Exposure To Lead From Intake Of Coffee

11m ago
18 Views
0 Downloads
1.16 MB
32 Pages
Last View : 1d ago
Last Download : n/a
Upload by : Axel Lin
Transcription

Exposure to lead fromintake of coffeeA LOUS follow-up projectEnvironmental project No. 1785, 2015

Title:Editing:Exposure to lead from intake of coffeeMax HansenJens Jørgen SlothRie Romme RasmussenPublished by:The Danish Environmental Protection AgencyStrandgade 291401 Copenhagen KDenmarkwww.mst.dk/englishYear:ISBN no.2015978-87-93352-66-7Disclaimer:When the occasion arises, the Danish Environmental Protection Agency will publish reports and papers concerningresearch and development projects within the environmental sector, financed by study grants provided by the DanishEnvironmental Protection Agency. It should be noted that such publications do not necessarily reflect the position oropinion of the Danish Environmental Protection Agency.However, publication does indicate that, in the opinion of the Danish Environmental Protection Agency, the contentrepresents an important contribution to the debate surrounding Danish environmental policy.Sources must be acknowledged.2

Contents1.Foreword . 42.Sammenfatning og konklusion . 53.Summary and conclusion . 74.Introduction . 95.Study design and methods . 14Introduction . 14Collected coffee samples . 14Determination of lead . 15Quality assurance parameters . 166.Results and discussion .17References . 243

1. ForewordThe List of Undesirable Substances (LOUS) was established by the Danish EnvironmentalProtection Agency (EPA) as a guide for enterprises. It addresses chemical substances of concern,based on their hazardous properties and the volumes used in Denmark. The latest version of LOUSfrom 2009 includes 40 chemical substances or groups of substances [DEPA 2010].During the period 2012-2015, all substances listed on LOUS will be surveyed and further need forrisk management measures will be evaluated. In certain cases, implementation projects will belaunched to achieve the goals laid down in the strategies for these substances.The present project ”Exposure to lead from intake of coffee” was initiated as a LOUS follow-upproject by the Danish EPA. The objective of this study was to estimate the exposure to lead fromcoffee in the Danish population. The investigation was conducted by chemical analyses of coffeebeans as well as brewed coffee. The consumption data was obtained from the Danish food database,which is hosted by the DTU FOOD.The project was carried out from September 2014 to April 2015 by the National Food Institute at theTechnical University of Denmark (DTU Food). Participants from DTU Food were senior advisorMax Hansen, senior scientist Jens Jørgen Sloth and scientist Rie Romme Rasmussen. The qualityassurance was performed by Folmer Eriksen.4

2. Sammenfatning ogkonklusionI DTU Fødevareinstituttets seneste overvågningsrapport over udvalgte kemiske forureninger ifødevarer blev det på baggrund af ældre data anslået, at den gennemsnitlige eksponering af bly frakaffe svarede til 40 % af blyindtaget fra drikkevarer og dermed næsten 20% af det totale indtag afbly fra drikke- og fødevarer. Da hovedparten af det totale indtag af bly stammer fra kosten, udgørdet estimerede bidrag fra kaffe således et betragteligt bidrag til danskernes samlede indtag af bly.En nylig tysk undersøgelse gennemført af BfR viste et højt indhold af bly i kaffe brygget påespressomaskiner af den type, der anvendes på cafeer. Nærværende projekt har derfor til formål atestimere indtaget af bly fra kaffe i den danske befolkning, samt at vurdere, hvor stor en del, derkommer fra henholdsvis kaffe, vand og kaffebryggere. Dette indtag er blevet sammenlignet med detsamlede indtag af bly fra drikkevarer/fødevarer. Det er desuden blevet undersøgt, om der er ensystematisk geografisk variation af blyindholdet i kaffe.For at sikre at de valgte kaffe typer, der blev udtaget, var repræsentative for det danske marked,blev der gennemført et udtræk fra Dansk Husstandspanel (GfK), der bygger på en meget detaljeretregistrering af en række husstandes indkøb af de 566 forskellige varenumre for kaffe, der findes pådet danske marked. Det blev desuden sikret, at de udtagne prøver repræsenterede kaffe fraforskellige dele af verden, således at en eventuel geografisk forskel i blyindholdet kunne fastslås. Ialt blev der analyseret 44 kaffebønne produkter, heraf 9 af de 10 mest solgte varenumre, og 7instant kaffe produkter.For at fastslå, om forskellige kaffebrygningsmetoder har indflydelse på indholdet af bly i drikkeklarkaffe, blev der analyseret på kaffe brygget efter følgende seks metoder: Filterkaffe brygget påkaffemaskine, stempelkande kaffe, kogekaffe, kaffe brygget på mokkakande, hjemmebryggetespresso og espresso fra cafeer. Der blev desuden udført analyser af det vand, der blev anvendt tilbrygningen.Til beregning af den samlede eksponering for bly via kaffe blev de fundne blykoncentrationersammenholdt med konsumdata for kaffe, der blev udtrukket fra den danske kost database afafdeling for risikovurdering og ernæring ved DTU Fødevareinstituttet.Det er almindelig kendt, at analyser af bly er vanskelige og let fører til overestimering af indholdetpå grund af en række diffuse kilder. Derfor blev der i dette projekt gjort et omfattende arbejde for atkvalitetssikre data.Analyserne viste et indhold af bly i de 44 prøver af kaffebønner på 4,5 – 65,3 ng/g med etgennemsnit på 15,9 ng/g. Dette er noget lavere end det, der er blevet fundet i de fleste andre studier,der er gennemført verden over. I disse undersøgelser er bly blot en enkelt af mange sporelementerder er analyseret, og kvalitetssikringen er ikke beskrevet i detaljer. Denne forskel i resultaterkompromitterer derfor ikke validiteten af det aktuelle studie. I instant kaffe blev der i de 7 prøverfundet i gennemsnit 17,9 ng/g. Med undtagelse af kogekaffen blev det fundet, at uanset hvilkenbryggemetode, der blev anvendt, blev alt den bly der var tilstede i kaffebønnerne ekstraheret ud iden drikkeklare kaffe. For kogekaffen ser det umiddelbart ud til at kun ca. halvdelen ekstraheres,5

men på grund af de relativt høje analyseusikkerheder kan det ikke udelukkes, at alt bly ogsåekstraheres ved denne bryggemetode.Der blev ikke observeret afsmitning fra bryggeudstyret i de kaffer, der blev brygget hjemme. På demålinger, der er udført på café kaffe, er der kun analyseret på den drikkeklare kaffe.Indtaget af kaffe blev for danske mænd bestemt til 706 ml/dag i gennemsnit med en 95 percentil på1657 ml/dag. For danske kvinder var de tilsvarende tal 550 ml/dag og 1371 ml/dag. På baggrund afdette blev det beregnet, at mænd i gennemsnit indtager 0,74 µg bly/person/dag og at 95 percentilenindtager 1,73 µg/person/dag. Kvinder indtager i gennemsnit 0,58 µg bly/person/dag med en 95percentil på 1,35 µg bly/person/dag. Det gennemsnitlige indtag svarer til 9,1 % af det samledeblyindtag fra drikkevarer for mænd og 7,1 % for kvinder. Dette er væsentligt lavere end det tidligereestimat på 40 %. Årsagen til dette er, at der i denne undersøgelse er fundet lavere indhold af bly ikaffe end i analyser af kaffe gennemført i de tidligere undersøgelser.I de 8 prøver af drikkeklar kaffe, der blev taget på cafeer, blev der generelt fundet et blyindhold påniveau med det der blev fundet i hjemmebrygget kaffe. Der var dog en enkelt prøve, hvor indholdetvar 67 ng/g. Der blev senere taget en ny prøve fra samme cafe, brygget på samme maskine og denviste et væsentligt lavere indhold. Dette indikerer, at det høje indhold kan skyldes en relativ nyligafkalkning da brugen af syre til afkalkning kan frigøre bly fra lodninger i kaffemaskinen, hvis det ertilstede. Fra en sundhedsmæssig betragtning er indholdet dog ikke så højt, at det vil væresundhedsmæssigt problematisk at indtage en kop kaffe dagligt med dette indhold.Indtaget af bly fra kaffe svarer samlet set til hhv. 4,2% og 3,3% af det totale blyindtag med kostenfor danske mænd og kvinder. Generelt kan det konkluderes, at indtaget af bly fra kaffe er lavt iforhold til indtaget fra andre fødevarer, og at det ikke udgør en væsentlig del af det samlede indtagaf bly med kosten.6

3. Summary and conclusionIn the latest monitoring report on chemical contaminants in food from DTU FOOD it was estimatedthat the average exposure of lead from coffee corresponded to 40% of the lead intake frombeverages and thus to almost 20% of the total dietary intake of lead. This estimate was based on olddata. A recent German study conducted by BfR showed a high content of lead in coffee brewed atprofessional espresso machines of the types frequently used in cafes. The aim of this project was toestimate the intake of lead from coffee in the Danish population, and to assess the amountcontributed from coffee, water and coffee brewers, respectively. This intake has been compared withthe total intake of lead from drinks/food. It has also been investigated whether there is a systematicgeographic variation of the lead content in coffee.To ensure that the selected coffee types were representative for the Danish market, data wereextracted from the Danish Household Panel (GfK) based on a very detailed records of a number ofhousehold purchases of the 566 different code numbers for coffee found on the Danish market. Itwas also ensured that the coffee selected represented beans from different parts of the world so thatany geographical difference in lead content could be determined. In total 44 coffee bean products(among them 9 of the 10 most sold coffees in Denmark) and 7 instant coffee products wereanalysed.To determine whether different coffee brewing methods influence the content of lead in ready todrink coffee, coffee samples based on each of the following six brewing methods were analysed:Filter coffee brewed coffee maker, French press coffee, boiling coffee, coffee brewed in mocha pot,home-brewed espresso and espresso from cafes. The water used for the coffee brewing was alsoanalysed.To calculate the total exposure to lead from coffee, the measured lead concentrations was comparedto consumption data for coffee in Denmark extracted from the Danish Food Database by theDivision of Risk Assessment and Nutrition at DTU FOOD.It is generally known that analysis of lead is difficult and easily leads to overestimation of thecontents due to a number of diffuse sources. Therefore extensive efforts were made in this project toensure quality control of the data.The analysis indicated that the content of lead in the 44 samples of coffee beans was ranging from4.5 to 65.3 ng/g with an average of 15.9 ng/g. This is somewhat lower than what has been found inmost other studies conducted worldwide. In these studies lead was just one of many trace elementsanalyzed and quality assurance is not described in detail. These differences in the results reporteddo thus not compromise the validity of the current study. In the instant coffee an average of 17.9ng/g was found in the 7 samples. With the exception of the coffee boiling it was found thatregardless of the brewing method used, all the lead present in the coffee beans was extracted tobrewed coffee. For boiling coffee, it initially appeared that only about half of the lead was extracted,but because of the relatively high analytical uncertainties in this analysis it cannot be excluded thatall lead present is extracted by this brewing method as well.There was no indication that lead was extracted from home brewing equipment. In thosemeasurements conducted on café brewed espresso only ready to drink coffee was analysed.7

The intake of coffee in Danish men is on average 706 ml /day on average, with a 95 percentile of1,657 ml/day. For Danish women, the corresponding figures are 550 ml/day and 1,371 ml/dayrespectively. Based on this, it was calculated that on average, men have an intake of from coffee of0.74 µg of lead/person/day with a 95 percentile of 1.73 µg/person/day. Women have an averageintake of lead of 0.58 µg /person/day with a 95 percentile of 1.35 µg/person/day. The averageintake is equal to 9.1% of the total lead intake from beverages for men and 7.1% for women. This issubstantially lower than the previous estimate of 40%. The reason for this is, that in the currentstudy the concentration of lead found in coffee was lower compared to previous studies.In the 8 samples of coffee from cafes the lead concentrations was generally at the same level as itwas in home-brewed coffee. A single sample had a concentration 67 ng/ml. In a new sample fromthe same cafe, brewed on the same machine at a later time the measured content of lead wassignificantly lower. This indicates that the high content may be due to a relatively recent descalingof the machine as the use of acid for de-calcification is assumed to release lead from weldings in themachine, if it is present. From a health perspective, the content is not so high that it would beproblematic from a health point of view to drink a cup of coffee every day even if the concentrationof lead remains at 67 ng/ml.The intake of lead from coffee corresponds to 4.2% and 3.3% of the total dietary intake for Danishmen and women, respectively. In general, it can be concluded that the intake of lead from coffee islow compared to the intake from other dietary sources, and that it does not constitute a major partof the total dietary intake lead.8

4. IntroductionLead (Pb) is an element which is found in the environment from both natural and industrialsources. Especially the previous use of tetraethyl lead as an additive in gasoline has been aconsiderable source of lead in the environment. Lead is absorbed in crops from the soil and is thusdirectly transmitted to humans by consumption of these crops or indirectly through consumption ofanimals given feed with a high lead content. Drinking water is another potential source of lead inhumans. Lead in drinking water partly arises from the ground water and partly from migrationfrom pipes, fixtures and appliances which contain lead either in pure form or as part of alloys usedfor weldings. Other sources of lead are work place air, migration from consumer products such astoys, clothes, electronics and food contact materials and from environmental pollution in general.In Denmark and EU lead is, however, heavily regulated in e.g. consumer products, buildingmaterials, drinking water and industrial emissions etc. It is assessed by the National Food Institutein Denmark (DTU FOOD) that food and beverages is the major contributor to the intake of lead inDenmark (A.Petersen et al., 2013).In the Danish population the intake of lead has been declining since lead was phased out as anadditive in almost all gasoline in the late 1970’ies. The latest assessment by DTU FOOD from 2013shows an average intake of lead of 0.25 µg/kg body weight/day with a 95 percentile of 0.46 g/kgbody weight (b.w.)/day (see Table 1) (A.Petersen et al., 2013). Beverages represent about half of thetotal intake of lead (see figure 1). Based on rather old analytical data DTU FOOD has estimated thatapproximately 40% of the lead exposure from beverages is from coffee. Based on these data theintake of lead from coffee thus corresponds to approximately 20% of the total dietary intake of lead.Lead in coffee may originate from the brewing water, extraction from the coffee bean itself or fromlead present in metal alloys in the brewing equipment (e.g. in weldings). A recent, but small studyfrom Germany suggests that lead migration from coffee brewing equipment is usually low, but thatthere may be a significant migration of lead from certain types of new professional espresso brewersor newly decalcified brewers (Bundesinstitut für Risikobewertung, 2014). The german studyincludes only eight differences coffee machines and it is thus not possible to draw generalconclusions based on these results.9

TABLE 1. INTAKE OF LEAD IN THE DANISH POPULATION BASED ON THE DANISH MONITORING PROGRAMME(A.Petersen et al., 2013).µg/kg he most important contributors to dietary exposure to lead in the Danish population are the foodgroups: beverages (47 %), fruit and fruit products (17.4 %) vegetable and vegetable products (9.1 %)and cereals and cereal products (8.4 %) (see figure 1) (A.Petersen et al., 2013).FIGURE 1. SOURCES OF LEAD IN PERCENT OF THE TOTAL INTAKE FROM FOOD IN DENMARK (A.Petersen et al., 2013)Toxicity of leadThe gastro intestinal absorption of lead is estimated to be 15 – 20 % in adults and a little higher inchildren. There are large intra individual differences in the absorption mainly due to differences innutritional status. Deficiency in elements like iron or calcium may increase the absorption of lead.After absorption lead is transported through the blood stream by metal transporters andaccumulates in the bones and to a lesser degree in the liver and the kidney (EFSA Panel onContaminants in the Food Chain (CONTAM), 2010).10

In children effects on the developing central nervous system are considered the critical effect oflead. EFSA has calculated that an intake of 0.50 μg/kg b.w./day may decrease the IQ at an age of 7by 1 %. EFSA concluded that it was not possible to define a lower level of exposure without anyadverse effect, and a sufficient margin of exposure (MOE) could thus not be established (EFSAPanel on Contaminants in the Food Chain (CONTAM), 2010). The MOE is by definition theBenchmark dose (BMDL)/actual exposure. The BMDL is the dose which is calculated to induce agiven magnitude of an effect. In this case the BMDL01 is the dose which may decrease the IQ inchildren by 1 %. If the MOE is 1 the actual exposure may decrease in IQ of 1 % in children. If theMOE is 10 the actual exposure is 10 times below the dose which is expected to decrease the IQ by 1%. There is no common agreement on what MOE should be considered acceptable. If a BMDL01 isbased on human data, a MOE of 10 will usually be sufficient. If the BMDL value is based on animalexperiments a MOE of 100 will usually be sufficient. However, the chosen response (in this case 1%), and the nature of the toxicological endpoint has large influence on which MOE should beconsidered sufficient.For adults EFSA consider effects on systolic blood pressure and kidney effects as the critical effectsof lead. Based on benchmark dose modelling EFSA calculated that dietary lead intake in adults of1.50 μg/kg b.w./day may affect the cardiovascular system by increasing the systolic blood pressureby 1 %. EFSA concluded that a margin of exposure of 10 or greater would be sufficient to ensure thatthere is no appreciable risk of a clinically significant effect on systolic blood pressure. Indeed, evenat MOEs greater than 1.0 the risk would be very low (EFSA Panel on Contaminants in the FoodChain (CONTAM), 2010).Based on animal studies EFSA calculated a BMDL10 for chronic kidney disease of 0.63 μg/kgb.w./day. EFSA concluded that a margin of exposure of 10 or greater would be sufficient to ensurethat there is no appreciable risk of a clinically significant change in the prevalence of chronic kidneydisease. Indeed, overall, the risk at MOEs of greater than 1.0 would be very low (EFSA Panel onContaminants in the Food Chain (CONTAM), 2010).With the use of European data for concentrations of lead in food and European consumption dataEFSA has calculated the MOE for the 3 critical effects in different part of the population. For nonbreast feed children the MOEs are below for 1 even for average consumers. For adults the MOEs fornephrotoxicity are around 1 for the average consumer and below 1 for what EFSA consider highconsumers (95 percentile) (see table 2). These data indicate an urgent need to decrease theexposure to lead from food in the whole population. If coffee is a significant contributor to the totalintake of lead consumer guidance could be a valuable tool in decreasing the exposure.The significance of in utero exposure is not yet been elucidated.11

TABLE 2. ESTIMATED MARGINS OF EXPOSURE (MOES) FOR DIFFERENT ENDPOINTS BY TYPE OF POPULATION. THEMOE RANGE FOR IN UTERO EXPOSURE WAS DERIVED FROM THE 20 TO 40 YEAR OLD FEMALE CONSUMER DATA(FROM (EFSA Panel on Contaminants in the Food Chain (CONTAM), 2010)AverageconsumerMOEHighconsumer(95 ar effects(a)1.2 - 4.20.62 - 2.1Nephrotoxicity(b)0.51 - 1.80.26 - 0.86Cardiovascular effects1.2 - 3.30.67 - 1.90.50 - 1.40.28 - 0.79VegetariansNephrotoxicitySpecific diet (Game meat)Cardiovascular effects0.760.61Nephrotoxicity0.320.262.41.60.79 - 1.90.53 - 1.3infants 3 months breast milkDevelopmental neurotoxicity(c)infants 3 months infant formulaeDevelopmental neurotoxicitychildren 1-3 yearsDevelopmental neurotoxicity0.16 - 0.450.09 - 0.29children 4-7 yearsDevelopmental neurotoxicity0.19 - 0.630.10 - 0.38In utero exposure (a)Developmental neurotoxicity0.39 - 1.30.19 - 0.74MOE: margin of exposure . (a) For cardiovascular effects the MOE was calculated by dividing the BMDL 01 intake value of 1.50μg/kg b.w. /day by the respective dietary exposure estimates taken from Table 29; (b) For nephrotoxicity the MOE wascalculated by dividing the BMDL10 intake value of 0.63 μg/kg b.w. per day by the respective dietary exposure estimates takenfrom Table 29; (c) For neurodevelopmental toxicity, the MOE was calculated by dividing the BMDL01 intake value of 0.50 μg/kgb.w. per day by the respective dietary exposure estimates taken from Table 29. The MOE range for in utero exposure was derivedfrom the 20 to 40 year old female consumer data;Literature studiesBeside the BfR study already mentioned, only 6 relevant studies published within the last 25 yearswith analytical data on the concentration of lead on coffee beans or coffee infusion have beenidentified (see table 3). 12In a Brazilian study 30 samples of coffee beans were measured using ICP-EAS.Concentrations of several hundred ng lead/g were found (C.A.P.Schmidt et al., 2009).Concentrations of metals, including lead were determined in 11 coffee infusions analysedby ICP-MS made from coffees purchased in Bosnia and Herzegovina, Brazil, Lebanon andPoland (Nedzarek et al., 2013).In a study assessing the daily intake of trace elements from foodstuffs in Rio de Janeiro, 3samples of coffee were analysed for lead using ICP-MS (Santos et al., 2004).Lead could not be detected (method AAS: detection limit 100 ng/g) in a study of groundcoffee, decaffeinated coffee, flavoured coffee, pure Arabica coffee or instant coffee(Grembecka et al., 2007).In a study on differentiate geographic growing origins the concentration of lead in coffeewas below the detection limit (500 ng/g) (Anderson and Smith, 2002).A study on nutritional and non-nutritional elements in Brazilian coffee did not detect leadabove the detection level of 450 ng/g using ICP-AES (Santos and Oleiveira, 2001).

TABLE 3. CONCENTRATIONS OF LEAD IN COFFEES IN OTHER SCIENTIFIC STUDIES PUBLISHED AFTER onng/gCommentReferencesCoffeebeans15ICP-EAS470 850Coffee beans grown inbasalt containing soilSmidt et. al,2009Coffeebeans15ICP-EAS350 540Coffee beans grown insandstone containingsoilSmidt et. al,2009Infusion1 g coffeein 27 mlwaterInfusion183 gcoffee in6 L water11ICP-MS858 193Coffee purchased indifferent countriesNedzarek et al20133ICP-MS230 - 308Assessment of traceelement intake in Riode JaneiroSantos et al2004F-AASNDDifferentioation ofmarked coffeeGrembecka etal 2007ICP-AESNDAnderson,2002ICP-AESNDDifferentiation ofgeografphic growingorigins of coffeeDetermination ofnutritional and nonnutritional elements inBrazilian Santos andOleiveira,200113

5. Study design and methodsIntroductionIn this chapter the study design with a description of the sampling, brewing and analytical methodsas well as the quality assurance applied in the project will be presentedCollected coffee samplesCoffee beans: In total 51 samples of coffee beans were purchased in 2014 on the Danish retailmarket and given internal sample codes (no. 1-51). These 51 samples included 29 ground coffeebeans, 15 whole beans and 7 instant coffees. Of the top 10 most used coffee beans in Danishhouseholds 9 of these beans were represented among the 51 samples. Ten of the samples wereorganic produced coffees, whereas rest was conventionally grown. Beans from areas with differentgeoactivity were targeted, including e.g parts of South America and Southeast Asia where relativehigh lead content in the soils was expected and parts of Brazil and Africa where lower lead levelswere expected. See Table 3A in the appendix for specific information on all the coffee beansincluded in this study.Coffee drinks: Six different home brewing methods (M1-M6) were tested for three coffee beanbrands (sample no. 17, 37 and 49, which represented normal (average) lead levels in the range 11-20ng/g).One were within the top 10 most used coffee beans (nr. 49) and duplicate brewing wereperformed for quality control (nr. 37). Different brewing principles were covered, includingfiltration, stepping, boiling and pressure brewing. For the brewing procedures (M1-M5) 65 g groundbeans per litre tap water were used. Furthermore ready-to-drink espresso coffee from 8 cafes andgas stations were purchased in the Copenhagen area. See TABLE 4 for more details on the coffeedrinks included in this study.Brew water: Tap water (Søborg, Denmark) was poured into in a 5 L glass flask to ensure uniformbrewing water for the brewing methods (M1-M5). The tap water had a mean lead (Pb) content of0.62 ng/ml (n 2, 7% relative standard deviation, RSD).TABLE 4. COFFEE DRINKS: DESCRIPTION OF BREWING METHODS, WHICH WERE TESTED FOR 3 SELECTED COFFEEBRANDS (AND SAMPLES FROM CAFES FROM WHICH READY-TO-DRINK ESPRESSOS WERE gM3FrenchpressBoilingM4NorwegianPressureM5Mocha pot6 g beans and 100 ml tap water heated in aluminium mocha poton a ricdrip – newAutomatricdrip – preusedDescription13 g beans pr 200 ml tap water, plastic filter13 g beans pr 200 ml tap water, paper filter200 ml boiling tap water was left in the pot without heating for 30sec. and was then poured over 13 g beans. The stepper waspressed after 3.5 min brewing.13 g beans and 200 ml tap water were boiled 5 min in a pot.

PressureCafé1-8Espresso,caféSamples were purchased from cafes and gas stations in Denmark.All cafes were sampled once (n 1) except cafe 5 which wassampled on two different occasions.Determination of leadThe method described below is accredited according to ISO17025 by the Danish accreditation body,DANAK.Reagents: Water was ultra-purified using a Millipore system (Molsheim, France) and all reagentsused were of per analysis quality or better. Nitric acid (HNO3) (PlasmaPure, SCP Science, France),hydrochloric acid (HCl) (PlasmaCal, SCP Science) and methanol (MeOH) (Rathburn, UK) wereused for sample preparation and analysis. Standard stock solutions of lead (Pb), tin (Sn) andbismuth (Bi) were obtained from SCP Science (Courtaboeuf, France). Sodium hydroxide (NaOH)from (Merck, Darmstadt, Germany), acetic acid (CH3COOH) from (Merck) and HCl were used forpH buffering of water for blind brewing.Coffee beans: Subsamples (homogenized beans 0.4-0.6 g) were digested in high-pressure quartzvessels using microwaves (Multiwave 3000, Anton Paar, Austria) with 2 ml water and 4 mlconcentrated nitric acid. The digests were diluted to a volume of 20 ml with water. Prior to analysissample aliquots were further diluted 5 times with a 2% HNO3, 1% HCl aqueous solution.Certified reference materials: The following certified reference materials (CRMs) were used in thestudy to evaluate the analytical quality.BCR281 Rye grass and BCR62 Olive leaves (Institute of Reference Materials and Measurements,Geel, Belgium) andNBS1572 Citrus leaves (National Institute of Standards and Technology, USA).Subsamples of the CRM were digested following the same procedure for the coffee.Coffee drinks: The coffee solutions were filtrated with single use hydrophilic syringe filters (0.45µm, Minisart, Sartorius, Göttingen, Germany) and prior to analysis aliquots (1 ml) were furtherdiluted to a volume of 6 ml with a 2% MeOH aqueous solution.ICPMS determination: The lead, tin and bismuth content was determined by inductively coupledplasma mass spectrometry (ICP-MS) using an Agilent 8800 ICP-MS instrument (AgilentTechnologies, Waldbronn, Germany) run in no gas mode (m/z Pb208, Sn118 and Bi209 with 0.2 sec.integration time/mass). The ICP-MS was equipped with a Micro mist glass concentric nebuliser(Agilent), and a Scott-type double-pass water-cooled spray chamber (Agilent). Typical plasmaconditions were 1,500 W RF power, 15 L min 1 plasma gas, 0.88 L min 1 carrier gas and 0.32 Lmin 1 makeup gas. Acceptable daily performance of the instrument was ensured by autotuningusing a solution of 1 µg/L each of Li, Mg, Y, Ce

drink coffee, coffee samples based on each of the following six brewing methods were analysed: Filter coffee brewed coffee maker, French press coffee, boiling coffee, coffee brewed in mocha pot, home-brewed espresso and espresso from cafes. The water used f