percentage change in wound volume at 6 weeks in favour of TNP (78% versus 30%, P=0.038). 2~ It was not clear whether mean or median values were provided. No standard deviations or ranges were provided to accompany these figures. They stated that follow-up beyond the 6 week study period was done until complete wound closure was demonstrated for each patient, all were offered operative wound closure for any remaining open wounds. Unfortunately, these time-to-complete-healing data were not reported. Adverse outcomes were: 3/18 wounds treated with TNP had osteomyelitis and/or calcaneal fractures. Two of the patients suffered calcaneal fractures whilst ambulating on the TNP dressing (which Joseph et al state is against the manufacturer's recommendations and medical advice). 2~ Both patients eventually required amputation. In the control group, 8/18 wounds had osteomyelitis, other wound infections or fistulas (P = 0.0028).
Topical negative pressure for treating chronic wounds: a systematic review McCallon et al reported a difference in the number of days to satisfactory healing in favour of TNP (22.8 + 17.4 days versus 42.8 + 32.5 days) although no statistical analysis was performed. 22 Definitive closure was achieved by delayed primary closure in four out of five TNP patients and two out of five control patients. One TNP patient and three control patients ultimately achieved healing by secondary intention. In addition, McCallon et al reported a difference in the percentage change in wound surface area at 2 weeks in favour of TNP (28.4_+24.3% decrease versus 9.5__+16.9% increase) although no statistical analysis was performed. 22 Adverse outcomes were: pain was reported by some TNP patients with initial collapse and/or with removal of the foam dressing. Both of these events were reported to be brief ( < 30 s) and no pain was reported with subsequent TNP therapy. The authors did not state how many patients reported pain and it was not clear whether a validated measure was used to assess pain. There were no RCTs evaluating the effectiveness of TNP on cost, quality of life, pain or comfort.
Was there was an optimum TNP regimen in terms of dressing type, mode of delivery of negative pressure, degree, application (continuous and~or intermittent) or duration of negative pressure? There were no RCTs evaluating the effectiveness of different TNP regimens. Discussion
Potential biases in the primary studies and the limitations they place on inferences Both eligible studies 2~ were at risk of selection bias. Selection bias can be eliminated by assembling comparison groups in such a way that the process is impervious to any subconscious influence by the individuals making the allocation. 29 This is most securely achieved if an assignment schedule generated using true randomisation is concealed and can always be implemented. 29 In the study by Joseph et al the assignment schedule appeared to be generated using true randomisation but the adequacy of allocation concealment was unclear, 2~ whereas in the study by McCallon et al it was completely absent. 22 There is some empirical evidence to suggest that trials in which concealment is unclear or inadequate, yield larger estimates of treatment effects. 3~ Both eligible studies 2~ were also at risk of performance bias. To protect against unintentional differences in care and placebo effects, those providing and receiving care can be blinded so that they do not know to which group the recipients of care have been allocated. 29 In the studies by Joseph et al 2~ and McCallon et a122 the experimental groups received TNP delivered via a foam dressing whereas the control groups received saline-moistened gauze dressings, so it was impossible to 'blind' those providing and receiving care. In the study by Joseph et al 2~ the outcome assessors were blinded regarding treatment allocation, whereas in the study by McCallon et a122 this was not made explicit, so the latter study was possibly also at risk of detection bias. 29 In studies where patients, providers of care and outcome assessors are blinded, the
241
RCT is double-blinded, which means the Joseph et al20 study was single-blinded whereas the McCallon et a122 study lacked blinding. There is some empirical evidence to suggest that trials that are not double-blinded yield larger estimates of treatment effects. 3~ It is uncertain whether or not both eligible studies 2~ were at risk of attrition bias. In the study by Joseph et al 2~ it was not explicitly stated whether intention-to-treat analysis was used or whether there were any withdrawals. In the study by McCallon et al22 it was also not explicitly stated whether intention-to-treat analysis was used but there did not appear to be any withdrawals. The approach to handling losses has a great potential to bias the results, however, empirical evidence to support this is unclear. 30 Another potential problem in the trials, 2~ which limits the inferences that can be drawn, was sampling error. Both studies were very small, which increases the probability that the samples were not completely representative of the populations from which they were drawn. Overall, the two small trials 2~ provide weak evidence to suggest that TNP may be superior to saline-gauze dressings in terms of healing chronic human wounds. However, the methodological limitations of these studies mean that the findings must be interpreted with extreme caution.
Potential biases in the review and the limitations they place on inferences The different databases searched have different commencement dates, so the search of databases was restricted to these dates. The reviewers relied on the good will of experts in the field to provide information on completed or ongoing, published or unpublished studies. When critically appraising the validity of the studies, the reviewers had to rely on adequate reporting of the trials, assuming that if something was not reported then it was not done, which is not necessarily correct. The reviewers did attempt to obtain additional clarifying data from the investigators. There is limited empirical evidence of a relationship between parameters thought to measure validity and actual trial outcomes. More research is needed to establish which criteria for assessing validity are important determinants of study results. 29
Conclusions
Implications for practice Practitioners ought to make patients aware of the limited trial-based evidence for the effectiveness of TNP on the healing of chronic human wounds and that further research is required to validate the preliminary findings.
Implications for research There is a need for well-designed, adequately powered, multi-centre RCTs to evaluate the contribution of TNP to
242 the healing o f chronic wounds. T h e strongest e v i d e n c e for w h e t h e r T N P w o r k s at all, in humans, is likely to be p r o v i d e d by trials in w h i c h the control dressing is the s a m e as the one used to transmit T N P but is not subjected to negative pressure. E v i d e n c e for the relative benefits o f TNP, compared to other dressings, is likely to be provided by trials in which the comparison dressing is one routinely u s e d for chronic w o u n d s based on the best available evidence. T h e r e is currently a relatively large o n g o i n g multicentre R C T c o m p a r i n g f o a m to transmit T N P with s a l i n e - m o i s t e n e d g a u z e dressings on healing grade 3 and grade 4 pressure ulcers. 21 This m a y establish the effectiveness o f T N P on healing grade 3 and grade 4 pressure ulcers c o m p a r e d to s a l i n e - m o i s t e n e d g a u z e dressings, but will not p r o v i d e information on the relative effectiveness o f T N P c o m p a r e d to m o d e m w o u n d dressings. T h e r e is still a n e e d for R C T s evaluating the relative effectiveness o f T N P c o m p a r e d to m o d e m w o u n d dressings for healing all types o f chronic h u m a n wounds, and the effectiveness o f T N P on cost, quality o f life, pain and comfort, and w h e t h e r there is an o p t i m u m T N P r e g i m e n .
Acknowledgements We thank the University of Central England for intramural support and the Cochrane Wounds Group.
References 1. Argenta LC, Morykwas MJ. Vacuum-assisted closure: a new method for wound control and treatment: clinical experience. Ann Plast Surg 1997; 38: 563-76. 2. Dealey C. The Care of Wounds, 2nd ed. Oxford: Blackwell Scientific Publications, 1999. 3. BanweU PE. Topical negative pressure therapy in wound care: a review of the development and use of sub-atmospheric pressures in the management of patients with different types of wound. J Wound Care 1999; 8: 79-84. 4. Fleischmann W, Becker U, Bischoff M, Hoekstra H. Vacuum sealing: indication, technique, and results. Eur J Orthop Surg Traumatol 1995; 5: 37--40. 5. Baxandall T. Healing cavity wounds with negative pressure. Elderly Care 1997; 9: 20-2. 6. Greer SE. Whither subatmospheric pressure dressing. Ann Plast Surg 2000; 45: 332-4. 7. Mendez-Eastman S. When wounds won't heal. RN 1998; 61: 20-4. 8. Fox JW IV, Golden GT. The use of drains in subcutaneous surgical procedures. Am J Surg 1976; 132: 673-4. 9. Fay ME Drainage systems: their role in wound healing. AORN J 1987; 46: 442-55. 10. Ovington LG. 1, 2, 3 s-t-r-e-t-c-h; vacuum enhances wound closure. In Advances in Wound Care, the 1999 edition source book, 1999: 125-7. 11. Morykwas MJ. External application of sub-atmospheric pressure and healing: mechanisms of action. Wound Healing Soc Newsletter 1998; 8: 4-5. 12. Morykwas MJ, Argenta LC. Use of negative pressure to increase the rate of granulation tissue formation in chronic open wounds. Presented at the Annual Meeting: Federation of American Societies for Experimental Biology, New Orleans, LA, USA, Mar 28-Apt 1, 1993. 13. Morykwas MJ, Argenta LC, Shelton-Brown EI, McGuirt W. Vacuum-assisted closure: a new method for wound control and treatment: animal studies and basic foundation. Ann Hast Surg 1997; 38: 553-62. 14. Fabian TS, Kaufman HJ, Lett ED, et al. The evaluation of subatmospheric pressure and hyperbaric oxygen in ischemic fullthickness wound healing. Am Surg 2000; 66:1136-43.
British Joumal of Plastic Surgery 15. Milliner T, Mrkonjic L, Kwasny O, Vtcsei V. The use of negative pressure to promote the healing of tissue defects: a clinical trial using the vacuum sealing technique. Br J Plast Surg 1997; 50: 194-9. 16. Banwell PE, Holten IW, Martin DL. Negative pressure therapy: clinical applications and experience with 200 cases. Wound Repair Regen 1998; A460. 17. Anonymous. Technology assessment of the V.A.C. for in-home treatment of chronic wounds. Washington: The Weinberg Group Inc., 1999. 18. Wysocki AB. Wound fluids and the pathogenesis of chronic wounds. J Wound Ostomy Continence Nursing 1996; 23: 283-90. 19. NHS Centre for Reviews and Dissemination. Undertaking systematic reviews of research on effectiveness. CRD Guidelines for Those Carrying Out or Commissioning Reviews. CRD Report Number 4, January 1996. 20. Joseph E, Hamori CA, Bergman S, Roaf E, Swann NF, Anastasi GW. A prospective randomized trial of vacuum assisted closure versus standard therapy of chronic nonhealing wounds. Wounds 2000; 12: 60-7. 21. Greer SE, Longaker MT, Margiotta M. Preliminary results from a multicenter, randomised, controlled, study of the use of subatmospheric pressure dressing for pressure ulcer healing. Wound Repair Regen 1999; 7: A255. 22. McCallon SK, Knight CA, Valiulus JP, Cunningham MW, McCulloch JM, Farinas LP. The effectiveness of vacuum assisted closure vs. saline moistened gauze in the healing of post-operative diabetic foot wounds. Adv Wound Care 2001 (Submitted). 23. Morykwas M, Schneider A, Argenta L, Fowler J, McHone J. Effect of VACT M therapy on total charge and length of stay of patients in DRG 263: a 21 month analysis. Presented at the 27th Annual Conference, Wound, Ostomy and Continence Nurses, Denver, CO, USA, May 13-18, 1995. 24. Deva AK, Buckland GH, Fisher E, et al. Topical negative pressure in wound management. Med J Aust 2000; 173: 128-31. 25. Philbeck TE Jr, Whittington KT, Millsap MH, Briones RB, Wight DG, Schroeder WJ. The clinical and cost effectiveness of externally applied negative pressure wound therapy in the treatment of wounds in home healthcare medicare patients. Ostomy Wound Manag 1999; 45: 41-50. 26. Das Gupta R, Twyman L, Morgan B, McGrouther DA. An introduction to VAC therapy in the treatment of chronic wounds. European Tissue Repair Society, vol. 3, 1996. 27. HSlmich LR, Slim E, Krag C, Dahlfeldt H, Norden A. Vacuumassisted wound closure clinical experience. Wound Repair Regen 1998; A470. 28. Ladin D, Hou Z, Philbeck TE Jr. Vacuum assisted closure in the management of lower-extremity wounds. Adv Wound Care 2001 (Submitted). 29. Mulrow CD, Oxman AD, ed. The Cochrane collaboration handbook (updated September 1997). In The Cochrane Library (database on disk and CD ROM). The Cochrane Collaboration. Oxford: Update Software, 1997: Issue 4. 30. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995; 273: 408-12.
The Authors Debra Evans RGN, BSc, PhD(Physiology), Lecturer in Research Methodology
Lucy Land RGN, DPSN, BSc, PGCE, Senior Lecturer in Research Methodology Health and Social Care Research Centre, University of Central England (UCE), Ravensbury House, Westbourne Road, Birmingham B15 3TN, UK. Correspondence to Dr Debra Evans. Paper received 15 September 2000. Accepted 29 November 2000.
9 2001 The British Association of Plastic Surgeons doi:10.1054/bjps.2001.3547
PLASTIC
SURGERY
REVIEW
Topical negative pressure for treating chronic wounds: a systematic review D. Evans and L. Land
Health and Social Care Research Centre, University of Central England, Birmingham, UK SUMMARY. Some wounds take a long time to heal, fail to heal or recur, causing significant pain and discomfort to the patient and cost to the National Health Service. This review assesses the effectiveness of topical negative pressure (TNP) in treating chronic wounds. The Cochrane Wounds Group Specialised Trials Register was searched for randomised controlled trials (RCTs) that evaluated the effectiveness of TNP on chronic-wound healing. Eligibility for inclusion, data extraction and details of trial quality were conducted by two reviewers independently. A narrative synthesis of results was undertaken as only two small trials, with different outcome measures, fulfilled the selection criteria. Trial 1 considered any type of chronic wound, trial 2 considered diabetic foot ulcers. Both trials compared TNP with saline-gauze dressings. Trial 1 reported a statistically significant difference in the percentage change in wound volume after 6 weeks, in favour of TNP. Trial 2 reported a difference in the number of days to healing and a difference in the percentage change in wound surface area after 2 weeks, in favour of TNE These two small trials provide weak evidence to suggest that TNP may be superior to saline-gauze dressings in terms of wound healing. However, due to the small sample sizes and the methodological limitations of the studies, these findings must be interpreted with extreme caution. The effects of TNP on cost, quality of life, pain and comfort were not reported. It was not possible to determine the optimum TNP regimen. Further high-quality RCTs that address these issues are required. 9 2001 The British Association of Plastic Surgeons
Keywords: topical negative pressure, chronic wound, healing, randomised controlled trial.
People are living longer, with more complex conditions, and are undergoing more extensive surgical procedures; therefore, the numbers with chronic wounds and wound complications continues to increase. 1 Despite the use of modem dressings, some of these wounds do not heal easily and cause much discomfort and pain. 2 Wound care in the UK costs in excess of s billion per year and therefore treatments that are both clinically effective and cost effective are required. 3 Since the biology of wound healing has become better understood many strategies have been developed to try to manipulate the wound-healing process. 3 One way of manipulating the wound environment is to apply a topical negative pressure (TNP) (measured in mmHg) across the wound surface. 1,4 To manipulate the wound in this way requires an open-cell dressing (e.g. foam, felt, surgical towel, gauze) to pack the wound, tubing connecting the dressing to a suction pump via a canister that collects any exudate, and an air-tight seal around the dressing. 5'6 All non-viable tissue needs to be removed beforehand. 1 TNP is generally contraindicated if the wound or surrounding tissues are cancerous, if there are fistulas to organs or
body cavities, if there is necrotic tissue or if there is untreated osteomyelitis.7 The concept of using negative pressure to create a suction force, enabling the drainage of surgical wounds in order to promote wound healing, is well documented.8,9 It has long been suggested that if excess fluid is not adequately removed from a wound following surgery, its components may serve as both physical and chemical deterrents to wound healing. 9 The basic concept of mechanical forces influencing the shape and growth of tissues is also well documented. 1~ TNP is reported to do both, i.e. remove excess interstitial fluid and transmit mechanical forces to the surrounding tissues with resultant deformation of the extracellular matrix and cells. 11 Both factors are thought to result in improved wound healing through a variety of mechanisms. 3 The transparent adhesive dressing usually employed to secure the dressing may also help to maintain a moist wound environment.3'7 Documented evidence of the effect of TNP on the wound environment has been obtained from several animal studies. Using an acute-wound swine model, Morykwas et al reported an increase in local functional blood perfusion, a significantly accelerated rate of granulation-tissue formation, a significant reduction in tissue bacterial levels and a significant increase in nutrient blood flow in TNP-treated wounds compared to wounds treated with saline-moistened gauze. 12,13 Fabian et al, using an ischaemic-wound model of the rabbit ear, reported a
A more detailed version of this review is available in electronic format on the Cochrane Library and will be continually updated with relevant RCTs establishing the effectiveness of TNP in treating chronic wounds. (The Cochrane Library (January 2001/1). Oxford: Update Software.)
238
Topical negative pressure for treating chronic wounds: a systematic review significant increase in the peak granulation tissue, a significantly smaller granulaton-fissue gap and a smaller epithelial gap in TNP-treated wounds compared to wounds treated with identical foam (but without TNP) and adhesive drape. 14 The use of TNP in patients with chronic wounds has been described by a number of clinicians. 1'15'16 The use of TNP for the in-home treatment of chronic wounds has been reviewed by referring to trials using non-concurrent/ historical control groups. 17 This review examines the impact of TNP on chronic wounds by referring to trials where the patients have been randomised to concurrent treatment and control groups.
Objectives The objectives of this study were to undertake a systematic review of all randomised controlled trials (RCTs) using TNP in the treatment of any patients with chronic wounds to determine: firstly, whether TNP was more effective than wound dressings in terms of improving healing rates, reducing cost, improving quality of life, minimising pain and providing comfort (summarised only when the original trials used validated measures); and secondly, whether there was an optimum TNP regimen in terms of dressing type, mode of delivery of negative pressure, degree, application (continuous and/or intermittent) and duration of negative pressure.
Methods The criteria for considering studies for this review were as follows.
Types of studies. Prospective RCTs that evaluated the effectiveness of TNP in treating chronic wounds. There was no restriction on the basis of language or publication status. Types of participants. Any patient with a chronic wound. For the purposes of this review, chronic wounds were as defined by Wysocki 18 and healing by secondary intention. Types of interventions. The primary intervention was TNP. All types of dressing, modes of delivery of negative pressure and negative pressure cycles that varied in degree, application and duration were considered. The intervention had to contrast with the control group. All types of control interventions (no treatment, sham, standard, other experimental) were included in the review. The strongest evidence was likely to be provided by trials in which the control dressing was the same as the one used to transmit TNP but which was not subjected to negative pressure. Types of outcome measures. The primary outcome measure of interest was the healing rate of the wound. As there is no consensus as to the most valid and reliable means of measuring healing rates of wounds, trials were included if they measured healing by some objective method, such as the time to complete healing or the rate
239
of change in wound area and/or volume expressed as an absolute or relative rate, or the proportion of wounds that were completely healed within the trial period. The secondary outcome measures of interest were cost, quality of life, pain and comfort.
Search strategyfor identification of studies The Cochrane Wounds Group Specialised Trials Register was searched for RCTs that evaluated the effects of TNP in the treatment of patients with chronic wounds. The Cochrane Wounds Group search strategy aimed to maximise the retrieval of completed or ongoing, published or unpublished studies with the appropriate design, wound type and intervention. For the intervention, the synonyms 'topical negative pressure', 'sub-atmospheric pressure therapy', 'sub-atmospheric pressure dressing', 'vacuum sealing', 'vacuum-assisted wound closure', 'vacuum-assisted closure (VAC or V.A.C.)', 'negative pressure dressing', 'foam suction dressing', 'vacuum compression', 'vacuum pack' and 'sealed surface wound suction' were used. Experts in the field and relevant companies were contacted to enquire about ongoing and recently completed relevant trials. In addition, citations within obtained papers were scrutinised to identify additional studies.
Methods of review References identified from searches were entered into a bibliographic software package (ProCite v. 4.0.3, Research Information System, USA). Titles and abstracts of studies identified by the Cochrane Wounds Group search strategy were assessed by one reviewer (DE), in terms of their relevance and design, according to the selection criteria. Translation was sought where necessary. Full versions of articles were requested if, from this initial assessment, they satisfied the selection criteria. Those titles and abstracts rejected were checked by another reviewer (LL). Full papers were checked to identify those studies that fitted the selection criteria (DE), again with translation being sought where necessary. This was repeated independently by another reviewer for verification (LL). Where uncertainty regarding the inclusion of abstracts or full papers remained, additional information was sought from the original authors. Any disagreement between the reviewers was resolved by discussion. Details of studies eligible for inclusion were summarised using a data-extraction sheet. The validities of the studies were assessed to detect potential sources of bias from the study design. This process was guided by a checklist to assess the validity of RCTs, outlined in the NHS Centre for Reviews and Dissemination Guidelines for Undertaking Systematic Rev,_'ews of Researcn on Effectiveness. 19 If any data were missing from trial reports, attempts were made to obtain that data by contacting the authors. Studies that had been published in duplicate were included only once. Data extraction and validity assessment were undertaken by one reviewer (DE) and repeated independently by another reviewer for verification (LL). Any disagreement was resolved by discussion.
240 Results
Description of studies No published full reports of RCTs were identified initially using the search strategy outlined above. Two relevant abstracts were identified and the authors contacted. One was subsequently published as a full paper2~ and the other is still ongoing. 21 Another relevant study was identified when an expert in the field made the reviewers aware of this study, which is awaiting publication. 22 Therefore, the two small trials eligible for inclusion in this review of the effect of TNP on the healing of chronic wounds were Joseph et al 2~ and McCallon et al. 22 Joseph et al assessed 24 patients with 36 chronic wounds (associated with pressure, dehiscence, trauma, venous stasis or radiation), defined as present for at least a month, in a randomised parallel-group study. 2~ Half (18) of the wounds received TNP (open-cell foam dressing with continuous suction (125 mmHg)) changed every 48h and the other half (18) received normal saline wet-to-moist gauze dressings (with an occlusive dressing used to secure the gauze) changed three times daily. It appears that where patients had multiple wounds, such wounds were treated during randomisation and data analysis as if they were independent from each other. At 3 and 6 weeks the percentage change in wound volume was measured by volume displacement of alginate impression moulds. Additional outcome measures were given by Joseph et al20 but they did not fulfil the selection criteria for this review. McCallon et al assessed 10 patients (age range: 18-75 years) with diabetic foot ulceration, defined as present for at least a month, in a randomised parallel-group pilot study. 22 Five patients received TNP (open-cell foam dressing with continuous suction (125mmHg) for the first 48h followed by TNP with intermittent suction (125mmHg)). Five patients received saline-moistened gauze dressings changed twice daily. The number of days taken to obtain satisfactory healing (calculated from the date of initial debridement to the date of definitive closure either by secondary intention or by delayed primary intention (suture closure, skin graft or muscle flap)) was recorded. In addition, at 2 weeks the percentage change in wound surface area was measured using sterile acetate, a pen and computer biometrics. Additional outcome measures were given by McCaUon et a122but they did not fulfil the selection criteria for this review. The two eligible trials both employed a commercially available device (VAC pump, Kinetic Concepts Inc, San Antonio, TX, USA) to apply the suction and McCallon et al22 stated that they used it in accordance with the manufacturer's general protocol for chronic wounds. Twelve studies reporting the effects of TNP on chronicwound heating, including cost, were excluded from the review, principally because they were not RCTs. In these studies either patients were not randomly allocated to the concurrent treatment and control groups, 23 or the control group was non-concurrent/historical, 24 ,25 or, as in the majority of studies, there was no control or comparison group at all. 1' 15' 16'26-28 Some of the studies were prospective RCTs on animals rather than humans. 12- 14 The remaining studies described TNP, proposed mechanisms of action or were case studies.
British Journal of Plastic Surgery Table 1 Quality assessment of randomised controlled trials evaluating the effect of topical negative pressure on chronicwound healing Joseph et al2~ patients in trial/arms inclusion/exclusion criteria stated a priori sample size calculation true randomisation allocation concealment appropriate baseline characteristics reported blinded outcome assessment main interventions well described adequate description of associated care appropriate outcome measures used withdrawals intention-to-treat analysis
McCaUon et a122
24 (36 wounds)/2 yes
10/2 yes
no yes a no b yes d
no no no c yes
yes yes
no yes
yes
yes
yes
yes
no no
N/A no
N/A: not appropriate (no withdrawals). aRandomised wounds rather than patients to treatments: inappropriate to randomise multiple wounds as if they were independent from each other when using a between-subjects design. bAllocation concealment unclear. CAllocation concealment absent. dReported for some variables, unclear how many patients in each arm of the study had multiple wounds.
The methodological quality of the included studies is summarised in Table 1.
Findings The findings of the review are presented with reference to the original objectives of the review. Synthesis of results using statistical pooling methods was not appropriate as only two small trials fulfilled the selection criteria and they used different outcome measures, so a narrative synthesis of the results is provided.
Was TNP more effective than wound dressings in terms of improving healing rates, reducing cost, improving quality of life, minimising pain and providing comfort? Joseph et al reported a significant difference in the percentage change in wound volume at 6 weeks in favour of TNP (78% versus 30%, P=0.038). 2~ It was not clear whether mean or median values were provided. No standard deviations or ranges were provided to accompany these figures. They stated that follow-up beyond the 6 week study period was done until complete wound closure was demonstrated for each patient, all were offered operative wound closure for any remaining open wounds. Unfortunately, these time-to-complete-healing data were not reported. Adverse outcomes were: 3/18 wounds treated with TNP had osteomyelitis and/or calcaneal fractures. Two of the patients suffered calcaneal fractures whilst ambulating on the TNP dressing (which Joseph et al state is against the manufacturer's recommendations and medical advice). 2~ Both patients eventually required amputation. In the control group, 8/18 wounds had osteomyelitis, other wound infections or fistulas (P = 0.0028).
Topical negative pressure for treating chronic wounds: a systematic review McCallon et al reported a difference in the number of days to satisfactory healing in favour of TNP (22.8 + 17.4 days versus 42.8 + 32.5 days) although no statistical analysis was performed. 22 Definitive closure was achieved by delayed primary closure in four out of five TNP patients and two out of five control patients. One TNP patient and three control patients ultimately achieved healing by secondary intention. In addition, McCallon et al reported a difference in the percentage change in wound surface area at 2 weeks in favour of TNP (28.4_+24.3% decrease versus 9.5__+16.9% increase) although no statistical analysis was performed. 22 Adverse outcomes were: pain was reported by some TNP patients with initial collapse and/or with removal of the foam dressing. Both of these events were reported to be brief ( < 30 s) and no pain was reported with subsequent TNP therapy. The authors did not state how many patients reported pain and it was not clear whether a validated measure was used to assess pain. There were no RCTs evaluating the effectiveness of TNP on cost, quality of life, pain or comfort.
Was there was an optimum TNP regimen in terms of dressing type, mode of delivery of negative pressure, degree, application (continuous and~or intermittent) or duration of negative pressure? There were no RCTs evaluating the effectiveness of different TNP regimens. Discussion
Potential biases in the primary studies and the limitations they place on inferences Both eligible studies 2~ were at risk of selection bias. Selection bias can be eliminated by assembling comparison groups in such a way that the process is impervious to any subconscious influence by the individuals making the allocation. 29 This is most securely achieved if an assignment schedule generated using true randomisation is concealed and can always be implemented. 29 In the study by Joseph et al the assignment schedule appeared to be generated using true randomisation but the adequacy of allocation concealment was unclear, 2~ whereas in the study by McCallon et al it was completely absent. 22 There is some empirical evidence to suggest that trials in which concealment is unclear or inadequate, yield larger estimates of treatment effects. 3~ Both eligible studies 2~ were also at risk of performance bias. To protect against unintentional differences in care and placebo effects, those providing and receiving care can be blinded so that they do not know to which group the recipients of care have been allocated. 29 In the studies by Joseph et al 2~ and McCallon et a122 the experimental groups received TNP delivered via a foam dressing whereas the control groups received saline-moistened gauze dressings, so it was impossible to 'blind' those providing and receiving care. In the study by Joseph et al 2~ the outcome assessors were blinded regarding treatment allocation, whereas in the study by McCallon et a122 this was not made explicit, so the latter study was possibly also at risk of detection bias. 29 In studies where patients, providers of care and outcome assessors are blinded, the
241
RCT is double-blinded, which means the Joseph et al20 study was single-blinded whereas the McCallon et a122 study lacked blinding. There is some empirical evidence to suggest that trials that are not double-blinded yield larger estimates of treatment effects. 3~ It is uncertain whether or not both eligible studies 2~ were at risk of attrition bias. In the study by Joseph et al 2~ it was not explicitly stated whether intention-to-treat analysis was used or whether there were any withdrawals. In the study by McCallon et al22 it was also not explicitly stated whether intention-to-treat analysis was used but there did not appear to be any withdrawals. The approach to handling losses has a great potential to bias the results, however, empirical evidence to support this is unclear. 30 Another potential problem in the trials, 2~ which limits the inferences that can be drawn, was sampling error. Both studies were very small, which increases the probability that the samples were not completely representative of the populations from which they were drawn. Overall, the two small trials 2~ provide weak evidence to suggest that TNP may be superior to saline-gauze dressings in terms of healing chronic human wounds. However, the methodological limitations of these studies mean that the findings must be interpreted with extreme caution.
Potential biases in the review and the limitations they place on inferences The different databases searched have different commencement dates, so the search of databases was restricted to these dates. The reviewers relied on the good will of experts in the field to provide information on completed or ongoing, published or unpublished studies. When critically appraising the validity of the studies, the reviewers had to rely on adequate reporting of the trials, assuming that if something was not reported then it was not done, which is not necessarily correct. The reviewers did attempt to obtain additional clarifying data from the investigators. There is limited empirical evidence of a relationship between parameters thought to measure validity and actual trial outcomes. More research is needed to establish which criteria for assessing validity are important determinants of study results. 29
Conclusions
Implications for practice Practitioners ought to make patients aware of the limited trial-based evidence for the effectiveness of TNP on the healing of chronic human wounds and that further research is required to validate the preliminary findings.
Implications for research There is a need for well-designed, adequately powered, multi-centre RCTs to evaluate the contribution of TNP to
242 the healing o f chronic wounds. T h e strongest e v i d e n c e for w h e t h e r T N P w o r k s at all, in humans, is likely to be p r o v i d e d by trials in w h i c h the control dressing is the s a m e as the one used to transmit T N P but is not subjected to negative pressure. E v i d e n c e for the relative benefits o f TNP, compared to other dressings, is likely to be provided by trials in which the comparison dressing is one routinely u s e d for chronic w o u n d s based on the best available evidence. T h e r e is currently a relatively large o n g o i n g multicentre R C T c o m p a r i n g f o a m to transmit T N P with s a l i n e - m o i s t e n e d g a u z e dressings on healing grade 3 and grade 4 pressure ulcers. 21 This m a y establish the effectiveness o f T N P on healing grade 3 and grade 4 pressure ulcers c o m p a r e d to s a l i n e - m o i s t e n e d g a u z e dressings, but will not p r o v i d e information on the relative effectiveness o f T N P c o m p a r e d to m o d e m w o u n d dressings. T h e r e is still a n e e d for R C T s evaluating the relative effectiveness o f T N P c o m p a r e d to m o d e m w o u n d dressings for healing all types o f chronic h u m a n wounds, and the effectiveness o f T N P on cost, quality o f life, pain and comfort, and w h e t h e r there is an o p t i m u m T N P r e g i m e n .
Acknowledgements We thank the University of Central England for intramural support and the Cochrane Wounds Group.
References 1. Argenta LC, Morykwas MJ. Vacuum-assisted closure: a new method for wound control and treatment: clinical experience. Ann Plast Surg 1997; 38: 563-76. 2. Dealey C. The Care of Wounds, 2nd ed. Oxford: Blackwell Scientific Publications, 1999. 3. BanweU PE. Topical negative pressure therapy in wound care: a review of the development and use of sub-atmospheric pressures in the management of patients with different types of wound. J Wound Care 1999; 8: 79-84. 4. Fleischmann W, Becker U, Bischoff M, Hoekstra H. Vacuum sealing: indication, technique, and results. Eur J Orthop Surg Traumatol 1995; 5: 37--40. 5. Baxandall T. Healing cavity wounds with negative pressure. Elderly Care 1997; 9: 20-2. 6. Greer SE. Whither subatmospheric pressure dressing. Ann Plast Surg 2000; 45: 332-4. 7. Mendez-Eastman S. When wounds won't heal. RN 1998; 61: 20-4. 8. Fox JW IV, Golden GT. The use of drains in subcutaneous surgical procedures. Am J Surg 1976; 132: 673-4. 9. Fay ME Drainage systems: their role in wound healing. AORN J 1987; 46: 442-55. 10. Ovington LG. 1, 2, 3 s-t-r-e-t-c-h; vacuum enhances wound closure. In Advances in Wound Care, the 1999 edition source book, 1999: 125-7. 11. Morykwas MJ. External application of sub-atmospheric pressure and healing: mechanisms of action. Wound Healing Soc Newsletter 1998; 8: 4-5. 12. Morykwas MJ, Argenta LC. Use of negative pressure to increase the rate of granulation tissue formation in chronic open wounds. Presented at the Annual Meeting: Federation of American Societies for Experimental Biology, New Orleans, LA, USA, Mar 28-Apt 1, 1993. 13. Morykwas MJ, Argenta LC, Shelton-Brown EI, McGuirt W. Vacuum-assisted closure: a new method for wound control and treatment: animal studies and basic foundation. Ann Hast Surg 1997; 38: 553-62. 14. Fabian TS, Kaufman HJ, Lett ED, et al. The evaluation of subatmospheric pressure and hyperbaric oxygen in ischemic fullthickness wound healing. Am Surg 2000; 66:1136-43.
British Joumal of Plastic Surgery 15. Milliner T, Mrkonjic L, Kwasny O, Vtcsei V. The use of negative pressure to promote the healing of tissue defects: a clinical trial using the vacuum sealing technique. Br J Plast Surg 1997; 50: 194-9. 16. Banwell PE, Holten IW, Martin DL. Negative pressure therapy: clinical applications and experience with 200 cases. Wound Repair Regen 1998; A460. 17. Anonymous. Technology assessment of the V.A.C. for in-home treatment of chronic wounds. Washington: The Weinberg Group Inc., 1999. 18. Wysocki AB. Wound fluids and the pathogenesis of chronic wounds. J Wound Ostomy Continence Nursing 1996; 23: 283-90. 19. NHS Centre for Reviews and Dissemination. Undertaking systematic reviews of research on effectiveness. CRD Guidelines for Those Carrying Out or Commissioning Reviews. CRD Report Number 4, January 1996. 20. Joseph E, Hamori CA, Bergman S, Roaf E, Swann NF, Anastasi GW. A prospective randomized trial of vacuum assisted closure versus standard therapy of chronic nonhealing wounds. Wounds 2000; 12: 60-7. 21. Greer SE, Longaker MT, Margiotta M. Preliminary results from a multicenter, randomised, controlled, study of the use of subatmospheric pressure dressing for pressure ulcer healing. Wound Repair Regen 1999; 7: A255. 22. McCallon SK, Knight CA, Valiulus JP, Cunningham MW, McCulloch JM, Farinas LP. The effectiveness of vacuum assisted closure vs. saline moistened gauze in the healing of post-operative diabetic foot wounds. Adv Wound Care 2001 (Submitted). 23. Morykwas M, Schneider A, Argenta L, Fowler J, McHone J. Effect of VACT M therapy on total charge and length of stay of patients in DRG 263: a 21 month analysis. Presented at the 27th Annual Conference, Wound, Ostomy and Continence Nurses, Denver, CO, USA, May 13-18, 1995. 24. Deva AK, Buckland GH, Fisher E, et al. Topical negative pressure in wound management. Med J Aust 2000; 173: 128-31. 25. Philbeck TE Jr, Whittington KT, Millsap MH, Briones RB, Wight DG, Schroeder WJ. The clinical and cost effectiveness of externally applied negative pressure wound therapy in the treatment of wounds in home healthcare medicare patients. Ostomy Wound Manag 1999; 45: 41-50. 26. Das Gupta R, Twyman L, Morgan B, McGrouther DA. An introduction to VAC therapy in the treatment of chronic wounds. European Tissue Repair Society, vol. 3, 1996. 27. HSlmich LR, Slim E, Krag C, Dahlfeldt H, Norden A. Vacuumassisted wound closure clinical experience. Wound Repair Regen 1998; A470. 28. Ladin D, Hou Z, Philbeck TE Jr. Vacuum assisted closure in the management of lower-extremity wounds. Adv Wound Care 2001 (Submitted). 29. Mulrow CD, Oxman AD, ed. The Cochrane collaboration handbook (updated September 1997). In The Cochrane Library (database on disk and CD ROM). The Cochrane Collaboration. Oxford: Update Software, 1997: Issue 4. 30. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995; 273: 408-12.
The Authors Debra Evans RGN, BSc, PhD(Physiology), Lecturer in Research Methodology
Lucy Land RGN, DPSN, BSc, PGCE, Senior Lecturer in Research Methodology Health and Social Care Research Centre, University of Central England (UCE), Ravensbury House, Westbourne Road, Birmingham B15 3TN, UK. Correspondence to Dr Debra Evans. Paper received 15 September 2000. Accepted 29 November 2000.