Human Tissue Engineered Products Today's Markets and Future Prospects

Transcription

1 Human Tissue Engineered Products Today's Markets and Future Prospects Final report for Work Package 2: Comparison of tissue engineering treatment costs with conventional treatment Dr. Bernhard Bührlen Dr. Bärbel Hüsing Fraunhofer Institute for Systems and Innovation Research Karlsruhe, Germany May 26, 2003

2

3 I List of Contents...Page List of Tables...i 1. Expected advantages of tissue engineering treatments compared to conventional treatments Terms of reference Methodology applied Selected case studies Search for relevant data Analysis of literature and data Case study "Wounds due to burns" Indications and treatment options Characterisation and assessment of the available database Analysis of the cost-effectiveness of treatments Case study "Wounds due to ulcers" Indications and treatment options Characterisation and assessment of the available database Analysis of the cost-effectiveness of treatments Case study "Articular cartilage defects of the knee" Indications and treatment options Characterisation and assessment of the available database...33

4 II...Page 6.3 Analysis of the cost-effectiveness of treatments Case study "Vascular grafts" Indications and treatment options Characterisation and assessment of the available database Analysis of the effectiveness of treatments Summary: The cost-effectiveness of treatments including tissueengineered products Literature...51 Annex: Strategy for literature search...57 Search in Medline Database...57 Wounds due to burns or ulcers...57 Cartilage defects of the knee...58 Vascular grafts...59 Search in Cochrane-Library Database...60

5 i List of Tables... Page Table 3.1: Categorisation system for health economic literature...7 Table 3.2: Overview of analysed studies on TE treatments...10 Table 4.1: Overview of analysed studies for the treatment of severe burns...15 Table 4.2: Cost analysis for TE products for skin burns...19 Table 4.3: Table 5.1: Table 5.2: Table 5.3: Table 5.4: Table 5.5: Table 6.1: Table 7.1: Table 7.2: Cost analysis for TE and conventional treatments for skin burns...20 Overview of the studies on tissue engineering options for ulcer treatment analysed for this report...22 Overview of the studies on conventional options for ulcer treatment analysed for this report...24 Product costs for tissue engineering and conventional skin ulcer treatments...29 Treatment costs for tissue engineering and conventional skin ulcer treatments...30 Treatment costs for conventional skin ulcer treatments...31 Cost analysis for TE treatments for cartilage defects...36 Overview of analysed studies for tissue engineered vascular grafts...39 Overview of analysed studies for conventional vascular grafts...41

6

7 1. Expected advantages of tissue engineering treatments compared to conventional treatments Modern medicine has developed several therapeutic strategies in order to compensate the loss of function in tissues of the human body which is due to disease, trauma or congenital defects. These strategies represent significant advances in the field of medicine, but have a number of inherent limitations (Lalan et al. 2001): Supplementation of metabolic products, mostly in the form of drugs, hormones and enzymes. Inherent limitation are side-effects of this medication, and the problem that the application of the supplemented substances lacks normal feedback mechanisms and lacks response to in vivo stimuli. This may result in unphysiological levels of the therapeutic substance, leading to acute or long-term complications (e. g. in the case of diabetes acute hypo- or hyperglycemic crisis, or long-term nephro-, vasculo- and retinopathy) (Lalan et al. 2001). Prosthetic substitution. Prosthetic substitution with non-biological materials and devices is used routinely in ophthalmology (glasses, intraocular lenses), traumatology (replacement of joints), cardiovascular surgery (mechanical heart valves, vascular prostheses, stents), and in reconstructive and cosmetic surgery, in dialysis and plasmapheresis. Major drawbacks are that often only a partial restoration of the complex tissues and functions can be achieved, that the quality of life of the affected patients remains impaired, the lack of biocompatibility of the nonbiological materials, leading to enhanced risk of infection and triggering immune reactions, loss of performance due to material failure, wear, absence of physiological self-repair, self-propagation and adaptation, absence of adaptation to changing physiological requirements (Grikscheit et al. 2002). Surgical reconstruction. Surgical reconstruction relies on tissue transfer from a healthy site to an affected site. Often, the availability of healthy donor tissue is limiting. If the transfer occurs between individuals, this requires immunosuppression in the recipient. Organ transplantation. Organ transplantation is severely limited by an increasing donor shortage relative to organ demand, by the requirement for life-long immunosuppression in the organ recipient, and also by the high costs of this type of treatment. Now tissue engineering is emerging as a significant potential alternative or complementary solution in modern medicine (Langer et al. 1993; Persidis 1999; Persidis 2000). The primary goal of all approaches in tissue engineering is the restoration of function through the delivery of living elements which become integrated into the patient. Treatments based on tissue engineering offer the potential to combine advantages of existing treatment strategies while avoiding their limitations. Tissue engineering offers the potential to provide treatments

8 2 according to medical needs and demand (as is possible with drugs or prosthesis/medical devices), which respond physiologically to in vivo stimuli, and have the ability of selfrepair, self-propagation and adaptation to changing conditions (as e. g. in organ transplantation). This suggests that tissue engineering could provide treatments that might be superior to existing ones in terms of effectiveness, prevention of side-effects or longterm complications, practical aspects (such as availability), and quality of life. However, these potential general benefits of tissue engineering have to be proven for individual applications and products. Regarding possible outcomes, three types of tissue engineering treatments can be distinguished (Brown et al. 2001): (1) Tissue engineering offers biology-derived solutions within existing applications, thus competing directly with established alternative treatment strategies, (2) Tissue engineering offers the potential to provide advantages over currently available treatments, e. g. regarding saving of lives, improvements in treatment outcomes and quality of life, reductions in health care costs (Advanced Technology Program 1999, 2002) (e. g. due to the availability of less expensive treatments for major health problems, due to achieving long-lasting, sustainable health effects (Petit-Zeman 2001), due to reducing the need for multiple, repeated intervention (Brown et al. 2001)), (3) Tissue engineering offers the potential to create totally new and unique treatment concepts for human health problems which could not be treated satisfactorily before (Brown et al. 2001). In type (1), tissue-engineered products compete directly with established alternative treatment strategies because their clinical outcome is comparable. Therefore, the treatment costs of the tissue-engineering option, compared with the treatment costs of the established conventional gold standard treatment, will most likely have a major impact on the competitiveness of the tissue-engineered products. Cost-costanalyses are the type of study designed to resolve questions of this type. In type (2) and (3), tissue-engineered products offer clinical, financial or other advantages over currently available treatments, but possibly at different costs. Therefore, it is necessary to not only compare pure cost data but to also take differences in the effects into account. For this type of questions, cost-effectiveness-analyses are the adequate type of research design.

9 3 2. Terms of reference Tissue-engineered products have to compete with conventional products and treatments in health care. The costs and cost-effectiveness of tissue-engineered products, compared to established conventional treatments, will be a major factor influencing this competitiveness. It is the aim of this report to gain insights into the present and future competitiveness of tissue-engineered products by the evaluation of published data on costs and cost-effectiveness of tissue-engineered products and treatments. The tissueengineered treatments are to be compared to a conventional treatment option that is accepted as the best option available for the respective disorder (conventional "gold standard treatment"). This analysis is to be carried out in the form of three casestudies.

10

11 5 3. Methodology applied 3.1 Selected case studies The analysis was carried out for three case studies. Case studies were selected in close cooperation with IPTS and DG ENTERPRISE according to the following criteria: Relevance for the present and future application potential of tissue engineering, Availability of empirical data on costs and effectiveness for the TE application as well as the corresponding conventional gold standard treatment. The results obtained in WP 1 and WP 3 indicated that the availability of empirical data was to be expected only for tissue engineered skin substitutes for the treatment of burns and ulcers, and for autologous cultured chondrocytes to be used for the repair of cartilage defects of the knee. These TE applications were chosen as case studies. In order to also cover a cardiovascular application of tissue engineering, tissueengineered vascular grafts were selected as an additional case study, because clinical experience with more than 100 patients has been accumulated with synthetic vessels covered with autologous endothelial cells. Clinical experience with other possible cardiovascular TE applications (e. g. heart valves, cell therapy to support impaired cardiac function) is not yet as extensive as for a certain type of tissue engineered vascular grafts. For each of these TE applications, the relevant indications were identified by literature analysis, and the corresponding "conventional" gold standard treatments for these indications were also identified by literature analysis. However, a single "gold standard treatment" could not be determined in all cases, and some of the TE treatments were compared in the primary studies to more than one conventional alternative.

12 6 3.2 Search for relevant data The literature search was carried out in the MEDLINE database1, in the Cochrane Library2, the NHS Centre for Reviews and Dissemination Database (NHS CRD)3, the German Agency for HTA4. It was complemented by manual literature search and internet searches. The MEDLINE database is the most extensive database on medical research literature and covers all fields of medicine, including TE as well as health economy. The Cochrane Library is a database that specialises in structured reviews (meta-analyses) of medical therapies and as a sub-division includes health technology assessments (HTAs) which normally report cost-effectiveness data. The NHS CRD database lists national (UK) and international HTAs and economic evaluations. The search strategy for all three case studies included a top-down approach based on the relevant key words (in MEDLINE: MeSH-Terms5) as well a bottom-up search for the names of relevant TE products identified in WP 1. These were combined with the respective key words and MeSH-terms for cost and costeffectiveness studies. From the Cochrane Library, also key papers on non-te treatments were identified as possible "gold standard" treatments. Details on the search strategy are presented in the Annex. In order to make sure that no important economic analysis of TE products had been overlooked, TE experts who were interviewed during WP 1 or were members of the EuropaBio Cells and Tissue Expert Group were asked to provide information on relevant studies known to them. In MEDLINE, 265 articles were identified as possibly relevant; because of less precise search options in the Cochrane Library 799 articles had to be scanned, a total of 1055 articles. After reviewing the abstracts, 42 articles from MEDLINE, 12 from the Cochrane Library and a few from the other sources were analysed in depth, nearly equally distributed over the three case studies. 3.3 Analysis of literature and data According to the research questions a system for the categorisation and evaluation of the literature was developed which helps in extracting the information from the via the German Network on Evidence-Based Medicine MeSH: Medical Sub-Heading in the structured hierarchical thesaurus.

13 7 articles. It is designed to make available all information to answer the research questions and is structured as outlined in table 3.1. Table 3.1: Categorisation system for health economic literature Source Treatment option(s) Indication Scope of the study Type of study Effectiveness/ Utility/ Benefit Costs Cost-effectiveness Influencing factors Comments Authors, year Short description Target group, disease Region: Single country EU USA... Review of effectiveness Cost Cost-cost Success rate Specific indicators Mortality Quality of life Unit: Per case For groups For whole region Time frame: Per year In 1 st year For surgery only For acute treatment For whole treatment Population: Patients General population Cost-effectiveness, costconsequences or HTA Cost-utility Cost-benefit QUALY Monetary benefit Qualitative rating (better/ worse than...) Perspective: Direct treatment costs Direct non-medical costs indirect costs Costs for whole society Qualitative rating (cheaper/ more expensive than...) Categories as for costs, in addition: Measure: $ or per successful case $ or per QUALY Qualitative rating (better/ worse than...) Factors reported that influence costs and cost-effectiveness e. g. Methodological quality of study The source of the literature and the treatment option(s) analysed together with their respective indications in terms of target groups or disease are named. Under the category "Scope of the study" the regions and populations from where the subjects come are important to compare the external validity of the publications.

14 8 The type of the study can be reviews of effectiveness (single studies, narrative reviews or meta-analyses) without cost information as well as the different types of evaluations in health economy: Cost-analyses give information on the costs of a treatment option, cost-cost-analyses compare the costs of two or more treatment options but without taking into account that the options might have different results. Cost-effectiveness-analyses (CEA) establish the relationship between costs and effects with the effects measured as laboratory indicators, mortality, questionnaire scores etc. Health technology assessments (HTAs) often contain the same type of information. In this category, also cost-consequences-analyses are included which analyse both effectiveness and cost of a treatment but do not compute a costeffectiveness-ratio. Cost-utility analyses (CUA) and cost-benefit-analyses (CBA) differ from costeffectiveness-analyses in the measurement of effectiveness: in cost-utilityanalyses, utility measures are computed normally in the form of QUALYs (qualityadjusted life years) which make comparable the cost-effectiveness of very different interventions. In cost-benefit-analyses, the effects are quantified in monetary units so that the costs and the outcomes of a treatment can be balanced and can be directly compared to another treatment. Studies of the CEA or CUA type are most valuable for the present analysis because they allow to compare TE with non-te applications jointly on both dimensions, costs and effectiveness. To compare only costs does not account for the fact that different treatments can yield different outcomes, the comparison of mere costs between two treatment options relies on the assumption that the results are the same for both options, an assumption that will normally not hold true. The measures for effectiveness as well as the cost categories have to be specified as different articles present various types of information. For the costs, their unit (per case, for groups or regions/countries), time frames and perspectives (i.e. completeness) have to be taken into consideration. Consequently, this is also true for the cost-effectiveness indicators. In some of the studies, factors that influence the costs and cost-effectiveness of the treatments are discussed, these factors are extracted in the next category of the analysis system. At last, comments from the authors of the present report on the quality of the primary study or HTA or other information that helps with the interpretation of the studies are documented. Table 3.2 gives an overview of the literature data on TE options that could be identified for the three case studies. For the three case studies, 38 relevant articles were included in the cost-effectiveness analysis. Ten of them considered only non-tetreatments and were identified to provide information on comparator treatments for

15 9 TE applications (see tables 4.1, 5.2 and 7.2). For skin and cartilage TE applications, at least one study in the category CEA or CUA was available (table 3.2). Cost data are available from six studies in the field of cartilage defects, from 13 studies on skin, thereof five on burns alone and seven on ulcers alone (table 3.2). In one study (Jones et al. 2002), the TE treatment is discussed for application in burns as well as chronic wounds but without providing clear criteria which product best to use for which indication. This study is analysed below together with the articles on burns. For vascular grafts, some comparative evidence is available on the effectiveness of TE- vs. non-te treatments (see table 3.2), whereas data on treatment costs are available only for non-te applications (see table 7.2), but not for TE applications. A major problem in the available cost data is the broad variance in the cost factors that are included. The scope of the analyses reaches from a society perspective including indirect costs, e. g. for absenteeism from work over the treatment costs in the hospital with or without inclusion of aftercare, to the mere costs for purchasing the TE product. Given that the treatments provided as well as their costs and outcomes vary broadly, the cost data cannot be easily compared between different studies. As the cost data are presented in the various studies in different currencies, the amounts in were computed. The exchange reference rates used for this purpose are of March 27, Because the cost information only have to serve as a landmark, they were taken "as is" from the publications without any discounting, i.e. without taking into account that costs and effects can be realised at different points in time. The exchange rates used in this report are as follows: AUD (Australian dollar) GBP (Pound sterling) BEF (Belgian Francs) SEK (Swedish krona) CAD (Canadian dollar) USD (US dollar) FRF (Francs Français) Source: European Central Bank, retrieved March 27, 2003, and for the conversion rates of the EURO- Member Countries: retrieved March 27, 2003

16 Table 3.2: Overview of analysed studies on TE treatments Case Type of study n of studies study Cost Cost-cost CEA CUA RE Experts survey Cost Total Skin - burn n=3 Rue et al. 1993; Still et al. 1994; Bello et al n=2 Monstrey et al. 1999; Parente 1997 n=1 Hermans 1998 n=6 n=7 - burn, ulcer, wounds n=1 Jones et al ulcer n=1 Bello et al n=6 Allenet et al. 2000; Augustin et al. 2002; Meaume et al. 2002; NHS Centre for Reviews and Dissemination 2001; NHS Centre for Reviews and Dissemination 2003k; Sibbald et al n=2 Altman 1997; NHS Centre for Reviews and Dissemination 2003j n=7 n=9 10

17 Case Type of study n of studies study Cost Cost-cost CEA CUA RE Experts survey Cost Total Cartilage n=2 Lindahl et al. 2001; NHS Centre for Reviews and Dissemination 2003c, (Jobanputra et al. 2001) n=1 Wildner et al n=1 Arbeitsgemeinschaft ACT und Tissue Engineering 2002 n=6 n=7 Vascular grafts Abbreviations: CEA: Cost-effectiveness analysis, cost-consequences analysis, or HTA; CUA: Cost-utility analysis; RE: Single study, narrative review or meta-analysis on effectiveness without cost information n=3 Jackson et al Minas 1998; Minas et al. 2000; NHS Centre for Reviews and Dissemination 2003a, (Jobanputra et al. 2000) n=5 Deutsch et al. 1999; Jensen et al. 1994; Laube et al. 2000; Lamm et al. 2001; Meinhart et al n=0 n=5 11

18

19 13 4. Case study "Wounds due to burns" 4.1 Indications and treatment options Indications considered in this chapter are mainly severe forms of burns, i. e. fullthickness burns in which all skin layers have been damaged, and which cannot heal by autoregeneration. Moreover, larger areas of the patient's body surface are affected. Improvements in resuscitation, ventilatory support, shock therapy and nutritional management have contributed to increased survival of patients suffering from severe burns. Therefore, early and timely closure of the burn wound has become one of the major problems in the treatment of severe burns (Monstrey et al. 1999; Rue et al. 1993). The "gold standard" treatment option is the application of split-thickness autografts harvested from uninjured areas. However, in severely burned patients, there are not enough donor sites available to provide enough autografts for coverage of injured areas. Due to the paucity of donor sites, multiple graft harvests from uninjured areas may be necessary, yielding grafts of progressively inferior quality. Moreover, even more time elapses until complete coverage of the injured sites can be achieved (Rue et al. 1993). Against this background, several products and treatment strategies have been developed for the coverage of burn wounds. For the following products and strategies information relevant for the assessment of costs and cost-effectiveness in burn treatment were identified in the present study: Non-TE treatments: "Conventional" split-skin autografting. Conventional split-skin autograft is the present gold standard in burn surgery and the treatment of chronic wounds (Jones et al. 2002). Cadaveric allografts Alloderm (Acellular de-epithelialised cadaver dermis) Biobrane (sheets of silicone, nylon mesh and collagen) Integra (sheets of silicone and collagen with glycosaminoglycan) TE treatments Transcyte (formerly Dermagraft-TC; extracellular matrix generated by allogeneic human dermal fibroblasts) as temporary skin replacement

20 14 Apligraf (= Graftskin; living allogeneic bilayered construct containing keratinocytes, fibroblasts and bovine collagen) Dermagraft (living allogeneic fibroblasts grown on degradable scaffold) EPICEL (cultured epidermal autograft) Unspecified cultured epithelial autografts Unspecified cultured autologous keratinocytes Unspecified cultured allogeneic keratinocytes. It has to be noted that some of the products can only be used as temporary skin substitute until the epidermis has healed. In this category fall cadaveric allograft and the products based on it (Alloderm ) and those including layers which are not biodegradable, e. g. Biobrane and Integra. However, some of these temporary wound covers include TE materials to support the healing process, e. g. Transcyte, and have to be removed only in part after the healing. Table 4.1 gives an overview of the literature used for the following analysis.

21 Table 4.1: Overview of analysed studies for the treatment of severe burns Reference Still et al Bello et al Rue et al Monstrey et al Scope (Country, year) USA, USA, ca USA, Belgium, ca Jones et al UK, ca Parente 1997 Hermans 1998 NHS Centre for Reviews and Dissemination 2003h USA, ca World, ca USA, ca Type of treatment option TE only Treatment option, product Type of study Type of available data unspecified cultured epithelial autografts Cost Only product costs, no treatment costs TE only EPICEL Cost Only product costs, no treatment costs TE only TE, non-te TE, non-te TE, non-te TE, non-te non-te unspecified cultured autologous keratinocytes TE: unspecified cultured keratinocytes non-te: cadaveric allograft TE: Transcyte, Apligraf, Dermagraft non-te: Biobrane, Integra, Alloderm, Cadaveric allograft TE: Transcyte non-te: cadaveric allograft TE: cultured keratinocytes non-te: Human cadaver skin, pig skin, hydrocolloid dressings, several gauze dressings, several creams unspecified skin grafting, comparison of inpatient versus outpatient treatment Cost Cost-cost Cost-cost Cost-cost Expert survey Cost-cost Only treatment costs, data derived from clinical trial with 16 patients Only product costs, no treatment costs, data derived from clinical trial with 11 patients in one centre Only product costs, no treatment costs Product costs for non-te, treatment costs for TE and non-te option results are based on model calculations with cost data derived from literature and expert assumptions Qualitative assessments of preferable burn treatment options by 36 opinion leaders in burn care Treatment costs based on actual data derived from a single-centre, prospective cohort study, involving 20 inpatients and 34 outpatients requiring skin grafts for burn treatment 15

22 Characterisation and assessment of the available database For TE applications in the treatment of burns, the data basis for the cost analysis has wide gaps. Most of the cost information available refers to the direct product costs of different burn wound covers. However, this information is only of limited value, because the direct product costs do not take into account how much of the product will be required to achieve the therapeutic aim, and which additional costs arise during the treatment (e. g. costs for surgery, nursing time, duration of hospital stay, etc.). These factors can only be taken into account if treatment costs are considered. However, only three studies could be identified which give information on treatment costs (table 4.1): Rue et al investigates treatment costs of the TE option "unspecified cultured autologous keratinocytes", Parente 1997 employs a computer model to compare the treatment costs arising from the TE option Transcyte versus conventional cadaveric allografts, and the NHS Centre for Reviews and Dissemination 2003h gives treatment costs for the conventional option of skin grafting (however unspecified whether autologous or allogenic). These economic analyses of TE skin products have major methodological shortcomings: the number of treated patients is small, controlled designs are rare, and in the case of Parente 1997) cost data are derived from model computations based on simplified models and experts' assumptions instead of prospective data. There is no analysis that studies costs as well as effectiveness (CEA). 4.3 Analysis of the cost-effectiveness of treatments The direct product costs for tissue-engineered products are shown in table 4.2. On the level of the product alone, costs fall into the range of 10 to 20 per cm² (table 4.2) for TE products, compared with 0.7 to 8.66 per cm² for conventional treatments. Apligraf is by far the most expensive skin substitute with 20/cm², followed by the other TE products. Alloderm, a product from de-epithelialized cadaveric skin costs 8.66/cm². Therefore, all tissue-engineered products have higher product costs than conventional treatments, exceeding them by a factor of 1.2 to 30. The direct product costs, however, do not take into account how much of the product will be required to achieve the therapeutic aim, and which additional costs arise during the treatment (e. g. costs for surgery, nursing time, duration of hospital stay, etc.). These factors are taken into account in the calculation of treatment costs (table 4.2). According to Parente 1997, treatment of large, full-thickness, graftable burns constitutes an extremely resource-intensive process. The necessary procedures are often

23 17 long and cannot be straightforward, because every patient's burn presents a unique challenge to a burn surgeon. Against this background, it is not astonishing, that irrespective of whether TE or conventional treatment options are applied, treatment costs for severe burns can vary in a very broad range, from several thousand to more than 100,000 per patient treated (table 4.3). This broad range of treatment costs makes it difficult to clearly establish cost advantages or disadvantages of TE options versus non-te options, especially as the data base for this is very narrow (see chapter 4.2). The only study which makes a direct comparison of treatment costs between TE and non-te options does not draw on actual data obtained in a clinical setting. Rather, it applies a model calculation (Parente 1997). The result of the model calculation is that the TE option compares favourably with the conventional option because the treatment costs with Transcyte are 18,815 to 94,550, depending on the severity of the burn injury whereas the conventional treatment causes costs in the range of 28,165 to 156,534 (table 4.3). Thus, the Transcyte treatment costs would amount to only % of the conventional treatment costs. However, the results of this model calculation cannot simply be generalized for two reasons: the model has not been validated with data from clinical settings, and in addition, it was assumed that the Transcyte product costs are the same as the product costs for the cadaver allografts. This assumption was due to the fact that the Transcyte product price was not yet known at the time of the publication. As can be seen from table 4.1, this assumption does not reflect the actual situation: Transcyte product costs exceed cadaver allograft costs by more than a factor of 10. Therefore, a new model calculation would be needed to test whether the predicted savings of the Transcyte treatment, compared to cadaver allografts, would still exist if realistic product prices were used in the model. Hermans (1998) presents no quantitative cost data, but indicates the costs for unspecified cultured keratinocytes as "high", resulting in infrequent clinical use. Monstrey et al. (1999) do not calculate treatment costs, but point out the substantial difference in graft costs for allogeneic cultured keratinocytes skin substitutes compared to cadaveric skin as second layers (see table 4.1). At the same time, they do not find significant improvements of the wound healing with the TE option, so that they come to the conclusion that the use of cryopreserved allogeneic cultured keratinocytes can hardly be advocated if the substantial differences in product costs are taken into account (Monstrey et al. 1999). Although tissue-engineered skin-replacements might have potential advantages, such as improved availability, quicker burn wound coverage, avoided shock, infection and reapplication, reduced hospital days, less blood products consumption, faster application, and reduced viral transmission (Parente 1997), the superiority in terms of effects of the TE treatment options to conventional treatments of burns has not yet been proven. The authors of the analysed studies could not find a positive

24 18 impact on wound closure compared to standard treatment (e.g. Rue et al. 1993), except faster epithelialization in the keratinocyte group after two weeks than in the group using cadaveric skin substitutes, an advantage that showed to be transient in the longer run (Monstrey et al. 1999). In addition, in one study cultured epithelial autografts proved to be more susceptible to loss than routine autograft (Still et al. 1994). Because there is no strong evidence for better results of the TE products compared to the cheaper standard treatments, the cost-effectiveness of the skin replacement products in the treatment of burns favours the conventional options. For these reasons, TE skin replacement options are presently only used for critical burn patients for whom additional surgical treatment options are welcomed, irrespective of their cost-effectiveness, in order to reduce their mortality.

25 Table 4.2: Cost analysis for TE products for skin burns Treatment option Costs Cost factors considered Scope Source TE treatment Cultured epithelial autografts (unspecified) /cm² Product alone USA, Still et al Cultured keratinocytes (unspecified) 9.92 /cm² Product alone Belgium, ca Monstrey et al EPICEL*) /cm² Product alone USA, ca Bello et al Transcyte*) /cm² Product alone UK, ca Jones et al Apligraf*) /cm² Product alone UK, ca Jones et al Dermagraft*) /cm² Product alone UK, ca Jones et al Conventional treatment Biobrane*) 0.70 /cm² Product alone UK, ca Jones et al Integra*) 4.87 /cm² Product alone UK, ca Jones et al Alloderm*) 8.66 /cm² Product alone UK, ca Jones et al Cadaveric allograft*) 0.37 /cm² Product alone Belgium, ca Monstrey et al Cadaveric allograft 0.80 / cm² Product alone USA, ca Parente 1997 Cadaveric allograft 0.88 /cm² Product alone UK, ca Jones et al *) Discussed for treatment of burns, ulcers, as well as chronic wounds. Source: Fraunhofer ISI

26 Table 4.3: Cost analysis for TE and conventional treatments for skin burns TE treatment Conventional treatment TE Treatment Costs Conv. treatment Costs Cultured 40,758 autologous keratinocytes (range 9,140- (un- 150,149) specified) Cost factors considered Scope Source - - Average cost per patient (n=16), cost factors not specified, mean of 15.9% of BSA6 treated with keratinocyte grafts Transcyte 18,815 Cadaver allograft 28,165 Model calculation of possible direct hospital cost offsets of Transcyte against cadaveric allograft; excludes costs for Transcyte calculation for burns of 40 % BSA USA, USA, ca Transcyte 53,879 Cadaver allograft 90,717 calculation for burns of 70 % BSA USA, ca Transcyte 94,550 Cadaver allograft 156,534 calculation for burns of 90 % BSA USA, ca Unspecified skin USA, grafting ca Unspecified skin grafting 15,853 Direct hospital costs for inpatient burn treatment (n=20), including surgeon's fees, pharmacy and supplies, antibiotics, anesthesia, X-ray and labs, recovery room, hospital stay no information on severity of burn wounds given 4,662 Direct hospital costs for outpatient burn treatment (n=30), cost categories as above no information on severity of burn wounds given USA, ca Rue et al Parente 1997 Parente 1997 Parente 1997 NHS Centre for Reviews and Dissemination 2003h NHS Centre for Reviews and Dissemination 2003h 20 *) Discussed for treatment of burns, ulcers, as well as chronic wounds. Source: Fraunhofer ISI 6 BSA: body surface area

27 21 5. Case study "Wounds due to ulcers" 5.1 Indications and treatment options Chronic wounds are defined as wounds which do not heal within six weeks. Chronic wounds can be devided into pressure ulcers, which form during sitting or lying without moving. Especially elderly and severely ill people are at risk. Ulcus cruris, venous ulcers, which are caused by venous insufficiency. Diabetic ulcers, diabetic foot, which can emerge in diabetic patients with an illcontrolled blood glucose level. For these different types of ulcers, the following conventional and tissue engineering options were analysed for this report: Conventional treatment options: Saline and other gauze dressings Hydroactive (=hydrocolloid) wound dressings, such as DuoDERM, Comfeel or other hydrocolloid dressings, also combined with enzymatic wound bed preparation Compression therapy (Unna's boot or bandage system) Tissue engineering treatment options: Apligraf Dermagraft BioSeed -S (autologous keratinocytes in fibrin glue) In some of the studies analysed for this report the treatments were restricted to severe forms, e. g. to ulcers that had not adequately responded to conventional ulcer therapy. Table 5.1 gives an overview of the studies on ulcer treatment analysed for this report.

28 Table 5.1: Overview of the studies on tissue engineering options for ulcer treatment analysed for this report Reference Bello et al NHS Centre for Reviews and Dissemination 2003k Sibbald et al Altman 1997 Meaume et al Scope (Country, year) USA, ca USA, ca Canada, ca Canada, USA, ca Typical European cohort, ca Indication NHS Centre for Reviews and Dissemination 2001 therapyresistent venous leg ulcers therapyresistent venous leg ulcers venous leg ulcers venous leg ulcers venous leg ulcers pressure ulcers, venous leg ulcers TE vs. Non-TE TE vs. Non-TE TE vs. Non-TE TE vs. Non-TE TE vs. non-te TE vs. Non-TE Non-TE only Treatment option, product TE: Apligraf (Graftskin) non-te: compression therapy (Unna's boot) TE: Apligraf (Graftskin) non-te: compression therapy (Unna's boot) TE: Apligraf plus compression therapy non-te: compression therapy alone TE: Apligraf and compression therapy (four-layered bandage system) non-te: compression therapy alone (four-layered bandage system) TE: Apligraf non-te: compression therapy Venous leg ulcers: TE: Apligraf non-te: saline gauze, DuoDERM Pressure ulcers: non-te: saline gauze, DuoDERM, Comfeel Type of study CEA cost-cost CEA CEA RE CEA Type of available data prospective, multicentre, randomised, controlled trial carried out at a single centre involving 240 patients Data on effectiveness Data on direct costs of treatment, based on actual data Data on direct costs of treatment, based on actual data Data on effectiveness Data on direct costs of conventional treatment, based on experts estimations Qualitative information on cost relationships between TE and conventional treatment, based on model calculations Computer-based model calculations to evaluate clinical outcomes (ulcers healed, time to heal, ulcer recurrences, infections) and associated costs Information on effectiveness Computer-based model calculations Comparison of TE vs. non-te only for venous leg ulcers, for pressure ulcers only comparison of several non-te options 22

29 Reference Augustin et al Allenet et al NHS Centre for Reviews and Dissemination 2003j Scope (Country, year) Germany, ca France, ca not reported Indication severe, therapyresistent leg ulcers diabetic foot ulcers foot ulcer patients with Type 2 diabetes mellitus TE only; TE vs. Non-TE TE vs. Non-TE TE vs. Non-TE Treatment option, product TE: BioSeed-S non-te: lipid gauze dressings, hydrocolloid dressings TE: Dermagraft non-te: standard therapy, not specified TE: Dermagraft non-te: unspecified control treatment Type of study CEA CEA RE Type of available data Clinical pilot study on the efficacy of BioSeedS in n=17 patients, clinical data on effectiveness cost-relevant data for non-te treatments from literature; comparative cost data based on model calculations Computer-based model calculations based on data from literature Review of two RCTs (n=331 patients) no information on costs and effectiveness Source: Fraunhofer ISI 23

30 Table 5.2: Overview of the studies on conventional options for ulcer treatment analysed for this report Reference Ramsey et al Marston et al Bergeman n et al Scope (Country, year) USA, USA, ; price year 1995 USA Germany 1997 Indication venous leg ulcers Diabetic foot ulcers venous leg ulcers pressure ulcers and venous leg ulcers NHS Centre for Reviews and Dissemination 2003i Non- TE only Non- TE only Non- TE only Non- TE only Treatment option, product non-te: saline gauze; hydrocolloid dressing + compression; Unna's boot treatments not specified, all treatments included which are attributable to diabetic foot ulcer treatment costs non-te: Unna's boot, four-layer compression bandages non-te: gauze, impregnated gauze, calcium alginate, hydroactive wound dressings, hydroactive wound dressing + enzymatic wound bed preparation Type of study CEA Cost cost cost-cost Type of available data Cohort study at single centre, n=81 patients Effectiveness data derived from single study, Data on direct costs of treatment based on actual data Analysis of actual treatments costs for diabetic foot ulcers in a total of 8,905 patients with type 1 or type 2 diabetes Analysis of actual treatment costs, based on 217 patients treated by compression therapy on an outpatient basis Costs include physician, laboratory and tests, wound dressing materials, home health nursing costs Treatment costs for 5 treatment strategies, based on data obtained in 4 hospitals from 120 patients in the inpatient setting Treatment costs consider material and personnel costs from the perspective of hospital administration No effectiveness data available 24

31 Characterisation and assessment of the available database A total of 11 publications was used for in-depth analysis (Table 5.1, 5.2). Eight of these publications address venous leg ulcers, 4 diabetic foot ulcers, and 2 pressure ulcers. All publications on TE treatment options make a direct or indirect comparison to at least one non-te treatment. Six of the 11 publications base their analysis on actual data derived from clinical trials or patients' databases run by a health maintenance organisation, while five publications use computer based model calculations. The six studies which investigate TE options analyse costs as well as effectiveness (CEA). In most of the studies the methodological quality can be assessed as good. So, publications on tissue-engineered products in ulcer care are more abundant than for burns. Especially on Apligraf, several good CEAs could be identified in which the costs and effectiveness of the TE treatment is compared to conventional treatment(s). This might be due to the fact that Apligraf has already been granted market approval for ulcer treatment in the USA in 1998 and in Canada for the treatment of venous leg ulcers in 1997, and for the treatment of diabetic ulcers in 2000 (Agence d'évaluation des technologies et des modes d'intervention en santé (AÈTMIS) 2001; Bello et al. 2001). For Dermagraft and BioSeedS, the other TE applications for ulcer treatment analysed here, the data basis for the cost analysis is scarce: for BioSeedS, only data from a clinical pilot study which involves 17 patients are available (Augustin et al. 2002). A larger, randomized comparative clinical study on BioSeedS involving 240 patients is underway, but the complete results will not be available before late For Dermagraft, a review of two randomly controlled trials was identified, which, however, does not give any information on costs and effectiveness (NHS Centre for Reviews and Dissemination 2003j). In addition, for all three TE options, i. e. Apligraf, Dermagraft and BioSeedS, some cost analysis stem from model computations based on simplified models and experts' assumptions instead of prospective data. In addition, in studies concerning venous leg ulcers, the TE treatment is often compared to compression therapy. However, experts have questioned compression therapy as a valid comparator because it is targeted to the cause and the skin replacement therapy to the consequences. Therefore TE skin and compression therapy are argued to be complementary and not alternative therapies in venous leg ulcers (Noel 2002). 7 Preliminary data from 62 patients have been made available to the project team by the company Baxter, but these information do not contain data on costs, and therefore were not used for the indepth analysis.

32 Analysis of the cost-effectiveness of treatments Table 5.3 gives an orienting overview of product costs for conventional and tissueengineering ulcer treatments. Conventional treatments such as gauze dressings have very low product costs, in the range of < 1/day of treatment. Advanced wound care products such as hydrocolloid wound dressings can be ten times more expensive regarding the material costs of ulcer treatment per patient. Even more expensive are tissue engineered products for ulcer care; the average product costs per patient are in the range of 763 (BioSeedS) to 2,774 (Dermagraft), depending on the product. However, a different picture emerges, if total treatment costs are considered. It has been shown in several studies, that the personnel costs required to apply and change the wound dressings by far exceed the product costs of the dressings. If personnel costs are also taken into account, advanced wound care products such as hydrocolloid dressings, although more costly in purchase, bring about significant reductions in total treatment costs, because of significant reductions in personnel costs and duration of treatment (see e. g. Bergemann et al. 1999, table 5.4). Against this background the questions arises how tissue engineering ulcer treatments compare on a total treatment cost basis, given the fact that their product costs are higher than for conventional and advanced wound treatments. Several studies have addressed this question (NHS Centre for Reviews and Dissemination 2001; Bello et al. 2002; NHS Centre for Reviews and Dissemination 2003k; Sibbald et al. 2001; Meaume et al. 2002; Augustin et al. 2002; Allenet et al. 2000; table 5.4). However, costs for treatment vary widely: Ulcer treatment costs with tissue engineering options cover a range of 1,282 per patient in 6 months to 18,690 per patient and year. However, also costs for the conventional treatments vary broadly: from 1,243 per patient in 6 months to 25,639 per patient and year (table 5.4). To a large extent, this variation is due to differences in the ulcer types (venous versus pressure versus diabetic ulcers, therapy-resistant and severe versus "normal" ulcers), in the cost factors included, in the time scale taken into consideration and the time dimension of the cost estimate (e. g. costs per month, per 12 weeks, per 6 months, per 1 year, until healing; which cannot be transformed into a uniform measure (e. g. costs/year) by simple calculation) and other methodological aspects. Therefore, only data for a TE-option and a non-te option, which were derived under identical conditions, can be compared with one another. Two studies, Meaume et al. 2002) and Allenet et al. 2000), find higher treatment costs for the tissue engineering option than for the conventional treatment options: Meaume et al. 2002) comes to the conclusion that for venous ulcer care in a typical European patient cohort, the treatment with hydrocolloid dressings is more costeffective than with saline gauze, while treatment with Apligraf causes treatment

33 27 costs which are higher by a factor of 5-7, at least in a period of twelve weeks. Allenet et al. 2000) expects costs per treated patient and year of 8,285 for treatment with Dermagraft, and 7,228 for "standard therapy". But since at the same time a better healing rate and shorter time to heal is reported for Dermagraft, lower average costs per healed ulcer result: 8,159 for Dermagraft compared to 8,641 with standard treatment (Allenet et al. 2000, table 5.4). Based on these results, Dermagraft could be attributed relative cost-effectiveness. However, the HTAreport by the NHS Centre for Reviews and Dissemination (2003j) which refers to Mason et al. (1999) finds only a non-significant increase in ulcer healing for Dermagraft compared to the control group. The study by Sibbald et al. (2001) shows treatment costs for venous ulcers nearly on the same level for Apligraf ( 1,282 to 1,370) and for a conventional bandage system ( 1,234 to 1,285). Two other studies find clear cost advantages of the TE treatment option in comparison to conventional treatments: An advantage of Apligraf compared to Unna's boot of approx. 7,000 per patient and year in the long run (one year) is indicated (NHS Centre for Reviews and Dissemination 2001; Schonfeld et al. 2000), which must in part be attributed to the high costs of the conventional alternative in this study. Augustin et al conclude that treatment of severe, therapy-resistant chronic wounds with BioSeedS appears to be good value in comparison to other therapies, as mean annual treatment costs of 4,370 compare favourably with 7,530 (hydrocolloid dressing) and 10,897 (lipid gauze dressing). However, this comparison was not based on actual data which were generated under identical conditions for TE- and non-te-options. A clinical trial is underway which aims at filling this information gap. However, in the treatment of venous ulcers, Apligraf seems to show not only higher costs, but also better effects (e. g. NHS Centre for Reviews and Dissemination 2001; Schonfeld et al. 2000), especially regarding the cumulative probability of ulcer healing after one year (Apligraf: 0.57; Unna's boot 0.31), and the probability of recurrence (Apligraf: 0.030; Unna's boot: 0.037). In a similar trial, 64 % of the cases treated with Apligraf showed a wound closure by six months and also a more rapid healing (median time to wound closure 57 days), compared to a rate of 44 % (median time to wound closure 181 days) for compression therapy alone (Altman 1997). In another study (Meaume et al. 2002), the rate of healed ulcers in 12 weeks is better for Apligraf (45 %) than for saline gauze (35 %), but worse than for Duo- Derm (51 %). However, if the costs per patient healed are calculated, the costs for Apligraf are much higher with 11,396 than for saline gauze ( 2,763) and Duo- DERM ( 1,436) (Meaume et al. 2002). In their model calculation of venous ulcer treatment by a bandage system alone versus treatment with Apligraf plus bandage system, Sibbald et al. (2001) find higher costs for the Apligraf treatment, but also a quicker healing of the ulcers: over the period of 3 months, the Apligraf treatment costs 193 more than the bandage sys-