Workshop on Sensory Analysis for better quality virgin olive oil

Newsletter Workshop on Sensory Analysis for better quality virgin olive oil 2 October 2015 Milan, Italy organised by the European Commission Directorate General Agriculture and Rural Development and the Joint Research Centre Institute for Reference Materials and Measurements

Workshop 2015 - Proceedings Introduction This newsletter presents the outcome of the scientific workshop on sensory analysis for better quality of virgin olive oil (VOO), which was held in Milan (IT) on 2nd October 2015. Olive oil (OO) quality is regulated at international level by the International Olive Council (IOC) trade standard (COI/T.15/NC No 3/ Rev. 10, November 2015) and organoleptic assessment methods (COI/T.20/Doc. No. 15/ Rev. 8, November 2015; see below), Codex Alimentarius (CODEX STAN 33-1981, Rev. 2015), and accordingly at EU level by Commission Regulation (EEC) No 2568/91 (consolidated in October 2015) establishing a list of chemical and organoleptic characteristics, as well as methods for their analysis, which are regularly updated to include new scientific findings thus improving control of quality and authenticity of OO. Despite this regular revision certain problems did not find appropriate solutions and the sensory analysis is sometimes challenged by putting the objectivity of tasting panels into question and by pointing to a potential lack of harmonization among panels. The EU, member of IOC, is worldwide by far the biggest producer (66%), consumer (62%) and exporter (63%) of OO and considers that the quality and image of OO are one of its most important assets which should be preserved. At the Agriculture Council meeting on 18th June 2012, Commissioner Cioloş presented an action plan for strengthening the European OO sector, which includes as main action item promoting the quality of OO and its control to preserve and promote the image of European OO and to protect / inform consumers better. OO quality is graded by applying certain chemical, physico-chemical and sensory parameters. Classification limits have been established on the basis of the opinion of scientific experts and quality is assessed by recognized chemical and sensory testing laboratories. Globally, the quality of oils present on the market has significantly improved over the last twenty years and the negative image inherited from the past faded. The sensory analysis is complementary to the chemical and physico-chemical analysis and enables a final classification of the VOO tasted including a possible downgrading. The European Commission saw this workshop as an opportunity to identify the difficulties in applying the sensory analysis method and to find ways to solve them and reinforce its use for assessing OO quality. Therefore, this workshop only focused on this method and exchanges were limited to the method itself. The general aim was to propose new ways of improving the sensory analysis to reach the EU s commitment to protect the quality and authenticity of the OO produced, consumed, exported from or imported by the EU. The international audience was composed by heads of sensory panels, representatives of producers, processors, traders and consumers. All these main points were introduced during the opening address by Mr. Jesus Zorrilla (DG AGRI). Human sensory analysis of food In this session moderated by Mr. Sélim Belkhodja (General Manager of Cogehuile, Tunisia), four presentations introduced the physiology of sensory sensation by humans and the applications of sensory analysis to honey, wine and VOO. Ms Catherine Peyrot des Gachons (Monell Chemical Senses Center, U.S.A.) reviewed the human taste and oral sensations, looking at how sensory stimuli are perceived in the oral and nasal cavities. Human evaluation of foods and 2

beverages is largely based upon inputs from three chemosensory systems (i.e. taste, olfaction and chemesthesis), integrated with sensory inputs from other physical attributes of foods such as temperature, texture, colour, and rheological properties. The chemical activations of taste receptor cells in the oral cavity (which generate the five main taste qualities - sour, bitter, sweet, salty, and umami) and olfactory receptor neurons in the upper nasal cavity (which generate myriad odour sensations) are responsible for recognising most of the food flavour. Humans have about 350 odorant receptors in the nasal cavity and the olfactory system uses a combinatorial code to identify odours. Because of the complexity of the olfactory system, only three to four components can be identified in complex mixtures and training does not improve this number. Mixtures with more than 30 equiintense odours tend to smell alike ( olfactory white ). The chemically-stimulated sensations of our epithelia (chemesthesis) also play key roles in human protection but also in the perception of many foods. These include the pungency of OO or the sting of alcoholic beverages. These sensory systems are heavily interactive and in fact the perception of flavour is based on our brain s ability to integrate such disparate and anatomically independent sensory cues as taste, smell, irritation, and touch. Flavour perception can be divided into two broad categories, the analytical part (e.g. intensity, quality of a flavour) and its pleasantness or desirability. The pleasantness of a flavour, which modulates ingestion or rejection, is determined by both innate (genetically-programmed) and learned factors. For example, sweet is innately attractive and bitter is innately aversive to humans and many other species. Yet, in the proper context, even innately negative components of flavour can become attractive. As an example, two of the most important sensory traits of high quality OO, bitterness and pungency, are innately aversive, yet they can become positive attributes if experienced in the favourable context of fine oils. Ms Angela Savino (Central Inspectorate for the Safeguard of the Quality and Fraud Repression of the Foodfarming Products in Salemo ICQRF, Italy) reported on the development of the sensory profiles of some Italian unifloral honeys. Sensory evaluation may be used to complement physico-chemical and pollen analysis and enables to distinguish the botanical origin of honey and to identify and quantify certain defects (fermentation, impurities, offodours and flavours). In earlier studies, the sensory group within the IHC (International Honey Commission) developed a harmonized glossary and two methods of evaluation: the descriptive semi-quantitative method and the yes/no classification method (ISO 6658:2005). Such methods are based on the ability of trained assessors to evaluate the correspondence of a declared unifloral honey to a memorized standard. An alternative method, the profile method (not yet an official one), was presented which takes into consideration twelve olfactory and olfactorygustatory attributes (e.g. floral, fruity, acidity, etc.). Their intensity for five Italian unifloral honeys: acacia, citrus, chestnut, eucalyptus and honeydew honey was analysed. In this way a typical sensory profile for each type of honey has been produced by a panel of specialized expert assessors. Consequently, a unifloral honey is considered to conform to the botanical declaration of origin if the profile is within the limits of the expected profile. Each unifloral profile evaluated in this study has highlighted specific characteristics. Principal component analysis clearly allows the discrimination of the studied unifloral honeys. The profiling method, created ad hoc for this research, was found to be easy to learn and to use and showed a strong discriminatory power in differentiating the five types of honeys studied here. Mr Yves Chevalier (Institut National des Appellations d Origine INAO, France) introduced the organoleptic analysis of French designation of origin of wines (AOC). Wine is a complex and extremely diverse product. Since many years, legislation requires all wines with a designation of origin to undergo sensory and chemical analysis for each vintage to keep the AOC designation resulting in an undeniable improvement of the wine quality. In 2008, the procedure for checking wine designations was harmonised with that of other agri-food designations including the organoleptic analysis. Sensory analysis of wines is now done just before their release for consumption or prior bottling. The organoleptic analysis is intended to check that a wine complies with all the specification requirements of the AOC designation. The commission in charge of carrying out this analysis is made up of at least five tasters from a group of people who serve as a reference for the particular wine production, namely: Technicians, Product memory holders, i.e. people who retain how the product should taste and smell and Product Users. These juries are assessed and trained on the specific characteristics and potential shortcomings of the designation. A list of the negative traits is drawn up by the particular organisation responsible for protecting and managing the designation concerned (the ODG ) on the basis of the national list validated by INAO. The specific characteristics of the designation, as defined in the specification and, where applicable, any negative traits are shown in the assessment sheet used by the tasters 3

Workshop 2015 - Proceedings evaluating the wine. Tasters give their individual appreciation of the acceptability of the wine with regard to characteristics of its designation. In its guidelines INAO recommends processing the results statistically in compliance with the binomial rule. For example, at least four out of the five tasters have to rate a wine sample as conforming to the specifications in order to grant the designation status. In cases where the sample is rejected, the jury has to indicate the degree of the organoleptic defect (low, medium, high). In practice, the main difficulties are posed by the reliability of the decision taken and the identification of the defects found. INAO resolved most of them by introducing a more stringent selection and assessment of the tasters and by providing them with specific training on the characteristics of the specific characteristics of the designation and its negative traits, but also on how to fill in the assessment sheet. They also simplified the assessment sheets, provided a reference sample and a sample at the lower quality end of the designation, and restricted the number to ten samples per tasting series with a rest period between series, etc. Lastly, increasing the number of tasters per commission and introducing statistical processing in accordance with the binomial rule are important factors for improving the reliability of the decisions taken. Ms Mercedes Fernández Albaladejo (IOC) summarized the sensory analysis of VOO. In 1987, the IOC developed and adopted a method for the organoleptic assessment of VOO (COI/T.20/Doc. No. 15/Rev. 8, November 2015), which is designed to classify VOO into one of its four categories (i.e. Extra VOO - EVOO, VOO, Ordinary VOO, and Lampante VOO). Since then, the method has been updated on several occasions. Four accompanying standards or guides cover aspects dealt with in the method: General basic vocabulary for sensory analysis (COI/T.20/Doc. No 4/Rev. 1); Glass for oil tasting (COI/T.20/Doc. No 5/Rev. 1); Installation of a test room (COI/T.20/Doc. No 6/Rev. 1); Selection, training and monitoring of skilled VOO tasters (COI/T.20/Doc. No 14); and Guidelines for the accreditation of sensory testing laboratories with particular reference to VOO according to standard ISO/IEC 17025:2005 (COI/T.28/DOC.1). The IOC organoleptic assessment method is aligned with international sensory analysis standards. The IOC method separates the sensory analysis of VOO from the subsequent classification of the oils, which is based on the analysis of the data obtained by univariate robust statistics. The assessment stage is performed by a sensory testing panel (8 to 12 members) selected, created, trained and qualified in line with the instructions given in the IOC method and guide. Each panel member shall smell and then taste the oil under consideration. They shall then enter the intensity with which they perceive each of the negative (i.e. Fusty/muddy sediment, Musty-humid-earthy, Winey-vinegary, Acid-sour, Rancid, Frostbitten olives, and others) and positive attributes (i.e. fruity, bitter, and pungent) on the 10-cm unstructured scale shown in the profile sheet provided. The value of the robust coefficient of variation for an attribute, which defines the classification of the sample and measures the reliability of the assessors, must not be greater than 20%. The oil is graded using the median of the defect perceived with the greatest intensity and the median for the fruity attribute. For an EVOO, the median of the defect must be 0 and the median of the fruity attribute above 0. For a VOO, the median of the defect can be above 0 but not more than 3.5 and the median of the fruity attribute must be above 0. IOC provides software packages that allow the grading to be displayed as a table of statistics or a graph. Consequently, the evaluation of the sensory quality of VOO involves perception of both favourable and unfavourable sensory attributes, with evaluation of sensory defects being used to classify oils into various grades. The IOC has also developed a method to determine and check the sensory profile of EVOO with designation of origin (DO) status (COI/T.20/ Doc No 22). World Café & Open Space The second part of the workshop was based on an interactive approach where participants were invited to identify any critical issues related to the IOC s method for the organoleptic assessment of VOO and to brainstorm on how this method could work even better. Four main issues have been highlighted during the world café session: i) Training of the trainers and trainees; ii) Accreditation of sensory panels and harmonisation of existing standards; iii) lack of reference materials (RMs); and iv) Statistics Interpretation of results & Reporting. 4

The participants were then invited to an Open Space session aiming to create time and space for people to engage deeply and creatively around these four issues. They were asked to participate in conversations related to each issue and answer the following questions: - How to improve the current limitations? - What are the obstacles? - Which support will be needed? - What is your commitment? - What is the timing needed? I. Training of the trainers and trainees Participants mentioned that some panels are stricter than others and consequently a harmonisation and a standardised training (on a continuous basis) of the sensory panels are recommended. Exchange of experts and samples among the existing panels should be more frequent as it was observed that some panels are not having access to a representative range of OO covering the existing OO diversity. It is indeed essential that the typicity of an OO is not confounded with a defect. An example of this confusion is the Koroneiki variety which originated from the southern Peloponese. This variety is harvested late in Israel and panels in the EU classify it with a musty defect because they are not familiar with its typicity. A second example is the Black Fruity Olive Oil which is obtained from green and black olives picked ripe and turned into oil after a certain storage time to promote a voluntary and controlled slight fermentation which is often classified as fusty by uninformed panels. Another critical issue is to improve the threshold of detection of the various panels. Indeed, due to the improvement of the global quality of OO, the assessors tend to over-estimate the defects. This is particularly evident amongst new assessors who have less experience with OO and therefore they have difficulty in putting the defects into perspective. Experts have to train the trainers in a standard manner as well as the trainers to train the tasters and panel leaders. Consequently, it was suggested that IOC should: Provide more didactical training documents explaining how the defects are generated, giving the opportunity to discover uncommon OO and providing information on the sensory aspect of OO; and Create a sub-commission of trainers for setting-up a programme to develop appropriate didactical training documents. II. Accreditation of sensory panels and harmonisation of existing standards The IOC standard for sensory assessment of OO was recognised as a good foundation; however, a harmonisation of the possible existing standards for the sensory analysis of OO is essential aiming to develop an ISO standard. A gap exists between official and private panels that need to be aligned by, for example, allowing private panels to participate in the IOC collaborative trials. The different OO grades have to be defined in a more rational way. The sensory panels should be accredited, including the IOC recognised panels, aiming to increase the repeatability and reproducibility of the method. IOC should recommend that panels follow strictly the standard during the collaborative trials. Some participants suggested that IOC should perform unexpected panel controls in order to evaluate their effectiveness and performance. IOC should give the opportunity to panellists to taste every year samples from all OO producing countries, including those with specific defects. It was also proposed to increase the number of panels and the number of specific controls on OO focusing on OO sold in bottles. 5

Workshop 2015 - Proceedings III. Lack of reference materials (RMs) A more efficient training of the panels relies on the availability of RMs for the recognition not only of all existing defects but also their intensities. Sensory attributes are directly ascribable to the strong stimulation of human sensory receptors by volatile and some non-volatile compounds present in VOO. Phenolic compounds stimulate the taste receptors and also the trigeminal nerve eliciting the sensation of bitterness by the former and pungency and astringency by the latter. Volatile compounds are low-molecular weight compounds that vaporise readily at room temperature, which by stimulating the olfactive receptors, are responsible for the whole aroma of the VOO. Consequently, RMs are needed for training a panel to recognise aroma and taste attributes. However, it was mentioned that RMs for fruitiness, bitterness and pungency are of less priority. Participants debated on which OO should be used as the matrix: manufactured vs. neutral vs. commercial. The addition of a chemical in the neutral OO could be difficult for its dilution in oil. It was also pointed out that a defect is not due to the presence of one volatile substance but several, and a defect is a moving target so it is important to investigate how to preserve ( freeze ) it at a certain development stage. In order to recognise an aroma and its evolution, continuous training and the provision of appropriate RMs to sensory panels is necessary. An OO having a defect at highest intensity should be prepared that can be diluted later on. Reference samples should also be issued from one source to give to panellists a consistent basis of comparison. The idea to create a chemical profile was suggested to support each defect. RMs can also be requested by other stakeholders like retailers, producers or bottlers. It is obvious that manufacturing the whole range of defects will be an enormous task but the OO community should start somewhere. IV. Statistics - Interpretation of results & Reporting Considering statistics, it was obvious that the challenges are OO at the borderline zones between EVOO and VOO and between VOO and lampante VOO for which specific criteria should be adopted to solve the classification of such OO into the right category based on their sensory attributes. At the EU level, the current limit between an oil that can be bottled (i.e. EVOO and VOO) and one that cannot (i.e. Lampante VOO) is a median of 3.5 which includes the uncertainty. As the official panels are guided by the in dubio pro reo ( when in doubt, in favour of the accused ), the category of EVOO and VOO could contain several OO of less quality. It was suggested that for EVOO strict statistical conditions should be applied (median of the defect is equal to 0 and the interquartile range is equal to 0). Experts are discussing this delicate topic within an IOC Committee. The experts advised also to improve the readability of the report on sensory analysis including the interquartile range as well as enhancing the precision of the sensory assessment in order to improve the interpretation of the results for subsequent enforcement actions, and to allow the use of the results by the producers and/or the press. Conclusions Mr Franz Ulberth (DG Joint Research Center) and Mr J. Zorrilla summarised and closed the workshop, respectively. Strategic directions were drawn for aspects related to training, harmonisation, RMs and statistics. Statistics was clearly a hot topic and improvements to the current system are awaited by producers, bottlers, distributors and importers. It was also highlighted that expectation and interpretation of sensory characteristics of VOO might be different for panellists / consumers. Therefore, a reflection on how to deal with typicity of an OO (e.g. black fruity OO) is needed. Finally, the EU made the proper working of the sensory panels a priority to ensure fewer conflicts between panels and other stakeholders having some concerns about the objectivity of the sensory analysis of OO. In the near future the EU will present the outcome of this workshop in different fora: in meetings with its Member States, as well as to the IOC with a view to set up working groups dealing with the four identified issues under the auspices of IOC. 6