Development of a multi-residue method for the determination of human and veterinary pharmaceuticals and some of their metabolites in aqueous environmental matrices by SPE-UHPLC–MS/MS
P. Paíga, L.H.M.L.M. Santos, C. Delerue-Matos
∗REQUIMTE/LAQV,InstitutoSuperiordeEngenhariadoPorto,InstitutoPolitécnicodoPorto,RuaDr.AntónioBernardinodeAlmeida,431,4200-072Porto, Portugal
Keywords:Human
andveterinarypharmaceuticals Multi-residue
Aqueousmatrices SPE
UHPLC-ESI–MS/MS
a b st r a c t
Theaimofthepresentworkwastodevelopandvalidateamulti-residuemethodfortheanalysisof 33humanandveterinarypharmaceuticals(non-steroidalanti-inflammatorydrugs(NSAIDs)/analgesics, antibioticsandpsychiatricdrugs),includingsomeoftheirmetabolites,inseveralaqueousenviron- mentalmatrices:drinkingwater,surfacewaterandwastewaters.Themethodisbasedonsolidphase extraction(SPE)followedbyultra-highperformanceliquidchromatography-tandemmassspectrometry (UHPLC–MS/MS)anditwasvalidatedfordifferentaqueousmatrices,namelybottledwater,tapwater, seawater,riverwaterandwastewaters,showingrecoveriesbetween50%and112%forthemajorityof thetargetanalytes.
Thedevelopedanalyticalmethodologyallowedmethoddetectionlimitsinthelownanogramsperliter level.Methodintra-andinter-dayprecisionwasunder8%and11%,respectively,expressedasrelative standarddeviation.Thedevelopedmethodwasappliedtotheanalysisofdrinkingwater(bottledandtap water),surfacewaters(seawaterandriverwater)andwastewaters(wastewatertreatmentplant(WWTP) influentandeffluent).Duetotheselectivityandsensitivityoftheoptimizedmethod,itwaspossible todetectpharmaceuticalsinalltheaqueousenvironmentalmatricesconsidered,includinginbottled wateratconcentrationsupto31ngL−1 (salicylicacid).Ingeneral,non-steroidalanti-inflammatory drugs/analgesicswasthetherapeuticgroupmostfrequentlydetected,withthehighestconcentrations foundinwastewaters(acetaminophenandthemetabolitecarboxyibuprofenatlevelsupto615and 120gL−1,respectively).
1. Introduction
Theconsumption of pharmaceuticals mayvary considerably fromcountrytocountry[1],andnoinformationforthetotaluse ofpharmaceuticalsisavailable.Theycanbesoldasprescription orover-the-countermedicines[2]andpharmaceuticals havean importantroleinthetreatmentandpreventionofdiseasesinboth humanandveterinarymedicine[3].Additionally,theycanbealso usedasgrowthpromotersinanimals[4].
Pharmaceuticalsarecontinuallyreleasedintotheenvironment mainlyasaresultoftheirexcretioninurineand/orfeces,manufac- turingprocesses,anddisposalofunusedorexpiredproducts[5].
Besidesthat,pharmaceuticalscanalsodirectlyentertheterrestrial
∗Correspondingauthor.
E-mailaddress:[email protected](C.Delerue-Matos).
environmentviadifferenttypesofmanure,slurriesorothertypes ofbiosolids[4]and,afterthat,reachtheaquaticenvironmentdue tosoilsrun-off.
The pharmaceutical market has been growing over the last decadesaswellastheknowledgeontheenvironmentalimpact ofpharmaceuticalstoecosystems[6].Oncereleasedintotheenvi- ronment,pharmaceuticalsandtheirbioactivemetabolitescanbe transported and distributed towater, soil orsediments, due to differentfactors, suchasthephysicochemical properties of the compoundsandthecharacteristicsofthereceivingenvironment.
Pharmaceuticalscanaccumulateinbiotaandinduceadverseeffects inaquaticorterrestrialorganismsaswell[5].
Wastewaters generated by hospital, industrial and domestic sourcesarepointedoutasoneofthemostimportantsourceofphar- maceuticalstotheaquaticenvironment[7].Manypharmaceuticals, theirmetabolitesand transformationproductsareincompletely removedin conventionalwastewater treatmentplants(WWTP)
and thus discharged into the environment [8]. Therefore, the presence of pharmaceuticals at tracelevels (nanogramsto low microgramsperliter)hasbeenreportedinwastewaters,surface waters, groundwater’s and, toa lesser extent, drinking waters [9–13].Inthisway,sensitiveandreliableanalyticalmethodsmust bedevelopedforthedetectionandquantificationofpharmaceuti- calsintheaquaticenvironment.
Samplepreparationisconsideredtobeafundamentalstepin environmentalanalyticalprocedures[14]andhastobeselective, cheap,quick,andenvironmentallyfriendly[15].Theuseofsolid phaseextraction(SPE)forthesamplepreparationhasincreased overthelast years, becauseitis easytooperate,hasincreased selectivitywithmanynewsorbents,andhasthepossibilitytointer- faceforautomationandrobotics[14].Itsversatility(purification, traceenrichment,desalting,derivatization,andclassfractionation) allowsSPEtobethefirstchoicetobeusedinsamplepreparation ofliquidsamples.
Tandem massspectrometry (MS) isconsidered tobea high resolutionandsensitivedetectiontechniqueduetoitsspecificity andlow limits ofdetection [16]. Eachyear,LC–MS/MSbecome evenmoresensitiveandisanimportantdetectionmethodinenvi- ronmentalanalysis[9].Numerousstudieshavedemonstratedits distinctadvantagesfortraceanalysisofpharmaceuticalsinenvi- ronmentalsamples[9,10,12].
Thedevelopmentofaccuratemulti-residueanalyticalmethod- ologiesforthesimultaneousanalysesoftracelevelsofhumanand veterinarypharmaceuticals in awide range ofaquatic environ- mentalmatricesisusefulandnecessary,inordertobepossible togatherdataondifferentways ofentranceofpharmaceuticals intotheenvironmentand,atthesametime,evaluatetheimpact anddistributionofeitherhumanandveterinarymedicinesinthe aquaticenvironment.Inthepresentstudy,anautomatedoff-line SPEprocedurefollowedbyUHPLCcoupledtotriplequadrupoletan- demMSwithelectrosprayionizationsource(ESI)wasdeveloped forthedeterminationof33humanandveterinarypharmaceuti- calsandsomeoftheirmainmetabolites,includingnon-steroidal anti-inflammatorydrugs(NSAIDs)/analgesics,antibiotics,andpsy- chiatricdrugs.Differentchromatographicandmassspectrometry parameters were optimized. A versatile SPE protocol was also developedin ordertobeappliedtoa greatdiversity ofaquatic environmentalmatrices(e.g.wastewater,surfacewater,drinking water).Finally,theoptimizedmethodwassuccessfullyappliedto sixteensamplesembracingdrinkingwater(bottledandtapwater), surfacewater(seawaterandriverwater)andwastewater(WWTP influentandeffluent),allowingtheevaluationofthedistribution oftheselectedhumanandveterinarypharmaceuticalsthroughthe aquaticenvironment.
2. Materialsandmethods
2.1. Reagents,solventsandmaterials
Allpharmaceuticalsandisotopicallylabelledstandardswereof highpuritygrade(≥98%)andtheirphysicochemicalcharacteristics andsuppliersarelistedinTableSM1ofSupplementaryMaterial.
Stockstandard solutions (ataconcentrationof 1000mgL−1) werepreparedonaweightbasisin acetonitrile,withexception of naproxen and diclofenac that were prepared in acetoni- trile:methanol(50:50,v/v),sincethesesubstancesareveryslightly solubleinpureacetonitrileand freelysolubleinmethanol[17];
andpsychiatricdrugsthatwerepreparedinmethanol.Theantibi- otics ofloxacin, ciprofloxacin, and enrofloxacin as well as the isotopicallylabelled standardciprofloxacin-d8werepreparedin methanoladdingNaOH1MasdescribedbyIbá ˜nez,M.etal.[18].
Carbamazepine-d10 (100mgL−1), venlafaxine-d6 (100mgL−1),
diazepam-d5(1000mgL−1),andnorsertraline(100mgL−1)were purchasedasfreebaseinmethanol.
Allstockstandardsolutionswerestoredat−20◦Candrenewed every3months,withtheexceptionofantibioticsthatwerepre- paredmonthlybecauseoftheirlimitedstability.Workingstandard solutionscontainingallpharmaceuticalswerepreparedinacetoni- trile:ultrapurewater(30:70,v/v)bymixingappropriateamountsof eachstockstandardsolution.Thesesolutionswerepreparedbefore eachanalyticalrun.
Amixturewithallisotopicallylabelledstandardswasprepared tobeusedforinternalstandardcalibration.
Acetonitrile LC–MS grade was supplied by Biosolve (Valkenswaard, Netherland) and methanol LC–MS Ultra CHROMASOLV®waspurchasedfromSigma-Aldrich.Hydrochloric acid37%wasobtainedfromCarloErba(Rodano,Italy),andformic acid98% PA-ACSandethylenediaminetetraacetic aciddisodium salt 2-hydrate (Na2EDTA) (assay 99.9–101.0%) were obtained from Panreac (Barcelona, Spain). Ultrapure water (resistivity of 18.2Mcm) was produced using a Simplicity 185 system (Millipore,Molsheim,France).
Allchromatographicsolventswerefilteredthrougha0.22m nylonmembranefilter(FioroniFilters,Ingré,France)usingavac- uumpump(DinkoD-95,Barcelona,Spain)anddegassedfor15min in an ultrasonic bath (Sonorex Digital 10P, Bandelin DK 255P, Germany). Before the UHPLC–MS/MS analysis, sample extracts werefilteredthrough0.22mPTFEsyringefilters(Specanalitica, Carcavelos,Portugal).SPEcartridgesStrata-X(200mg,3mL)from Phenomenex(California,USA)andOasisMCX(150mg,6mL)from Waters(Milford,Massachusets,USA)wereusedinthestudyofthe SPEsorbent.
2.2. Samplingsitesandsamplecollection
Sixteensampleswereanalyzedusingtheoptimizedanalytical methodology.Twosampleswerecollectedforeachtypeofwater withtheexceptionofseawaterforwhichsixsampleswerecol- lected.Tapwatersampleswereobtainedfromthetapoflaboratory andin aprivatehouselocatedin thePortoarea, whilstbottled waterswerepurchased inthelocalmarket.Riversampleswere takenfromLisriver,whichcrossesthecityofLeiriainthecenter regionofPortugal.SeawatersamplesfromtheAtlanticOceanwere collectedinPortocoastalareafrombeacheswithdifferentbathing waterquality(excellent,goodand sufficient)[19].Twobeaches foreachclassificationwereselected.Wastewatersamples(influent andeffluent)werecollectedintwoWWTPslocatedinthecenter regionofPortugal(Leiria,Portugal).WWTPinfluentandeffluent were24hcompositesamples,whereastheothertypesofwaters weregrabsamples.Sampleswerecollectedin2015.
Amber glass bottles pre-rinsed with ultrapure water were used for sample collection. After reception in the laboratory, wastewaters and river water were filtered through 0.45m nylonmembrane filters (Fioroni Filters, Ingré, France) followed by0.22mnylonmembranefilters(FioroniFilters,Ingré,France), whiletapwaterandseawaterwereonlyfilteredthrough0.22m nylonmembranefilters(FioroniFilters,Ingré,France).
2.3. Analyticalmethod
SPEcartridges wereconditionedwith5mLof methanolfol- lowed by5mLof ultrapurewater and 5mL ofultrapure water atpH2using avacuumsystemmanifold(Chromabond,Düren, Germany). Asuitable volume ofa 0.1MNa2EDTAsolution was addedtothewatersamplestoachieveafinalconcentrationof0.1%
(gsolute/gsolution).250mLoftapwater,bottledwater,seawater, andriverwater,100mLofWWTPeffluentor50mLofWWTPinflu-
Fig.1.MRMchromatogramofastandardmixtureat100gL−1oftheselectedpharmaceuticalsanalyzedinthenegative(i)andpositive(ii)ionizationmode.
entwithpHadjustedto2withconcentratedHClwerepercolated throughthecartridge.Then,thecartridgeswererinsedwith5mL ofultrapurewateranddriedundervacuumfor60mintoremove theexcessofwater.Finally,theelutionwasperformedwith10mL ofmethanol.Extractswereevaporatedunderagentlestreamof nitrogenand reconstitutedwith500Lofacetonitrile:ultrapure water(30:70,v/v).Lastly,5Lofamixtureofisotopicallylabelled standards was added in order to obtain a final concentration of7.5gL−1forsalicylicacid-d4,150gL−1foracetaminophen- d4,75gL−1 foribuprofen-d3,10gL−1forcarbamazepine-d10 andforfluoxetine-d5,20gL−1forvenlafaxine-d6,40gL−1 for diazepam-d5,100gL−1 forciprofloxacin-d8and azithromycin- d3,and50gL−1forsulfamethoxazole-d4.
QuantificationofpharmaceuticalswasperformedinaNexera Ultra-High Performance Liquid Chromatography system (Shi- madzu Corporation, Kyoto, Japan) equipped with two solvent deliverymodules, a degasser, an autosampler, a column oven, and coupled to a triple-quadupole mass spectrometer detec- tor LCMS-8030 with an electrospray ionization source (ESI).
NSAIDs/analgesics were analysed in negative ionization mode, whileantibioticsandpsychiatricdrugsweredeterminedinthepos- itiveionizationmode.Differentchromatographicconditionswere usedforeachionizationmode.NSAIDs/analgesicswereanalysed using previously optimized conditions [20]. Briefly, chromato- graphic separation was achieved using a Kinetex C18 column (150×2.1mmi.d.,1.7mparticlesize)fromPhenomenex(Cali- fornia,USA),usingultrapurewateraseluentAandacetonitrileas eluentBataflowrateof0.22mL/min.Thegradientelutionwas:
0–1.0min,30%–35.6%B;1.0–2.0min,35.6%–100%B;2.0–6.0min, 100%B;6.0–6.5min,returntoinitialconditions;6.5–10.5min,re- equilibrationofthecolumn.
Forpositiveionizationmode,thechromatographicconditions wereoptimizedandchromatographicseparationwascarriedout inaCortecsTMUPLC®C18+column(100×2.1mmi.d.;1.6mpar- ticlesize)fromWaters(Milford,Massachusets,USA).EluentAwas 0.1%formicacidinultrapurewaterandeluentBwasacetonitrile ataflowrateof0.3mLmin−1.Thegradientelutionstartedwith5%
ofeluentB,increasingto100%Bin3min,maintained100%Bdur- ing0.5minandthen,returnedtoinitialconditionswithin0.5min.
Thecolumnwasre-equilibratedfor3minbeforethenextinjection.
Inboth modes,theautosamplerwasoperatedat4◦C,aninjec- tionvolumeof5Lwasusedandcolumntemperaturewaskeptat 30◦C.LabSolutionsLC–MSsoftware(ShimadzuCorporation,Kyoto, Japan)wasusedforsystemcontrolanddataprocessing.
Sourcedependentparameters wereoptimized by the direct injectionofastandardmixturesolution10mgL−1forpositiveion- izationmode,whereasfornegativemodetheMS/MSconditionsare describedinapreviouswork[20]andarethefollowing:desolvation linetemperature(DLT)wassetat250◦Candheatblocktempera- ture(HBT)at300◦C;interfacevoltage(IV)at5.0kV;nebulizinggas (NGF)anddryinggas(DGF)ataflowrateof2.6and12.5Lmin−1, respectively.Forpositiveionizationmode,theoptimizedcondi- tions(seesubsection 3.1.1) wereas follows: DLT: 300◦C; HBT:
425◦C;IV:5.0kV;NGF:2.6Lmin−1andDGF:15.0Lmin−1.Inboth ionizationmodes,nitrogenwasusedasnebulizinganddryinggas, andargonwasusedascollisioninduceddissociationgasatapres- sureof230kPa.Adwelltimeof25msandof15mswasusedin negativeandpositiveionization,respectively.
Multiplereactionmonitoring(MRM)isamethodusedinMS/MS inwhichanionofaparticularmassisselectedinthefirststageof atandemmassspectrometerandanionproductofafragmenta- tionreactionoftheprecursorionisselectedinthesecondmass spectrometerstage,formedbycollision-induceddissociation,for detection.Thesignalrepresentstheprecursor-to-productiontran- sitionfor a specificion pair.Themostintense productwasset asquantifierion whereasthesecondmostintensewasusedas
qualifierion.Detaileddataontheoptimizedmassspectrometry parameters(precursorions,quantifierandqualifierions,andion ratio)aswellasthecorrespondinginternalstandardusedforquan- tificationpurposesisgiveninTableSM2(Supplementarymaterial).
2.4. Methodvalidation
The performance of the methodwas evaluated throughthe estimationofthelinearity,extractionrecoveries,methoddetec- tionlimits(MDLs),methodquantificationlimits(MQLs),precision asrepeatabilityexpressedasrelativestandarddeviation(%RSD), reproducibilityandmatrixeffects(ME)foreachtypeofwatersam- ple.
Linearitywasestablishedbysettingcalibrationcurves(solvent andmatrixmatched)usinglinearregressionanalysiswithconcen- trationsintherangeof5–250gL−1.Quantificationoftheanalytes wasperformedbytheinternalstandardapproach.Recoverieswere determinedbycomparingconcentrationsobtainedaftertheSPE procedure, calculated byinternal standard calibration,withthe initialfortificationlevels.Sincewatersamplescancontaintarget compounds,blanks(non-spikedsamples)werealsoanalyzedand thelevelsfoundsubtractedfromthoseobtainedforspikedsamples.
MDLsandMQLsweredeterminedastheminimumamountofana- lytedetectablewithasignal-to-noiseratioof3and10,respectively.
Thelimitsweredeterminedusingrealsampleswheneverpossible, otherwisespikedsampleswereused.Methodprecisionwasdeter- minedbyintra-andinter-day analysis(%RSD).Threestandards mixturescontainingalltheanalytesatafinalconcentrationof25, 50,and100gL−1wereusedandsixsuccessiveinjectionsinone dayandsixconsecutivedays(triplicateinjections)wasperformed, respectively.ToassesstheMEinalltypesofwaters,theslopeof thematrixmatchedcalibrationcurvewascomparedwiththeslope ofthecalibrationcurvepreparedinsolvent(acetonitrile:ultrapure water(30:70,v/v)).Ablank(samplewithoutadditionofstandards) wassimultaneouslyassayedinordertosubtracttheconcentration ofthetargetanalytespresentinthesample.MEwascalculated accordingtoEq.(1)[9,21],respectively.Avalueofzeroindicates that there is noME, while for a positive value there is an ion enhancementsignalandforanegativevalueanionsuppression signal.
Signalsupression(%)=
slopematrix-matched
slopesolvent −1
×100 (1)
3. Resultsanddiscussion
3.1. UHPLC–MS/MS 3.1.1. Ionsourceparameters
Source-dependentparameterssuchasDLT,HBT,IV,NGF,and DGFarekeyparameterstoenrichtheinstrumentalsensitivity[22]
and,thereforetheywerestudied.ForNSAIDs/analgesicspharma- ceuticalsthesource-dependentparameterswereusedasdescribed inPaíga,P.etal.[20],whileforpharmaceuticalsanalyzedinpos- itiveionizationmodetheparameterswereoptimized.Inthecase ofDLT,temperaturesfrom200to300◦C weretestedanditwas observedanincreaseofthesignalfrom220to225◦C,beingcon- stantthereforeforpsychiatricdrugs.However,forantibioticsthe signalincreasedwiththeincrementofthetemperature.Thus,a temperatureof300◦Cwasselected,sinceantibioticsshowedlower sensitivitywhencomparedwithpsychiatricdrugs.ForHBT,tem- peraturesbetween200and500◦Cwerestudiedand,ingeneral, highsensitivitywasobservedforallthecompoundsusingatem- peratureof425◦C.TheinfluenceofIVintheanalyticalsignalwas evaluatedbetween0and5kVanditwasnoticedanincreaseinthe signalwithanincrementinIV.Amaximumsignalwasachieved
for5kV,whichwasselected.NGFwasstudiedbetween0.5and 3.0Lmin−1anditwasobservedanincreaseoftheanalyticalsignal from0.5to2.6Lmin−1andafterthatadecrease.Thehighestsig- nalintensitywasobtainedusing2.6Lmin−1.Lastly,DGFwasvaried between10and20Lmin−1andamaximumresponsewasobtained foraflowrateof15Lmin−1.
3.1.2. Dwelltimeoptimization
UsingMRMforquantificationQ1andQ3aresettofilterspecific ionsastheypassthroughtheionoptics.Itisthedwellandthepause timestogetherwiththenumberof MRMtransitions monitored thatdeterminethecycletimeandthusthenumberofdatapoints obtainedacrossagivenpeak[23].Whileveryshortdwelltimes canbeusedforextendedcompoundscreening,higherdwelltimes aredesirableforbetterSignal-to-Noiseratio.Thus,settingproper dwelltimeisnecessarytoachievethebestqualityandquantityof data.
Differentdwelltimeswerestudiedintheintervalof1.0–100ms.
Forantibioticsandpsychiatricdrugs,itwasobservedaconstant analyticalsignalandagoodreproducibility(RSD<5%)intherange ofdwelltimebetween1.0(RSD=4.54%)and15ms(RSD=3.09%).
Afterthat,anincreaseofthedwelltimeledtoadecreaseofthe analyticalsignalwithRSDvaluesreaching28%.Therefore,thebest compromisebetweentheanalyticalsignalandthereproducibility wasachievedusing15msforantibioticsandpsychiatricdrugs.
3.1.3. Chromatographicseparation
Thirty-three human and veterinary pharmaceuticals (NSAIDs/analgesics, antibiotics, and psychiatric drugs) were studied in the present work. For NSAIDs/analgesics (negative ionization mode), chromatographic analysis was performed as describedinPaíga,P.etal.[20].Fortheremainingpharmaceuticals (positiveionizationmode),thechromatographicconditionswere optimized by evaluating different chromatographic columns, mobilephases,modesofelution,flowratesandcolumntemper- atures.Thecomplexityofsomecompoundsmaycauseproblems duringLCanalysis,whichoftenresultsinbroadortailingpeaks, poorresolutionorsomeshapeissuesduetoanincreasedretention [24].Itwasobservedthatforpsychiatricdrugs,25%ofacetonitrile wastheinitialmaximumamountoforganicsolventallowedfora goodpeakresolution,otherwisefrontingandtailingpeakswere obtained.Intheotherhand,forantibioticsatleast50%oforganic solvent(acetonitrile)intheinitialconditionswasneeded.Itshould be noted that even with that initialproportion of eluents,the tailingandfrontingisnotobservedbutlargepeakswereachieved.
Theperformanceoftwochromatographiccolumnswaseval- uated (Fig. SM1, Supplementary material). Using Kinetex C18 column,thebestpeakresolutionwasachievedwith10%ofACN intheinitialconditionsofthegradientelution.Peakswelldefined wereobtainedforpsychiatric drugs,howeveralmost allantibi- oticsshoweda badpeak resolution(Fig.SM1-A,Supplementary material).ReplacingthestationaryphasebythecolumnCortecsTM UPLC®C18+(100×2.1mm,1.6m)andusingthebestchromato- graphicconditionsachievedforKinetexC18column,ingeneral, amuchbetterpeakresolutionwasobservedforallpharmaceu- ticals.Therefore,thecolumnCortecsTMUPLC® C18+wasusedfor thechromatographicanalysisinthepositiveionizationmode.Dif- ferentmobilephasescomprisingseveralcombinationsofaqueous andorganiceluentsweretestedinordertoprovideabetterpeak resolutionandagoodsensitivityaswell.Bothelutionmodes(iso- craticandgradient)weretested.Thebestperformanceforallthe pharmaceuticals was obtained using 0.1% formic acidin ultra- purewater (A) and acetonitrile (B)in themobile phase and a gradientelutionasfollows,initialconditions:5%B;0.0–3.0min, 5–100%B;3.0–3.5min,100%B;3.5–4.0min,returntoinitialcon- ditions;4.0-7.0min,re-equilibrationofthecolumn.Theflowrate
wasoptimizedto0.3mLmin−1.Arepresentativechromatogramof a100gL−1standardmixtureoftheselectedpharmaceuticalsis presentinFig.1.
3.2. Solidphaseextractionoptimization
3.2.1. OptimizationoftheSPEsorbent,samplepHadjustment andadditionofNa2EDTA
TheperformanceofthepolymericsorbentStrata-Xandofthe mixedmodepolymericandcationexchangesorbentOasisMCXon theextractionoftheselectedhumanandveterinarypharmaceuti- calswasevaluated.Theeffectofsample’spHintherecoveriesof theselectedpharmaceuticalswasassessedforpH2,9,andwithout pHadjustment.ConcentratedHClorNaOHwasusedtoadjustsam- ple’spH.Differentsamplevolumes(25,50,100,250,and500mL) werealsostudied.Extractionefficiencyofcertainpharmaceuticals, likeantibiotics,canbeimprovedbyaddingNa2EDTAtothesam- ples,becausesolublemetalsboundtothechelatingagent,releasing theanalyteandincreasingtheextractionefficiency[25].Thus,the effectofaddingNa2EDTA,priortoextraction,intherecoveryofthe selectedpharmaceuticalswasalsoevaluated.Reconstitutionsol- ventusingdifferentorganicsolvents(acetonitrileandmethanol) andmixturesofsolvents(methanol:ultrapurewater(1:1,v/v),ace- tonitrile:ultrapurewater(1:1,v/v),methanol:ultrapurewater(3:7, v/v), acetonitrile:ultrapure water (3:7, v/v), methanol:ultrapure waterwith0.1%formicacid(3:7,v/v),andacetonitrile:ultrapure waterwith0.1%formic acid(3:7,v/v))wasalsostudied.Allthe optimizations wereperformedusing 250mLof ultrapurewater andtheextractswerereconstitutedwith1mLofacetonitrile.The obtainedrecoveriesforthedifferenttestedconditionsarepresent intheradarchartsshowninFig.2.
Using Strata-X, recoveries increased for carboxyibuprofen, ibuprofen,salicylicacid,clarithromycin,azithromycin,norsertra- line, andsertralineatsamplepH2,acetylsalicylicacid, salicylic acid, trimethoprim,citalopram,andvenlafaxinewithoutsample pHadjustment,and trimethoprim, sulfadimethoxine,norfluoxe- tine and venlafaxine at sample pH 9. A decrease of recoveries wasobtainedfor sulfapyridine, norfluoxetine,and trazodoneat samplepH2,hydroxyibuprofen,sulfadimethoxine,azithromycin, and 10,11-epoxycarbamazepinewithoutsamplepHadjustment, anddiclofenac,ibuprofen,naproxen,clarithromycin,azithromycin, norsertraline,sertraline,paroxetine,andtrazodoneatsamplepH 9,when Na2EDTAwasadded.Recoveries wereconstantfor the remainingpharmaceuticals.
On the other hand, for Oasis MCX it was observed that, in all pH range, low recoveries were obtained for almost all NSAIDs/analgesicsandantibioticswhenNa2EDTAwasaddedtothe sample,withtheexceptionofacetaminophenthathadthehighest recovery(between81.6to96.3%),whereascarboxyibuprofenwas notdetected(Fig.2).Inthecaseofpsychiatricdrugs,thebehaviour observedwasopposite andhighrecoverieswereobtainedwith the addition of Na2EDTA. Low recoveries were only obtained fornorfluoxetineatsamplepH2and10,11-epoxycarbamazepine withoutsamplepHadjustment.Forciprofloxacin,enrofloxacin,and ofloxacin,lowrecoveries(<10%)wereachievedusingbothtypesof sorbentsinalltestedconditions.Fortrimethoprim,alowerrecov- ery(inallpHrange)wasachieved(<30%)usingOasisMCXwithout Na2EDTAadditionthanwhenthechelatingagentwasadded.Using Strata-X,trimethoprimshowedagoodrecoveryinallexperiments.
ForallthestudiedpH’sthesamebehaviourwasobtainedforsul- fonamideantibioticsandthehighrecoverieswereachievedusing Strata-X. Afterthe additionofNa2EDTA,sulfonamides’recover- iesremainedconstantusingStrata-X,butdecreasedwhenOasis MCXwasused.Carbamazepine,citalopram,andvenlafaxinehad constantrecoveriesinallstudiedconditionsandnoinfluenceof
Fig.2. Recoveriesobtainedinthestudyofsolidphaseextractionsorbent,samplepHadjustment(pH2,withoutpHadjustment,andsamplepH9),andadditionornotaddition ofNa2EDTA(NSAIDs/analgesics:1-Acetaminophen,2-Acetylsalicylicacid,3-Carboxyibuprofen,4-Diclofenac,5-Hydroxyibuprofen,6-Ibuprofen,7-Naproxen,8-Nime- sulide,9-Ketoprofen,10-Salicylicacid;Antibiotics:11-Ciprofloxacin,12-Enrofloxacin,13-Trimethoprim,14-Sulfamethoxypiridine,15-Sulfapyridine,16-Sulfamethazine, 17-Sulfadimethoxine,18-Sulfadiazine,19-Sulfamethoxazole,20-Ofloxacin,21-Clarithromycin,22-Azithromycin;Psychiatricdrugs:23-Norsertraline,24-Norfluoxetine, 25-Carbamazepine,26-Fluoxetine,27-Sertraline,28-Citalopram,29-Venlafaxine,30-Paroxetine,31-Trazodone,32-Diazepam,33-10,11-Epoxycarbamazepine).
sorbent,sample’spHoradditionofNa2EDTAwasverifiedinthe recoveryresults.Inthecaseofparoxetine,forbothtypesofSPE sorbentsatpH2,nodifferencewasobservedbytheadditionof Na2EDTA.Forsertralineandtrazodonerecoverieswerehighusing OasisMCXwiththeadditionofNa2EDTA.
AmarkeddecreaseintherecoveryofNSAIDs/analgesicsand antibiotics usingOasis MCX was observed when Na2EDTA was added(Fig.2).Althoughseveralstudiesmentionedtheimportance ofaddingacationcomplexingagent(Na2EDTA)tochelatemetals andtominimizeinterferencesforsomepharmaceuticals[9,26],our resultsareinagreementwithliterature[27,28],whereadecrease ofrecoveriesforsomepharmaceuticalswasalsoobservedwhen Na2EDTAwasadded.Apossiblejustificationforthisisthatwhen theNa2EDTAispresentinexcess,itchelatesnotonlymetalsbut alsoorganiccompounds[27].
Comparingthe performance of both sorbents could becon- cludedthatStrata-Xallowedhigherrecoveriesforthemajorityof theanalysedpharmaceuticals.Thiscouldbeduetoalowreten- tionoftargetcompoundsinOasisMCX,sincethismixedsorbent wasaloweramountofreversed-phasesorbentcomparativelyto Strata-X.
For almost all pharmaceuticals, the highest recoveries were obtainedwhen samplepHwasadjustedto2.Theseresultscan beexplainedbythepresenceof acidicfunctionalgroups inthe molecularstructureofmanypharmaceuticals,thereforelowering pHundertheirpKavaluesenhancesthepresenceofneutralforms andtheirinteractionwiththereversed-phasesorbent.
Despiteofthegoodrecoveriesfoundforsomepharmaceuti- calsusingOasisMCX,forthemajorityoftheselectedcompounds thebestrecoverieswereachievedusingStrata-Xcartridges,with
sample pH adjusted to 2 and adding Na2EDTA to the sam- ple.The recoveriesof NSAIDs/analgesicsranged from72.5% for acetylsalicylic acidto 111.5% for nimesulide with exception of acetaminophen(recoveryof27.8%).Formostofpharmaceuticals, ourrecoverieswereinaccordancewithresultsobtainedbyGros, M.et al.[9]andWeigel,S. etal.[29]usingpolymericsorbents.
A similarrecoveryfor acetaminophenwas obtainedbyWeigel, S.etal.[29]andthelowrecoveryobtainedmightbejustifiedby itsreadywatersolubilitylimitingtheretentionofacetaminophen.
Forantibiotics,recoverieswerebetween67.2%forclarithromycin and97.8%fortrimethoprim,withtheexceptionofciprofloxacin, enrofloxacin,andofloxacin,withrecoverieslowerthan11%.Low recoveries(about 30%)forciprofloxacinandofloxacinwerealso obtainedbyS.Castiglionietal.[30],whichpointedoutthestepof evaporationtodrynessasthereasonforthelowrecoveriesforthese antibiotics.Forpsychiatricdrugs,recoveriesrangedfrom52.7%to 111.5%fortrazodoneanddiazepam,respectively.
3.2.2. ReconstitutionsolventafterSPEextraction
None of the studied conditions previously described in subsection3.2.1.,allowedobtaininggoodrecoveriesforthefluo- roquinoloneantibioticsciprofloxacin,enrofloxacin,andofloxacin.
Fluoroquinolones have an amphoteric behaviour (pKa=5 and pKa=8–9)andtheirsolubilitydependofthepH[31].Thus,differ- entreconstitutionsolventsweretested,embracing100%oforganic solventsormixturesofsolvents.Theobtainedresultsarepresent inFig.SM2(Supplementarymaterial).Using100%ofacetonitrileor 100%ofmethanollowrecoverieswereachieved,namely:1.8,17.1, and13.5%for100%ofacetonitrileand11.9,53.2,and50.1%for100%
methanol,for ciprofloxacin, enrofloxacin, andofloxacin, respec-
Fig.3. Samplevolumebreakthroughin(A)riverwater;(B)WWTPeffluent,and(C)WWTPinfluent(1-Acetaminophen,2-Acetylsalicylicacid,3-Carboxyibuprofen,4- Diclofenac,5-Hydroxyibuprofen,6-Ibuprofen,7-Naproxen,8-Nimesulide,9-Ketoprofen,10-Salicylicacid,11-Ciprofloxacin,12-Enrofloxacin,13-Trimethoprim,14-Sul- famethoxypiridine,15-Sulfapyridine,16-Sulfamethazine,17-Sulfadimethoxine,18-Sulfadiazine,19-Sulfamethoxazole,20-Ofloxacin,21-Clarithromycin,22-Azithromycin, 23-Norsertraline,24-Norfluoxetine,25-Carbamazepine,26-Fluoxetine,27-Sertraline,28-Citalopram,29-Venlafaxine,30-Paroxetine,31-Trazodone,32-Diazepam,33- 10,11-Epoxycarbamazepine).
tively.Therefore,mixturesoforganicsolventswithultrapurewater atdifferentproportions(1:1and3:7(v/v))weretested.Themix- tureofultrapurewaterandacetonitrileimprovedtherecoveriesof allfluoroquinolones(>80%)andrecoveriesremainedconstantsfor bothtestedproportions(1:1and3:7(v/v)).Whenultrapurewater wasmixedwithmethanolintheproportion(1:1,v/v)recoveries wereonlyhigherforciprofloxacin,while fortheproportion3:7 (v/v),recoveriesincreasedforallfluoroquinoloneswith49.1,58.2,
and 65.4%for ciprofloxacin,enrofloxacin, and ofloxacin,respec- tively.
Sincethechromatographicseparationofantibioticsandpsychi- atricdrugsembracingtheuseof0.1%offormicacidintheaqueous eluent, it isimportant tocheck therecoverieswiththeacidifi- cationof theaqueousphasein thereconstitutionsolvent.Thus, mixturesof0.1%offormicacidinultrapurewaterwithacetoni- trileormethanolwerealsotested,usingtheproportion3:7(v/v), duetothehighrecoveriesobtainedwiththisproportion.Nosignifi-
Fig.4. Matrixeffectfortheselectedpharmaceuticalsinthedifferentaqueousenvironmentalmatrices.
cantimprovementswerenoticedwiththeacidificationofultrapure waterandrecoveriesbetween85and102%andfrom52to61%
wereobtainedusingacetonitrileandmethanolasorganicsolvent, respectively.Therefore,themixtureofacetonitrileandultrapure waterintheproportionof3:7(v/v)wasselectedasreconstitution solvent.
3.2.3. Samplevolumebreakthrough
Thestudyofthebreakthroughvolumewasperformedusing surfacewater(riverwater)andwastewaters(WWTPinfluentand effluent).Volumesbetween25and500mL(riverwater)andfrom 25to250mL(wastewaters)weretested.Resultsofrecoveriesare showninFig.3.
Themeanrecoveriesobtainedwere:82.9(25mL),78.7(50mL), 81.3(100mL),79.0(250mL),and68.2%(500mL)forriverwater;
91.0(25mL),81.0(50mL),76.6(100mL),and69.0%(250mL)for WWTPeffluent;and76.8(25mL),75.1(50mL),53.7(100mL),and 44.0%(250mL)forWWTPinfluent,withRSDvalueslowerthan10%.
Therecoveryofacetaminophendecreasedwiththeincreaseof thesamplevolume,whichisinagreementwithpreviousworks [20,32].Acetaminophenshowedthesamebehaviourinalltypesof watertested.Evenwitharecoverylowerthan50%,acetaminophen isdetectedinenvironmentduetoitshighconsumption.Therefore 250,100,and 50mLwerethesamplevolumesselectedforsur- faceswaters,WWTPeffluentandWWTPinfluent,andrecoveries between30.3%(acetaminophen)and101%(ibuprofen,naproxen, and nimesulide)for river water, 28.4(acetaminophen)to 116%
(acetylsalicylicacid)forWWTPeffluent,and32.6(acetaminophen) to122%(acetylsalicylicacid)forWWTPinfluentwereachieved.
3.3. Methodperformance
Linearity,MDLs,MQLs,repeatability,reproducibility,recoveries andMEwereperformed.Thelinearityofthemethodwasestab- lishedbysettingcalibrationcurves(solventandmatrixmatched) usinglinearregressionanalysiswithconcentrationsintherangeof 5–250gL−1.Allpharmaceuticalsgavegoodfits(R>0.994).Cal- ibration standards weremeasured at thebeginning and at the endofeachsequence.Tocheckthesignalstability,onecalibra- tionstandard(100gL−1)wasinjectedrepeatedlythroughoutthe sequence.Solventblanksconsistingofacetonitrilewereprepared
torunaftereveryfivesamplesformonitoringtheinstrumental background.
Recoveries wereevaluatedat three levelsof fortificationfor eachtypeofwatersample.Theobtainedresultsaresummarized intablesSM3aandSM3b(Supplementarymaterial).RSDvalues lowerthan10%wereachievedforallwatersamples.Generally, goodrecoverieswereobtainedforall thestudiedpharmaceuti- cals(>50%),exceptforacetaminopheninalltypeofwatersamples andforsulfapyridine,sulfamethazine,sulfamethoxypiridazineand sulfadimethoxineinseawater.Recoveriesrangedfrom50.5(sul- famethoxazole)to100%(ketoprofen,nimesulide,andparoxetine) forbottledwater,50.1(sulfamethoxazole)to98.0%(ciprofloxacin) fortapwater, 56.6(sulfadiazine) to109%(naproxen)inseawa- ter,47.8(sulfamethoxazole)to102%(ciprofloxacin)forriverwater, 50.8(10,11-epoxycarbamazepine)to129%(acetylsalicylicacid)in WWTP effluent, and 51.6 (10,11-epoxycarbamazepine)to 122%
(acetylsalicylicacid)inWWTPinfluent,respectively.Thedeveloped methodshowed good reproducibility, withRSD values ranging from0.05to10%inallmatrices(TableSM3,Supplementarymate- rial).SimilarrecoverieswereachievedbyM.Grosetal.[33]for surfacewater,WWTPeffluent, andinfluent,rangingfrom60to 142%,from62 to121%, andfrom50 to151%,respectively, and byCaietal.[11]fordrinkingwater,showingrecoveriesbetween 61.4-124.3%formostoftheselectedpharmaceuticals.Theobtained recoveries for seawaters are also in agreement withliterature, whererecoveriesfrom26.6to229%werereported[34].
MDLsandMQLsvarieddependingontheaquaticenvironmen- talmatrixconsideredandhighervalueswereachievedforWWTP effluentandinfluent(TableSM4,Supplementarymaterial).Ingen- eral,MDLs rangedfrom 0.02ngL−1 forsalicylic acid, naproxen, andnimesulide,inbottledwater,to185ngL−1 for norsertraline inWWTPinfluent,andMQLsvariedbetween0.04ngL−1 forsal- icylicacid, inbottledwater, and562ngL−1 fornorsertraline, in WWTPinfluent.ItshouldbehighlightedthatthehighestMDLsand MQLswereobtainedtothemetabolitescarboxyibuprofen,hydrox- yibuprofen,andnorsertralineandthepharmaceuticalsibuprofen andciprofloxacin(upto185and562ngL−1,respectively).
Threedifferentconcentrations(25,50, and100gL−1)were usedand twelvesuccessiveinjectionsin one dayandtriplicate injectionsinsixconsecutivedayswereperformedfortheintra- andinter-dayprecision.Theoverallmethodprecisionwassatis- factory,withRSDvaluesrangingfrom0.6(hydroxyibuprofen)to
Table 1a
Concentration ofpharmaceuticals,inngL−1,detectedinbottledwater,tapwater,andseawater.
Pharmaceuticals Concentration(ngL−1)±RSD(%)
Bottledwater Tapwater Seawatera
BW1 BW2 TW1 TW2 SW1-B SW2-B SW3-G SW4-G SW5-Y SW6-Y
NSAIDs/analgesics
Salicylicacid 30.6±5.12 21.2±4.23 66.0±5.10 39.4±1.91 169.1±4.28 91.3±2.37 73.7±5.90 92.6±5.15 73.4±4.18 137.5±0.89
Acetylsalicylicacid n.d. n.d. n.d. n.d. n.d. n.d. n.d. 5.12±7.13 n.d. n.d.
Acetaminophen n.d. n.d. n.d. n.d. 224.6±0.37 98.6±3.81 53.2±6.37 156.3±6.04 269.7±0.61 59.4±3.06
Hydroxyibuprofen n.d. n.d. n.d. n.d. 29.0±2.87 30.8±7.71 27.3±1.79 30.3±8.12 98.9±5.62 27.6±8.57
Carboxyibuprofen n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 270.4±3.97 n.d.
Ketoprofen n.d. <MDL <MDL <MDL 11.3±6.41 17.7±6.98 11.2±0.52 11.9±5.94 12.9±4.97 12.3±5.61
Naproxen n.d. n.d. n.d. n.d. 17.5±1.05 n.d. n.d. 23.9±2.87 177.7±6.59 17.4±0.16
Nimesulide n.d. n.d. n.d. n.d. 5.2±9.12 n.d. n.d. n.d. n.d. n.d.
Diclofenac n.d. n.d. n.d. n.d. n.d. n.d. 2.36±2.87 n.d. 3.99±4.76 n.d.
Ibuprofen <MDL <MDL <MDL <MDL 27.6±0.47 23.8±0.63 32.7±3.39 33.6±7.73 40.6±4.10 9.4±4.3
Psychiatricdrugs
Venlafaxine n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Trazodone n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Citalopram n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Paroxetine n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Norfluoxetine n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Norsertraline n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Fluoxetine 0.27±5.66 n.d. n.d. 1.90±1.24 n.d. 0.27±5.96 0.25±6.24 0.74±5.16 0.27±1.81 n.d.
Sertraline n.d. n.d. n.d. n.d. n.d. <MDL n.d. n.d. n.d. 8.09±9.18
10,11-Epoxycarbamazepine n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Carbamazepine 22.1±2.32 n.d. 22.3±1.52 20.0±0.580 n.d. n.d. n.d. n.d. n.d. 28.3±3.12
Diazepam n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Antibiotics
Trimethoprim n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Ofloxacin n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Ciprofloxacin n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Enrofloxacin n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Sulfadiazine n.d. n.d. n.d. n.d. <MDL <MDL <MDL <MDL <MDL <MDL
Sulfapyridine n.d. n.d. n.d. n.d. <MDL <MDL <MDL <MDL <MDL n.d.
Sulfamethazine n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Sulfamethoxypiridazine n.d. n.d. n.d. n.d. n.d. n.d. <MDL <MDL <MDL <MDL
Azithromycin n.d. n.d. n.d. n.d. <MDL <MDL n.d. <MDL <MDL <MDL
Clarithromycin n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
Sulfamethoxazole n.d. n.d. n.d. n.d. n.d. n.d. 1.65±17.4 <MDL n.d. n.d.
Sulfadimethoxine n.d. n.d. n.d. n.d. n.d. <MDL <MDL <MDL <MDL n.d.
aBeachesclassifiedas:B-“excellent”withablueflag,G-“good”withagreenflag,andY-“sufficient”withyellowflag.
Table1b
Concentrationofpharmaceuticals,inngL−1,detectedinriverwater,WWTPinfluentandeffluent.
Pharmaceuticals Concentration(ngL−1)±RSD(%)
River WWTPInfluent WWTPEffluent
RW1 RW2 IW1 IW2 EW1 EW2
NSAIDs/analgesics
Salicylicacid 89.2±9.13 128.2±8.93 6332.3±3.37 33535.5±5.15 186.7±4.25 126.9±7.90
Acetylsalicylicacid n.d. n.d. n.d. n.d. n.d. n.d.
Acetaminophen <MDL 4.9±12.57 30030.4±1.36 615134.9±6.94 736.0±6.80 2139.0±7.79
Hydroxyibuprofen <MDL n.d. 190.2±0.44 198.4±1.27 284.7±0.97 358.7±1.23
Carboxyibuprofen n.d. n.d. 41554.0±5.03 120365.0±4.68 n.d. n.d.
Ketoprofen <MDL <MDL <MDL <MDL 22.3±0.21 55.9±0.35
Naproxen n.d. <MDL 2078.7±5.17 533.3±2.89 <MDL 110.7±8.82
Nimesulide n.d. n.d. n.d. n.d. n.d. n.d.
Diclofenac n.d. n.d. n.d. n.d. n.d. n.d.
Ibuprofen <MDL <MDL 4389.3±7.44 14124.8±8.52 517.4±8.23 323.7±9.68
Psychiatricdrugs
Venlafaxine n.d. n.d. <MDL 15.4±4.19 91.9±6.45 170.9±0.19
Trazodone n.d. n.d. <MDL 17.7±8.96 7.5±7.94 21.7±11.7
Citalopram n.d. n.d. n.d. 15.1±1.37 26.1±7.06 61.4±7.98
Paroxetine n.d. n.d. n.d. n.d. n.d. n.d.
Norfluoxetine n.d. n.d. n.d. n.d. n.d. n.d.
Norsertraline n.d. n.d. n.d. n.d. n.d. n.d.
Fluoxetine 3.3±0.85 3.7±3.92 5.2±1.52 8.8±1.15 12.9±6.78 27.5±4.93
Sertraline <MDL <MDL n.d. n.d. <MDL n.d.
10,11-Epoxycarbamazepine n.d. n.d. n.d. n.d. n.d. 88.0±3.36
Carbamazepine 32.9±3.27 34.4±0.16 66.2±4.54 110.9±6.12 98.5±6.88 244.9±3.34
Diazepam n.d. n.d. n.d. n.d. n.d. n.d.
Antibiotics
Trimethoprim n.d. n.d. n.d. n.d. n.d. 59.3±13.0
Ofloxacin n.d. n.d. n.d. n.d. n.d. n.d.
Ciprofloxacin n.d. n.d. 118.9±0.30 n.d. 96.6±9.29 n.d.
Enrofloxacin n.d. n.d. n.d. n.d. n.d. n.d.
Sulfadiazine n.d. n.d. n.d. <MDL n.d. n.d.
Sulfapyridine n.d. n.d. n.d. n.d. n.d. n.d.
Sulfamethazine n.d. n.d. n.d. n.d. n.d. n.d.
Sulfamethoxypiridazine n.d. n.d. n.d. n.d. n.d. n.d.
Azithromycin n.d. n.d. n.d. 67.0±4.59 n.d. 11.4±12.3
Clarithromycin n.d. <MDL n.d. n.d. n.d. 70.4±5.75
Sulfamethoxazole n.d. n.d. n.d. 224.1±4.30 n.d. 73.4±9.42
Sulfadimethoxine n.d. n.d. n.d. n.d. n.d. n.d.
7.73%(norsertraline)forintra-dayand4.29(salicylicacid)to11.1%
(norsertraline)forinter-dayprecision(TableSM5,Supplementary material).
Matrixeffects(ME)isonesignificantdrawbackinESIMSquanti- tativeanalysis,becausetheESIsourceishighlysusceptibletoother componentspresentinthematrix.TheMEcouldbedefinedasthe changeinUHPLC–MS/MSresponseofananalyte,bysuppression orenhancementofthesignal,causedbycoeluting matrixcom- pounds,relativetoaninjectionofapurestandard[35].MEwere evaluatedfor thedifferent aquaticenvironmental matrices and resultsareshowninFig.4andFig.SM3(Supplementarymaterial).
ItwasobservedMEinallthestudiedmatrices,which,ingeneral, wasexpressedasanionsuppressionforalmostallthepharma- ceuticals.Althoughcompoundslikeciprofloxacin,sulfamethazine, sulfamethoxine,ofloxacin,norsertraline,norfluoxetine,venlafax- ine,diazepam,ibuprofenandnimesulideshowedionenhancement fordifferentmatrices,mainlywastewaters(WWTPinfluent and effluent).Inthecaseofibuprofenandnimesulide,theionenhance- mentwasmore pronouncedin seawater(Fig.4).Thiscouldbe justifiedbythehighsaltcontentofthismatrix, suggestingthat thesaltresiduesmightstillbepresentinthesampleextract,co- eluting withtheselected analytes[36–39]. On the otherhand, ionsuppression wasusuallymore evidentfor matrices suchas WWTPeffluentandriverwater,beingthiseffectmorepronounced forNSAIDs/analgesics(Fig.4).Nevertheless,antibioticslikeclar- ithromycin,azithromycinandmostofthesulfonamidesaswellas theantidepressantsvenlafaxineandcitalopramandthemetabolite
norsertralinealsohadamarkedionsuppressioninWWTPeffluent (Fig.4).
EvaluatingthegeneralMEobservedfor thedifferentaquatic environmentalmatrices, it wasnoticedthat MEwere lesspro- nounced(<20%)inbottledwaterandtapwaterforthemajorityof theselectedpharmaceuticals(67%and58%ofpharmaceuticalsfit inthislevel,respectively),whileforsurfacewaters,WWTPinflu- entandeffluent,morethan50%ofthepharmaceuticalshad ME higherthan20%.Infact,duetothehighcomplexityofwastewa- ters,approximatelyone-thirdofthecompoundsshowedMEhigher than50%(Fig.SM3,Supplementarymaterial).
3.4. Applicationtorealsamples
Atotalofsixteensamplesembracingdifferenttypesofwater were analyzed using the developed analytical methodology. In Tables1aand1b,theconcentrationofthepharmaceuticalsdetected in each analyzed samples is in bold. It was possible todetect pharmaceuticalsinalltheconsideredaquaticenvironmentalmatri- ces, including drinking water. In fact, pharmaceuticals suchas carbamazepine, fluoxetine, ibuprofen and ketoprofen, and the metabolitesalicylicacidwerefoundinbottledandtapwaters,at concentrationsupto30.6and66.0ngL−1,respectively(Table1a).
Concentrationswerehigherintapwaterwithsalicylicacid,carba- mazepine,andfluoxetinereachingconcentrationsof66.0,22.3,and 1.97ngL−1,respectively.Similarconcentrationsofcarbamazepine weredetectedindrinkingwaterofChina[11],whileforsalicylic
