Value of animal models for predicting hypersensitivity reactions to medicinal products

Non-IgE-mediated hypersensitivity induced by multivitamins containing Tween-80

Multivitamins have been widely used for years. Adverse reactions, especially hypersensitivity, to multivitamins are becoming noteworthy. However, the classification of hypersensitivity is confusing, and the trigger is unknown. Multivitamins consist of two vials labelled vial 1 containing Tween-80 and vial 2. Multivitamins without Tween-80 were used as a contrast. Behaviouristics, histamine, IgE, and blood pressure of beagle dogs and guinea-pigs were investigated by observation, ELISA and sphygmomanometer, and degranulation and apoptotic of RBL-2H3 cells were assayed by spectrophotometry and flow cytometry. The results showed that dogs suffered from multiorgan anaphylactoid symptoms, and dramatically decreased blood pressure, and high plasma concentrations of histamine after the first administration of multivitamins and multivitamins vial 1, which contains Tween-80, compared to the control, multivitamins vial 2 or multivitamins without Tween-80. In anaphylaxis assay, guinea-pigs did not display any anaphylaxis symptoms and there were no changes in plasma histamine and IgE concentrations in the multivitamins and multivitamins vial 1 groups or in the multivitamins vial 2 and multivitamins without Tween-80 groups except ovalbumin. Compared to the control, the release of β-hexosaminidase and histamine, and the apoptosis of non-antigen-sensitized RBL-2H3 cells significantly increased in the Tween-80 and multivitamins and multivitamins vial 1 groups in a concentration-dependent manner. However, there was no alteration in multivitamins vial 2 and multivitamins without Tween-80 groups. The results indicate that the hypersensitivity induced by multivitamins may be anaphylactoid reaction, but not anaphylaxis. Multivitamin-induced release of inflammatory factors is triggered by Tween-80 through a non-IgE-mediated pathway.

Biocompatibility of cross-linked hyaluronate (Gel-200) for the treatment of knee osteoarthritis

OBJECTIVE

To compare the biocompatibility and immunogenicity of two intra-articular hyaluronan formulations, Gel-200 (Gel-One(®)) and hylan G-F 20 (Synvisc(®) series).

EXPERIMENTAL DESIGN

A comparison of the biocompatibility of Gel-200 and hylan G-F 20 was made using a rat subcutaneous air pouch model and the knee joint of normal rabbits. Immunogenicity was evaluated using a homologous passive cutaneous anaphylaxis (PCA) assay in guinea pigs.

RESULTS

In the air pouch model in rats, characteristic fibrous belts formed in the subcutaneous tissue. Injection of hylan G-F 20 into the air pouch induced granulomatous nodules primarily composed of macrophages, multinucleated giant cells, and eosinophils accompanied with the test material in the center of the nodules in the fibrous belt. Furthermore, the thickness of the fibrous belt in the hylan G-F 20 group increased significantly compared to the saline group. Injection of Gel-200 into the air pouch induced neither granulomatous inflammation nor significant thickening of fibrous belt, while foamy macrophages containing the test material were observed. Intra-articular injection of hylan G-F 20 into the rabbit knee joints induced granulomatous inflammation, eosinophil infiltration, and significant increase in the number of cells in the synovial fluid, while these findings were absent in the Gel-200 group. In the immunogenicity assay, hylan G-F 20 induced a positive PCA reaction, but the Gel-200 did not.

CONCLUSION

Gel-200 showed more favorable biocompatibility and less immunogenicity compared to hylan G-F 20. Gel-200 is expected to be a single injection hyaluronan product with less safety concerns for the treatment of knee osteoarthritis (OA) pain.

Methods of evaluating immunotoxicity

Immunotoxicology is an important aspect of the safety evaluation of drugs and chemicals. Immunosuppression, (unspecific) immunostimulation, hypersensitivity and autoimmunity are the four types of immune-mediated adverse effects. However, the nonclinical assessment of immunotoxicity is at present often restricted to animal models and assays to predict unexpected immunosuppression. There is, however, no general consensus that a variety of assays can be considered depending on the compound to be tested. A major issue is whether histological examination of the thymus, spleen, lymphoid organs and Peyer's patches is a reliable predictor of immunosuppression or whether immune function should also be assessed. A T-dependent antibody response assay, either the plaque-forming cell assay or anti-keyhole limpet haemocyanin enzyme-linked immunosorbant assay, is recommended as a first-line assay. A variety of assays, including lymphocyte subset analysis, natural killer-cell activity, lymphocyte proliferation, delayed-type hypersensitivity, cytotoxic T-lymphocyte activity and macrophage/neutrophil function assays, can also be used. In certain circumstances, host resistance assays can be considered. With the exception of contact sensitisation, very few animal models and assays can reliably predict the potential for (unspecific) immunostimulation, hypersensitivity or autoimmunity. A major limitation of immunotoxicity risk assessment is the lack of human data. Immunological end points and clinical criteria to be included in clinical trials and epidemiological studies have to be carefully standardised and validated.

Popliteal lymph node assay: facts and perspectives

The popliteal lymph node assay (PLNA) derives from the hypothesis that some supposedly immune-mediated adverse effects induced by certain pharmaceuticals involve a mechanism resembling a graft-versus-host reaction. The injection of many but not all of these compounds into the footpad of mice or rats produces an increase in the weight and/or cellularity of the popliteal lymph node in the treated limb (direct PLNA). Some of the compounds known to cause these adverse effects in humans, however, failed to induce a positive PLNA response, leading to refinements of the technique to include pretreatment with enzyme inducers, depletion of CD4(+) T cells or additional endpoints such as histological examination, lymphocyte subset analysis and cytokine fingerprinting. Alternative approaches have been used to improve further the predictability of the assay. In the secondary PLNA, the test compound is injected twice in order to illicit a greater secondary response, thus suggesting a memory-specific T cell response. In the adoptive PLNA, popliteal lymph node cells from treated mice are injected into the footpad of naive mice; a marked response to a subsequent footpad challenge demonstrates the involvement of T cells. Finally, the reporter antigens TNP-Ficoll and TNP-ovalbumin are used to differentiate compounds that induce responses involving neo-antigen help or co-stimulatory signals (modified PLNA). The PLNA is increasingly considered as a tool for detection of the potential to induce both sensitization and autoimmune reactions. A major current limitation is validation. A small inter-laboratory validation study of the direct PLNA found consistent results. No such study has been performed using an alternative protocol. Other issues include selection of the optimal protocol for an improved prediction of sensitization vs autoimmunity, and the elimination of false-positive responses due to primary irritation. Finally, a better understanding of underlying mechanisms is essential to determine the most relevant endpoints. The confusion resulting from use of the PLNA to predict autoimmune-like reactions as well as sensitization should be clarified. Interestingly, most drugs that were positive in the direct PLNA are also known to cause drug hypersensitivity syndrome in treated patients. This observation is expected to open new avenues of research.

Immunotoxicology: role in the safety assessment of drugs

The immunotoxic effects of drugs are divided into immunosuppression, immunostimulation, hypersensitivity and autoimmunity. The major adverse consequences of immunosuppression are infectious complications and virus-induced malignancies. Flu-like reactions, more frequent autoimmune diseases and hypersensitivity reactions to unrelated allergens, and inhibition of drug-metabolising enzymes are the adverse effects related to immunostimulation. Hypersensitivity reactions are the most frequent immunotoxic effects of drugs. They include immune-mediated ('allergic') and non immune-mediated ('pseudoallergic') reactions. Drug-induced autoimmune reactions, either systemic or organ-specific, are seemingly rare. A review of drug-induced immunotoxic effects demonstrates that immunotoxicity is a significant cause of morbidity and even mortality. As immunotoxicologists have long focused on immunosuppression, the nonclinical immunotoxicity safety assessment of unexpected immunosuppression is based on a number of relatively well standardised and validated animal models and assays. However, there is no general consensus regarding the minimal requirement for this assessment. Many different assays can be used to extend the assessment case by case. Few animal models and assays have been validated for use in the nonclinical safety assessment of unexpected immunostimulation. The situation is worse regarding the prediction of hypersensitivity and autoimmune reactions. Our limited understanding of the molecular and cellular mechanisms of immunotoxicity accounts, at least partly, for this situation. Recent guidelines for the immunotoxicity safety assessment of drugs, even though conflicting on several points, will serve as an impetus not only to refine current animal models and assays, but also to search for better alternatives. The new data generated will have to be interpreted and extended to animal species other than just rodents. Likewise, animal results will have to be compared with findings in humans. The search for immunological endpoints that can be used in several animal species and in humans will therefore become essential. Specific endpoints and clinical criteria that can be included in clinical trials to further investigate the potential for immunotoxicity of new drugs will have to be defined. Because immunotoxicity plays a key role in drug-induced adverse effects, the role of immunotoxicology in drug safety assessment is indisputable and the systematic nonclinical as well as clinical immunotoxicity assessment of every new drug is deemed essential.

Genetic aspects of immune-mediated adverse drug effects

Adverse drug effects (ADEs) are of great importance in medicine and account for up to 5% of all hospital admissions. ADEs can arise from several mechanisms and a wide range of drugs can cause immune-mediated ADEs (IMADEs). For a drug to elicit an IMADE, it must be both immunogenic (that is, able to sensitize the immune system) and antigenic (that is, able to evoke a response from a sensitized immune system). Unlike protein therapeutics, small-molecule drugs (or xenobiotics) are usually neither immunogenic nor antigenic. IMADEs are therefore the result of complex interactions between drug-metabolizing enzymes, immune sensitization and immune effectors. The genetic aspects of this interplay are discussed in this review.

Detection of systemic hypersensitivity to drugs using standard guinea pig assays

The most commonly used assays designed to detect either skin or systemic immune-based hypersensitivity reactions are those using guinea pigs (GP). We obtained data from various FDA records to evaluate the correlation between GP assay results and reported post-marketing systemic hypersensitivity reactions. We examined the new drug application (NDA) reviews of approved drugs for the results of GP assays. Post-marketing human data were extracted from the FDA adverse event reporting system (AERS). Drug usage data were obtained from a commercial database maintained by IMS Health Inc. We found 83 (21%) of 396 drugs approved between 1978 and 1998 had reported GP test results. Among these 83 drugs, 14 (17%) were found to have positive results in at least one GP assay. Simple reporting index (RI) values for systemic hypersensitivity reactions were calculated from AERS data and usage to produce the index of adverse event reports per million shipping units of drug. A variety of definitions of positive human response were examined. A statistically significant association was seen for rash between post-marketing and clinical trials adverse event reports. No statistically significant associations between human data and GP test results were observed. These data suggest that standard GP assays have limited ability to predict human systemic hypersensitivity potential for pharmaceuticals.

Evaluation of hypersensitivity to anthraquinone-related substances

Evaluation of hypersensitivity to anthraquinone dyes, which are widely used as color additives, was performed in guinea pigs. In animals sensitized with 1% commercial-grade D&C Green No. 6 (Quinizarin Green SS, CI61565), subsequent challenge with the same dye showed a dose-dependent hypersensitivity reaction, whereas challenge with D&C Green No. 6 purified by recrystallization did not. Guinea pigs sensitized with commercial-grade D&C Green No. 6 showed a positive reaction to challenge with quinizarin, an intermediate in dye synthesis, but not to p-toluidine, another intermediate. It was confirmed that quinizarin itself induced hypersensitivity. The cross-reactivity of four kinds of purified dyes, in which the quinizarin contents were 1.2ppm or less, was evaluated. Purified D&C Green No. 5 (Alizarin Cyanin Green F, CI61570), D&C Green No. 6 and Ext. D&C Violet No. 2 (Alizurol Purple, CI60730) did not produce a hypersensitivity response, while D&C Violet No. 2 (Alizurin Purple SS, CI60725) did. We found that D&C Green No. 5, D&C Green No. 6 and Ext. D&C Violet No. 2 do not intrinsically induce hypersensitivity, and we conclude that the reaction to challenge with the commercial-grade dyes is due to quinizarin contaminating them. In contrast, D&C Violet No. 2 per se induced hypersensitivity. A structure-activity study indicated that 1-hydroxyl and 4-hydroxyl groups in the anthraquinone structure are key factors in hypersensitivity induction by anthraquinone-related compounds.

From clinical to human toxicology: linking animal research and risk assessment in man

Since the 1960s, clinical toxicologists have primarily focused on acute poisonings. This proved very successful as the prognosis markedly improved with the use of resuscitation methods, evidence-based management and new antidotes. This latter area was the first major instance linking animal research and clinical toxicology, as illustrated with N-acetyl-cysteine or specific antibodies. Simultaneously the evolution of poison centers was a critical turning point as '2nd generation' centers are increasingly involved in risk assessment and toxicovigilance. Human toxicology is a broader area in that it is also involved in the toxicity evaluation of xenobiotics with the resulting need to link animal research and risk assessment to match the results of preclinical studies with clinical observations. However, this is not an easy task as experimental and clinical toxicologists seldom share ideas and expertise. Immunotoxicology is an example of this situation. Most of the available data on immunosuppression was obtained in animals and not in man, whereas allergic reactions have been extensively investigated in man, but overlooked in animals until recently. One of the major challenges facing toxicology is to bridge the gap between animal research and risk assessment in man. Human toxicology is expected to play a role in taking up this challenge.

[Preclinical evaluation of the safety of biotechnology products: specific aspects]

Biotechnology-derived products represent a class of increasingly numerous drugs. One of their major characteristics is extreme diversity, which requires specific approaches for the preclinical evaluation of their safety. The selection of relevant animal species is not easy, as most of these products are human-specific. Thus, only one species will often be used, i.e. primates. As most of these products are large molecules, they can be directly immunogenic. When they are human-specific, no animal model is available to predict the risk. Many biotechnology-derived products have an expected influence on the immune system. This must be taken into account in the preclinical strategy of immunotoxicity evaluation that is now required for every new drug. As conventional toxicity testing is generally limited, safety pharmacology studies should include more than the core battery of assays required by current guidelines in order to complement missing data as much as possible. Because of these particularities, a comprehensive investigation of metabolism and pharmacokinetics is not usually needed. Some products can cross-react with cellular components not intended as therapeutic targets. It is, therefore, essential to rule out the risk of possible cross-reactions that can result in adverse effects. Finally, viral safety is a crucial component of the preclinical safety evaluation of these products. Overall, biotechnology-derived products raise specific issues because of their innovative and original characteristics, and it is difficult to address all these issues if not by using a case-by-case approach.

Diclofenac activates T cells in the direct popliteal lymph node assay and selectively induces IgG(1) and IgE against co-injected TNP-OVA

Non-steroidal anti-inflammatory drugs (NSAIDs) are frequently associated with immune-mediated hypersensitivity reactions. The NSAID diclofenac is associated with several distinct allergic and autoimmune-like reactions including anaphylaxis, idiosyncratic hepatotoxicity and autoimmune hemolytic anemia. The aim of this study was to examine the immunostimulating potential of diclofenac in the direct popliteal lymph node assay (PLNA) and reporter antigen PLNA. In BALB/c mice, diclofenac caused dose-dependent increases in PLN weight and PLN cellularity in the direct PLNA; 0.25 mg was non-immunostimulating whereas 0.50-1.00 mg caused a significant PLN reaction. In the direct PLNA, diclofenac also increased the percent of T cells in the PLN with activated phenotypes (CD44(high)CD62L(low) and CD44(high)CD62L(high)). Finally, the magnitude of the diclofenac-induced direct PLN reaction was significantly reduced when the assay was conducted in T-cell-deficient mice. When co-injected with the reporter antigen TNP-Ficoll (trinitrophenyl Ficoll), 0.50 mg diclofenac caused significant increases in PLN weight, PLN cellularity, and induced IgM and IgG(1) anti-TNP antibody forming cells (AFCs) in the PLN. In a final set of studies, a TNP-OVA PLNA was conducted using diclofenac, phenobarbital (negative control) and streptozotocin (positive control). As expected, phenobarbital (1.00 mg) failed to cause an increase in PLN cellularity or induce AFCs in the PLN. Streptozotocin (1.00 mg) caused significant increases in PLN cellularity, IgM AFCs, and selectively induced IgG(2a) and IgG(2b) AFCs against TNP-OVA. Likewise, diclofenac caused dose-dependent increases (0.25-1.00 mg) in PLN cellularity and IgM AFCs. However, in contrast to streptozotocin, diclofenac caused a selective dose-dependent increase in both IgG(1) and IgE AFCs. Finally, an increase in the intracellular level of IL-4, but not INFgamma, was detected in CD4(+) PLN cells following the injection of diclofenac mixed with TNP-OVA. Collectively, these data suggest that diclofenac: (i) induces a T-cell-dependent direct PLN reaction that; (ii) provides non-cognate help for IgG AFC production when co-injected with TNP-Ficoll, possibly through the formation of neo-antigens; and (iii) possesses intrinsic adjuvant activity that selectively induces IL-4 mediated production of IgG(1) and IgE against co-injected TNP-OVA.

Vaccines: predicting the risk of allergy and autoimmunity

The field of vaccines is markedly evolving with the introduction and development of many new concepts and formulations, as well as new indications. Based on the current clinical experience, vaccines can be considered safe in most cases. Nevertheless, allergy and, to a lesser extent, autoimmunity have repeatedly been described or suspected as rare adverse consequences of human vaccines. The mechanisms of these adverse reactions are ill-elucidated, if at all. No animal models have been adequately standardized and validated to predict the risk of allergy and autoimmunity associated with vaccines. However, a number of existing models can be considered for use, but need refinement to be applied to vaccine evaluation. Finally, because the preclinical safety evaluation has not received much attention in the past, efforts should be paid to design specific and cost-effective procedures to meet the current expectations.

Investigating the TNP-OVA and direct popliteal lymph node assays for the detection of immunostimulation by drugs associated with anaphylaxis in humans

Using current animal models, it is not possible to identify low-molecular-weight compounds (LMWCs) that are likely to be associated with anaphylaxis. It is generally accepted that the ultimate effector mechanism involves drug-induced IgE antibody. The objective of the present study was to determine if diclofenac, zomepirac and glafenine, which are associated with anaphylaxis in humans, have immunostimulating potential in the murine TNP-OVA (trinitrophenyl-ovalbumin) popliteal lymph node assay (PLNA), and more specifically to determine if the immunostimulation caused by these LMWCs results in IgE antibody production. These LMWCs were chosen because both zomepirac and glafenine were removed from the market due to high association with anaphylaxis, and diclofenac, which remains on the market, is frequently associated with anaphylaxis. In addition to conducting a TNP-OVA PLNA, the immunostimulating potential of these compounds was examined in the direct PLNA. When co-administered with TNP-OVA, all three LMWCs caused dose-dependent (0.25, 0.50, 1.00 and 1.25 mg) increases in popliteal lymph node (PLN) weight and cellularity that were observed beginning with the 0.25-mg dose. In addition, beginning with the 0.25-mg dose, all three compounds caused dose-dependent increases in TNP-OVA specific IgM and IgG(1) antibody-forming cells (AFCs). Diclofenac induced an isotype switch and caused a dose-dependent increase in the number of IgE AFCs with no detectable IgG(2a) AFCs and minimal high-dose-only IgG(2b) AFCs. Zomepirac induced IgE, IgG(2a) and IgG(2b) AFCs following the injection of 0.50 mg only, and glafenine induced IgE, IgG(2a) and IgG(2b) AFCs following the injection of 0.50-1.00 mg. In the direct PLNA, diclofenac caused dose-dependent increases in PLN weight and cellularity that were observed beginning with dose of 0.50 mg, whereas zomepirac failed to increase any PLN parameter and glafenine only increased the PLN weight. These results suggest that diclofenac, zomepirac and glafenine are immunostimulating LMWCs in the TNP-OVA PLNA with the potential to induce IgE antibody against a co-administered hapten-conjugate. Furthermore, these results suggest that the TNP-OVA PLNA offered significant advantages over the direct PLNA. Although it is not realistic to suggest that a single assay, based on a low number of test compounds, can identify all LMWCs with the potential to cause anaphylaxis in humans, these observations do demonstrate the potential utility of the PLNAs in examining LMWC-induced immunomodulation and support further development and investigation of the assays.

Task force report: future research needs for the prevention and management of immune-mediated drug hypersensitivity reactions

Immune-mediated drug hypersensitivity reactions (IDHR) have a significant impact on clinical practice, drug development, and public health. However, research to understand IDHR mechanisms and to develop diagnostic and predictive tests has been limited. To stimulate more research, a task force with representatives from the key stakeholders (research clinicians, regulatory scientists, and immunotoxicologists from the pharmaceutical industry) was assembled to identify critical data gaps and opportunities and to make recommendations on how to overcome some of the barriers to IDHR research and address research needs. It is hoped that this report will act as a springboard for future discussions and progress toward increased funding and development of organizational structures for IDHR research.

Induction of delayed-type hypersensitivity to sulfamethoxazole in mice: role of metabolites

Although cutaneous adverse drug reactions (ADRs) are relatively frequent and potentially severe, their mechanisms are poorly understood and no validated predictive experimental model is available. Sulfamethoxazole (SMX) is commonly used to treat infections in HIV-positive patients and severe cutaneous ADRs have been described. This study was undertaken to test whether sensitization to SMX could be achieved in mice using a combination of in vivo and in vitro endpoints. No delayed-type hypersensitivity (DTH) response could be evidenced following SMX injection in the back and subsequent challenge into the footpad or onto the ear. Pretreatment with the enzymatic inducers phenobarbitone and betanaphtoflavone, or depletion in CD4(+) T-lymphocytes were not successful either. In contrast, the injection of SMX/S9 mix in the back and challenge with SMX/S9 mix induced a significant increase in footpad thickness. A significant proliferation of spleen cells from SMX- or SMX/S9 mix-treated mice was evidenced following incubation with SMX/S9 mix, but not SMX alone. This study provides indirect evidence that SMX metabolites are involved and confirms previous in vitro results obtained with lymphocytes from patients with a history of SMX-induced ADRs cultured with murine microsomes. Further investigations using other drugs known to induce similar ADRs are warranted to establish the predictive value of this murine model.

Use of transgenic animals to investigate drug hypersensitivity

Hypersensitivity reactions to drugs and environmental agents are often due to exaggerated humoral (Th(2)) or cell mediated (Th(1)) immune responses with typical cytokine profiles. Overexpression of Th(2) cytokines, such as IL-4, IL-5 or IL-13 in mice, enhances an IgE antibody mediated response, while deletion of these cytokines attenuates and/or prevents allergic responses. Conversely, modulation of Th(1) cytokine gene expression may affect cell-mediated immune responses. Therefore, cytokine transgenic mice are used as investigative tools to study potential chemicals and/or drug allergies. In addition to cytokines and chemokines, other factors are important for the development of allergic responses, such as IgE, Fc receptors, vasopressin and several other factors, which can be tested in transgenic mice.

Drug allergy diagnosis in humans: possibilities and pitfalls

Due to the potential hazards of drug allergies, an early and reliable diagnosis is crucial. The use of in vivo tests is not recent but, because of the hazards of skin testing in patients with a history of anaphylaxis, they had been abandoned for a while. Recent reevaluations have shown that for some drugs, e.g. antibiotics-reliable skin tests can ensure the diagnosis of drug allergy in up to 70% of cases. Many in vitro tests based on well-defined mechanisms, e.g. the basophil degranulation test have been used for the diagnosis of totally unrelated allergic mechanisms. It is almost impossible to interpret their validity as diagnosis tools. Nevertheless, other tests, e.g. the lymphocyte transformation test which have been evaluated in well-conducted recent studies, seem to have a good predictive value. Their use is still restricted to clinical trials or research studies. A reliable clinical approach as well as a detailed examination of the drug intake remains obligatory to diagnose drug allergy. Available in vivo and in vitro tests are sometimes used to confirm the diagnosis. The sensitivity and specificity of these tests is evaluated in clinical studies. Research to improve the existing tests and to develop new diagnostic tools is still of paramount importance.

Autoimmunity and toxicity testing

Major difficulties when addressing autoimmunity today within the context of regulatory toxicology are the following: (i) the incidence of auto-immunity related to chemical exposure is not known; (ii) the mechanisms involved are not understood in most instances; and (iii) no fully validated models are available. Although no relationship has been firmly established between the serum levels of autoantibodies and the development and/or the severity of autoimmune diseases, these are considered the hallmarks of autoimmunity. It has proved impossible so far to detect reproducibly autoantibodies indicative of organ-specific autoimmune reactions induced by chemicals in conventional toxicity testing. The detection of autoantibodies suggestive of a systemic autoimmune reaction has been successful with very few compounds only, and most often using non-conventional strains of animals. Genetically-deficient animal strains were sometimes helpful, but these models should be standardised and validated. This also applies to experimental autoimmune disease models. The popliteal lymph node (PLN) assay is potentially helpful, but more research and validation efforts are warranted. From a regulatory toxicology perspective, the search for serum autoantibodies, as it is performed today, does not seem to be a reliable tool, and progress is more likely to be expected from the design, standardisation and validation of dedicated models.